It's an opinion piece this week! Been a while since I've done one of these. ■ AD: 👨💻 NordVPN's best deal is here: nordvpn.com/tomscott - with a 30-day money-back guarantee.
It's weird that 10 years ago, people typing into Google "can you please tell me what the weather will be like this week thank you" was seen as weird and not appropriate, because why would you talk to your search engine? But reading the transcript of the conversation between Tom and ChatGPT, that's how it talks back to you...
chatgpt is like a parrot who talk it know how to get the right reaction but it does not know what it says it can't it only is using machine learning that was learned by giving it a curated list of what to learn from & what not
@@abyssaljam441 robots are expensive due to resources needed to build them, meat on the job turning knobs and bolts are probably not going away for a while
People mine and Tom's age were like that too. Our schooling was in an analogue world for a digital one that began to exist almost exactly at the time we entered the workplace. Our life training was literally in the wrong world for that time.
I feel, like a 3d artist, that my days might be numbered. Companies wouldn't skip a beat if they could replace artists with an algorithm. I hope I'm wrong. I actually love what I do...
Teleprompters make it a lot easier. I hate the vlogging trend of jump cuts. Tom's style is very much the opposite and it screams authenticity, I trust that he knows what he's talking about because he can riff on it non stop. I know he's a brilliant human being but I do think he uses a teleprompter, that's not a bad thing though.
@@mntucket7410 They definitely do, but it's still very difficult to read without stumbling on words or pausing too long while using inflections at the right time and all the other things that make the reading and video seamless. The longer the video, the harder it is. He's been doing this for many years now and he's got it down really well
@@mntucket7410 Yeh, I think he might. But he does it in a way where he could just be speaking off the cuff. @TomScottGo do you use a teleprompter or is it all off the cuff?
It's amazing how quickly they went the way of the dodo after the settlement, while nowadays we can still intensively torrent without issue using different clients.
I'll never forget when my ex accidentally deleted all of my music I downloaded over about 3 years! I actually cried! With a 56k modem it took thousands of hours to download them
I went onto chatGPT and got it to write a simple powershell script to find the sizes of folders and it did a workmanlike job of it, coming up with one that did the job, but slowly. I wasn't impressed. Then I asked it to find a quicker way of doing it, and it correctly used a hash table making it about 10x quicker. That was when I realised how important this is.
About 20 years ago I was taking a programming class and had a conversation with the professor about how cool it would be to be able to just tell the computer what you are trying to do, and it does it. He said that would be impossible to code a program to do that because it would have to understand your language and dialect. The future is now.
I have wished that computers would be like in Star Trek and we could just tell it to do things and it would do them. Some thing seem like they should be simple, but they aren’t. A number of years ago, I was going to put a bunch of folders full of articles on CDs; a task that I thought would be easy. But then I discovered that a bunch of the articles had names that were too long and they all had to be re-saved with shorter titles before burning on CDs. It would have been so much easier to just tell the computer to find those articles, shorten the titles, and burn to CDs.
@@keirfarnum6811 Well, we are not far off from that. Give it a few years and the newest versions of speech assistents are going to be frightening good.
A programming language is literally the way to tell the computer what to do. The problem with that is, that the computer dies what you tell him and not what you mean.
He wasn't wrong, they still can't understand language and dialect. It's only that processing power, storage, connectivity, and bandwidth over the internet has become so much exponentially larger and relatively cheaply available that a program can now pull from a massive data set, run calculations based on given parameters and by probability and past failures (that it was told was failures) can produce what most humans ( by their recorded "that's correct" responses to it's output) most likely expect based on given parameters.
I'm a nuclear engineering student, I used chatGPT while studying and that's my experience: On one hand, it was really good for generic stuff, like "ask me something about nuclear engineering" (I needed some random questions to prepare an oral exam). On the other hand, it was awful at giving any specific knowledge, responding in a vague way or just completely and absurdly wrong.
right now its just a language model trained on allready written/crawl ans searchable text in the internet. Verry specific knowledge is often behind paywalls, behind DRM in onlinebooks and behind university access barriers... and also sometimes still on paper in books in libaries. and yes it can be wrong in a verry strong way. there is no recursive algorithm to check for facts or a knowlege database ii uses. but.. give the system a basic database of known knowllege as a strong data point to use. give it a feedback option from the users you can simply say nope you are wrong here... Its clear this will be the next step and at least a feedback loop will be implementet to harvest a lot of big data out of it.
12:36 weird question... but do you listen to Weird Al? You listed the exact same music downloading websites and in the exact same order as Al Yankovic does in "Don't Download This Song".
@@jeremykothe2847 Yea, but will that still be the case in 2030? 2024? Maybe next month even? The second it can get itself "unstuck", the potential becomes unimaginable.
I know this wasn't the point you were making but in the back of my mind I kept thinking of the Douglas Adams quote: "I've come up with a set of rules that describe our reactions to technologies: 1. Anything that is in the world when you’re born is normal and ordinary and is just a natural part of the way the world works. 2. Anything that's invented between when you’re fifteen and thirty-five is new and exciting and revolutionary and you can probably get a career in it. 3. Anything invented after you're thirty-five is against the natural order of things."
@@MyName-tb9oz yes. Good comedy is not inventing humour. Just pointing out how funny and ironic life on this planet actually is. And adams was one those really good at pointing.
This feels so honest. In my opinion it's rare to see this vulnerability, specially from such a prominent person, and I think there's a great deal of beauty in it. Thank you Tom
My primary fears are not what jobs it will take away or what it might get wrong, but how it will be essentially weaponized and commercialized by those who would like to profit off of its misuse.
You're not wrong. This is the beginning of a major shift. As a person with no coding skills, I was not able to get ChatGPT to come up with the correct code to make a very simple, but functional indicator for a program called MT4. It said it could code in the required language, but no luck. So, I think it takes a human with coding skills, to be able to evaluate the output and makes fixes as you did.
Programming is both a art and a science. ChatGPT helps the programmer with repetitive tasks but can not entirely create an app made of millions of code lines ! The prompt for this app would contain thousands of lines ! But prompt engineering is now sought and pay well ($ 200 000 per year) 😮😊
I think the big difference is that now the barrier to producing code in a new programming language is much reduced. I know how to program in Python, but I don't know JavaScript and CSS that well. Now that I've been working on a web app I need those languages for my front end. ChatGPT can create a program that gets me most of the way there. If there's an issue with the syntax, then I'm going to take a long time to notice the issue, syntax is what I'm trying to get ChatGPT to do for me, but I can assure myself the flow of logic is correct because I learned about generic programming topics like algorithms and data structures in Python, but are equally applicable to all programming languages.
My pit in the stomach moment was yesterday when my mother, who is not a big tech at all, easily won an argument against my sister about whether or not to allow the dog on the bed. by asking chat gpt for good arguments and comebacks. my sister, who did not realised what was happening was complettly flabargasted and left speachless.
I feel the anger about Gmail labelling, you are not alone Scott, I ALSO realized that labels only affects a single message in the thread and it KILLS me.
The problem goes away if the emails are unthreaded instead of being joined into a single thread. Each email becomes its own unit that can then be sorted.
"if you're under 25 you don't understand how fast this all happened" it's so true. I0m 23, bit younger, but i can feel the difference when i talk to someone who is 16/17. The way they are one thing with social media and their phone it's absurd, but what's even weirder is that it does NOT directly translate to tech skills. We're managing to spend so much time on devices not learning anything about them
very interesting point; we seek information more than ever before in today's society yet most also don't seem to ask "okay, but how/why does this work?" I think my favourite question ever is the simple "why?" just that. why don't people look at the fact they have stared at instagram or tiktok or whatever and go hummmm, why is it soo addicting or take up soo much of my time? how does it know what to feed me? It shocks me how little of my peers say that (i'm 20)
That's because manufacturers and developers are trying as hard as they possibly can to obscure the inner workings of their devices. The solution is to use older software & devices.
It's predictable really that using phones rather than computers would lead to decreased tech literacy, as the smartphone hides a lot more of its functionality than a PC does.
I've long been skeptical of the contention that new machine functionality won't eliminate jobs but will merely make current employees more productive. Greater productivity from each employee clearly leads to needing fewer employees for the same task. e.g. automated telephone switchboards may have initially just made operator jobs easier, but they eventually replaced operators. (Have you met a telephone switchboard operator lately?)
Exactly this. Funnily enough, it reminds of a Frankie Boyle joke - 'in a capitalist society, technology isn't going to make life better - instead of earning good money as a sex worker, that same person will instead be earning minimum wage cleaning the c*m-grates of a hundred sex robots'. If technology makes our jobs easier/does our jobs for us, the powers that be will just see that as a reason not to pay us anything.
@@GamePlays_1230 He said he was skeptical of the idea that new machine functionality won't eliminate jobs, then backed up that opinion by giving an example of new machine functionality eliminating jobs. How is that a contradiction?
For the record, you're definitely not the only one annoyed by Gmail labels, there's dozens of us! And as an aside i too, got the same feeling, so if you're wrong, you won't be the only one.
The one thing that strikes me about this video is the fact that nowhere in his path did Tom question that Google just might have implemented it the wrong way :D
"That horror, that dread is that in a couple of years, my world is about to change. And despite everything, I will still want my emails to be in folders" The sheer power of that sentence, it's unbridled. You put into words what I couldn't. The Dread of Change.
Exactly! This sentence reminds me of how generations work. Millennials pick on boomers for being stuck in the past. Gen Z picks on Millennials for being stuck in the past. Gen Alpha will pick on Gen Z for being stuck in the past. Some day Gen Beta will pick on Alphas. Change is the way of the world but we dont change as quickly. This sentence alone encapsulates politics and business in a beautifully succinct way
I think it's more like the dread of a very specific change for instance, when was the last time you were just randomly out and ran into a friend because they just happen to frequent the same places you do because you share their interests?
You know... I'm an old programmer since I was a little boy with an Altair... We used to say about computers; "computers do what you tell them to do, not what you want them to do." It has held true all these decades. It may not be valid much longer.
nah - this kind of stuff works for very limited scripts... something that effectively has to do one thing and will work with "off the shelf" bits as tom mentioned...as soon as it gets more complex it completely falls apart and the thing is... you always have to verify whatever it produces... and the bigger the codebase, the harder that gets... and automatic optimizations can and will produce incredibly obtuse sourcecode thats even harder to verify... and problems like that compound
That I think won't be solved by ChatGPT. Automation, or even human delegation, always walks a fine line between correcting minor mistakes and substituting the original specifications for a wrong interpretation based on the agent's mistakenly confident assumptions.
you're a smart guy tom. if things change, you will change with them. something similar has happened to me a few times in my 72 years, and i was lucky/clever enough to jump ship, roll with the punches, or whatever cliche/metaphor you like, and find a new direction. not always easy, but the best things aren't. Courage mon frere😀
Sucks to struggle for 10 years and any time I've "made it", something happens within 6 months to lose it again. Meanwhile other people are like why are you trying so hard, just get the job I did and you can do the same thing for 30 years. Awesome.
I started to realize the gravity of the situation when Open AI's CEO started talking about how underdeveloped ChatGPT is compared to what they're planning to release, even calling it their "worst product." Such crazy times we're living in.
@@grieferoncamera4600 Not really, it's more that GPT-3 is much more powerful but requires the user to adapt to it, and ChatGPT is the other way around. It made it available to people who didn't know / believe how amazing it was
Came in expecting another "OMG CHATGPT IS SO SMART NO WAY!" video, instead got a view into a very personal and abstract feeling that spans entire generations. Great stuff.
This just hit me last week. I work in the government recreation industry, and we just started using ChatGP for writing grants, support letters, and press releases. It's insane. What used to take us hours to to write out, is now filled accurately in seconds. Our world is about to shift so radically, we cannot even imagine what's about to happen
That's right, now even fewer people will be creative and successful. We will have mindless, brainless, chatgpt monkeys, and then a few men still using their own minds. How can it benefit you if there is no barrier to entry and everyone can spit out the same generic chat GPT trash? Things are shifting, and they're shifting in favor of men that can still engage in unique creative activity.
Do double check them. One flaw I've found with ChatGPT is that it can plagiarise quite readily. I asked it to write me a story about a boy who goes to a magic school and it literally paraphrased Harry Potter with all the actual names (Hogwarts, Hagrid etc.) while initially passing it off as its own original work.
@@sophiedogrun I hope so. But I expect that the fact checking will get less and less until chatGPT can rig any founding as it pleases, since we couldnt be bothered mistrusting it.
It's strangely comforting to hear that Tom, the person whose opinion I trust when it comes to qurstions regarding technology is concerned about the same thing that I am. I'm 22 y.o. linguistics student and I am afraid that soon my degree might be barely worth the paper it's printed on.
I think it really depends on what you’re planning to do with your linguistics degree. If you are going further into linguistics, and you’re studying phonetics or sociophonetics, programs like ChatGPT don’t have the capability of doing acoustic analysis of vowels and consonants. I’m a graduate student in linguistics right now, and at least that component isn’t directly replaceable by transformers like ChatGPT yet. However, another student in my cohort did ask ChatGPT to write a Praat script for him (which is used to perform acoustic analysis), and it spit it right out for him! He still had to know what questions to ask of it, and even after using the script, that just helps to acquire the data that he needs, and does not perform any of the actual analysis. So, at least in that area of linguistics, models like ChatGPT have fit more into the category that Tom mentioned of “making people’s jobs easier.” Still, you’re right that everything I’ve said must have a “not yet” and “thus far” appended to the end of it.
I feel your pain, pal. I'm also 22. I wanted to become a linguist since the primary school, yet I chose to get a journalism degree just because it looked more viable at the time. "Machines can already semi-competently translate the texts, so soon enough they will improve so much, so they will be able to create translations nearly undistinguishable from professional ones" was I thinking five years ago, "but bot can not possibly produce a good news report, right? Surely, journalists will have a plenty of job in a future, because events are happening each day and somebody needs to cover them!". However, in hindsight, considering the quality of the texts ChatGPT produces even today, it seems like I'm also going to be out of commission soon
@@VasiliyOgniov There's an issue when using ChatGPT or similar language models as a journalist. As they do not understand the words they are writing, there are large risks of misinformation, which is only amplified the more people depend on it, so I do not believe journalists are going to be fully replaced. It's important to understand the limits of these programs, and how there are almost certainly going to be limits on what they can do, which is mostly storage of data, and crucially: the inability to understand, adapt, compare and learn
I have been in a similar situation where social media has destroyed the Internet I grew up on, and it is rightfully dreadful because I can tell you it's miserable. It isn't that social media changed the Internet, it's that it changed the people using it and their values.
I spent 2 years away from home serving a church mission and smartphones got popular while I was gone. I remember thinking it was so strange to see my entire family sitting together but all just absorbed in scrolling. Especially my parents, who has always been quite anti-video game my whole childhood. Whelp they got me one too and now several years later I find myself in the exact same place as them… it’s totally absorbed me without me even realizing it. The compulsion is so strong. Very strange to see how these forces can change people so profoundly, and so quietly
@Carrie Bartkowiak no. Being single now with things like tinder mean many people now find the idea of approaching someone or just getting to know someone in person that you meet without extensive digital communications, off-putting or wrong. That never existed before. It's changed how we interact with each other and TRAINED people into think traditional face to face communication can be wrong or improper. It's also had positive effects and deep negative effects with the ability to just block someone and effectively erase thier existence from your life.
Good to know I’m not alone with the feeling of dread. I’m a software engineer and used to be excited by new technology and thinking about the future. Recently I’ve been feeling deeply nostalgic about the old days and trying to get away from technology and spend more time in nature.
I watched mr. wizard Saturday s and the went outside to play. Making a snow person in winter. And riding my bike in summer. Best of both until I met an Altair 8080.
I trained to become an illustrator for the last 10 years and I get what you mean. I just want to leave civilization at this point. Feels very distopian
A year later your video seems as fresh as yesterday. You explain complex issues simply with personal examples and convey your insights with clarity and compelling emotion. That certainly deserves a like 👍 & subscribe. Well done!
Hey Tom, you're not alone in being EXTREMELY annoyed with how labels work. I deal with those headaches at least twice a week providing support for individuals who can't find their emails because labels don't work like folders. The thing is that GMAIL could actually ADD folders and it would not disrupt their label system at all, as their label system is meant to work more like tags for individual emails. You're supposed to load up an email with as many tags as is relevant to you. NO ONE does this, as it would just be a huge time-sink and the search function works halfway-decently.
"Maybe Siri and the Google Assistant are going to become the things they were always promised to be." I have a feeling this quote is going to age like fine wine.
I can't put into words how refreshing it is that Tom will actually say "I was wrong" when he was wrong about something, even something as minor as "my prediction of the future of natural language processing was incorrect." So many people can't say those words under any circumstances.
@@ziwuri most of us aren’t brought up to learn how to own up to our mistakes. Most parents don’t even have the tools to to do it themselves! Tom is a breath of fresh air that I think hopefully can inspire people beyond the educational information.
Год назад+1
He did a compilation video of his predictions, many of which were wrong: "Ten years ago, I predicted 2022. Did I get it right?"
I always wondered what it was that was going to make me feel like my parents with the internet. This is probably it. In 10 years, I'll have no idea how any of this works anymore and some 15 year old kid will have to hand hold me through it :(
I'm 17 rn and sometimes it's weird how easy it is for me to not understand a tech. My 11 yr old brother's tech is sometimes already too weird to actually understand and interact with well :/
Then try chatgpt, try any of the alternatives. Do it today. It's going to take effort not to become a luddite, but quite frankly, you're already ahead of any of today's 5 year olds, you have no excuse.
I'm presently most worried that people (of low-to-middle skill) won't know enough to catch ChatGPT when it's confidently incorrect, and that lives will be lost as a result. And it's not like that bad interaction doesn't happen with it's just people -- but there does seem to be some bit of one's natural (healthy) skepticism that turns off when they're dealing with tech, rather than people.
My grandpa is very poor in health at the moment. He can barely stand or take care of himself, and his house is infested with all kinds of bugs. So my uncle of course asked ChatGPT for advice. Rather than taking obvious action such as calling an exterminator out there and getting his dad some nursing care, he explained the whole situation just for ChatGPT to tell him roughly the same thing. For someone as tech-savvy as my uncle, it's a bit concerning to see him so uncritical of new technology like this.
Didn't the same happen with wikipedia? People used to say that it was all wrong because anyone could write there whatever they though it was "true". And yet here we are now: wikipedia is the de facto "truth" of knowledge (with citations, pictures and in several languages).
I'm more worried that traditional search engines will die, replaced with chat bots that recite information vetted by the highest bidder/automatically censored by the government.
“I've come up with a set of rules that describe our reactions to technologies: 1. Anything that is in the world when you’re born is normal and ordinary and is just a natural part of the way the world works. 2. Anything that's invented between when you’re fifteen and thirty-five is new and exciting and revolutionary and you can probably get a career in it. 3. Anything invented after you're thirty-five is against the natural order of things.” ― Douglas Adams, The Salmon of Doubt
Or another way I've heard it put: Anything from before you were born is antiquated and barely usable, anything from when you are born till your thirties is normal and leaps and bounds above what came before, and anything after that point is totally unnecessary and superfluous.
I know Tom has Flown into the sunset, but i would find it very interesting to see a commentary video from Tom about Ai now a year later after this. Love this channel and everything about it
I've started to experience chat GPT. It was kind of scary, and you've articulated it perfectly. I never imagined that I'd want to see a sci-fi like technology flatten out and not change the world.
I think it's not that we're scared of change. It's that we're scared of fast change that we might not be able to adapt fast enough to. The same change we humans have caused on the environment that have made many organisms go extinct because they couldn't adapt fast enough.
@@WanderTheNomad Most changes to technology in the past were to help and assist us. This doesn’t feel like it will help us as much as it will compete with us. Compete with us for jobs, for creativity, and who knows what else. It seems as though we are simply finding more ways to make ourselves obsolete with each new “advance” in technology.
@@bderrick4944 that's ridiculous. It is fundamentally impossible to make a human obsolete as a human has no inherent raison d'etre, only circumstantial ones. What you render obsolete is human labor, which is a good thing. The other thing you render obsolete is a system of economic and societal distribution that >requires< active input of labor from every individual, for most of their waking hours, for most of their lives. Which is also good, because it is only because of that strange system that needing less human labour suddenly becomes a bad thing. Humans are not made to work, no matter what the capitalist productivity machine wants. Work, quite literally, was made to benefit humans. we need to move beyond capitalism where the fruits of labour are awarded chiefly to those who already have the advantage (trending invariably towards monopoly and oligarchy), and enter a new society where humanity is actually allowed to flourish through the promised shared bounty of emergent technology, not threatened by its potential.
For Google to mess up sorting in their own app it is both ironic and concerning. I'm guessing they won't ever adhere to their original, expected design.
When I was at university studying computer science 10 years ago we used to be like "It really sucks for all those people who are about to lose their jobs to automation. But surely programmers will always be in demand!" Welp.
@@flybeep1661 you say it so aggressively (maybe by accident) that it looks like you're mad at them for it, when they weren't the ones making the changes and they have no blame in it, and you also don't have proof they made fun of some person because of the programs replacing humans stuff
The weirdest thing is, when I was using ChatGPT I found myself talking to it in a formal and polite mannerism when there was no such need for me to do so. It just feels.. wrong in a way but that just shows how human like it is. Crazy..
Yes, I use "please" when I ask it to do something... then wonder why and the continue to do so. Am I worried that if I stop using the word when I'm asking ChatGPT to something, will I forget when asking a person. Footnote, just checking the sentence and following the prompts of Grammarly to make corrections...
That's... very interesting. To be completely honest I've found myself doing the opposite, hurling abuse at it when it gets stuff wrong because I know it won't react in any other way other than apologizing. I'd never in a million years do that to a person but maybe because I know it isn't real I feel like I'm allowed to express my feelings. Being polite or being rude are equally weird though really.
I usually say "please" too, at least once in the interaction. I really don't think that it feels or anything, but it so often says are things that a creative, intelligent, conscious being with emotions would say (especially if you ask it to pretend to have "heightened emotional sensitivity") that I feel like I have to just in case. I've found myself creating personalities for ChatGPT to emulate too, so that makes it even harder not to feel that way. I've asked it to be Uncle Iroh from Avatar to ask for advice, I've asked it to become "Rodney the Rapping Robot", I've asked it to only respond in Garfield comics. The robot is my friend.
@@ktburger659 38, and I'm definitely on the other side now, and I work with computers. Just catching up is hard, I feel like I just learned a new trick and it's already obsolete, every week.
ok, it's an age competition, my bid: I am more than twice your age, yes, things keep upsetting my world view. Cutting edge, honing my skills etc. long gone. On the other hand bits of my youth (like inflation) keep coming round again - not saying that inflation is economically good but to my generation it is familiar.
A personal note of sympathy with your Gmail labels frustration! I was recently engaged in civil litigation, in which a great deal of evidence was hundreds of emails exchanged between myself, a landlord and various city officials. I discovered when trying to find all of the relevant exchanges - many, but not ALL of which I had originally labelled - the exact same issue you had with missing portions of threads due to labels not having automatic global application. (I ended up spending several trdious days with the advanced search function in the end.) I imagine other folks have needed to collect all exchanges on a certain topic for some reason or other, and run into a similar frustration! I know that's not that point of this video, just the set-up. But I FELT that complaint!
I hate surprises but I don’t mind being wrong so I’m not sure about that. Some people just feel uncomfortable, or maybe it’s some sort of control issue. I don’t know.
The crucial thing, of course, is the chatbots are insensitive to when they start confabulating or bullshitting. Your videos are carefully researched, and you'll pass up on the chance for a cool video if you can't actually confirm things- sometimes you'll make the less flashy video that's about the uncertainty. That's because you're an agent embedded in the world, not just a language model that generates words that sound like they go together.
I don't think that's hard. Just train it to crosscheck a certain amount of scientific sources. You might even give it the thesis you're trying to prove and it'll select the sources accordingly. The focus of chatGPT is to sound human and be creative, not to be correct. Doesn't mean chatbots in general can't be correct. In fact chatbots as they are used in customer service today only give fact checked answers. It's a different underlying technology, of course.
A lot of people are unaware of their own bs too. It's only a matter of time before this language model gets access to GAN-like learning with the help of all users, similar to how little kids are learning when they try to extract general rules out of everything while being supervised and corrected by their parents.
A couple of Saturdays ago, I spent a little over an hour writing a roughly 70 line program in PHP that converts a number like "7261" into "2 hours, 1 minute, and 1 second" minding the plurals, putting the "and" in where required, and with the oxford comma. Then I asked ChatGPT do to it and it did, in seconds, in just 27 lines of code, and so elegant, it was like art. Long story short, I know exactly what Tom means when he talks about the existentially dready pit in the bottom of his stomach. I felt the exact same way. Sure ChatGPT missed the oxford comma, but I can live with that.
“…sure, ChatGPT missed the Oxford comma, but I can live with that.” This sentence gives me that existential dread sensation that Tom was talking about. Sums up the last 200 yrs of technological innovation and the overall human willingness to give up little things that mattered more than we understood, to see what known fears and discomforts our tech might take off our lives. The changes Tom is talking about have more to do with human developmental insufficiency, than the tech we create as we grow.
This video is oddly comforting. Having a direct comparison to things I know (Napster, iPhones, and Gmail) as well as the model of the sigmoid curve makes me feel like there's structure to the existential dread. It removes some element of surprise, and for someone like me, that makes all the difference in the world
Kolard Greene: This video is oddly comforting. Having a direct comparison to things I know (Napster, iPhones, and Gmail) as well as the model of the sigmoid curve makes me feel like there's structure to the existential dread. It removes some element of surprise, and for someone like me, that makes all the difference in the world ChatGPT: It's great to hear that you found comfort in the video and that it provided you with a helpful perspective on the changes happening in the world around us. The idea of a sigmoid curve can indeed provide a sense of structure and predictability to seemingly chaotic situations. It's also great that you were able to connect with the examples mentioned in the video, which helped you relate the concept to your own experiences. Understanding and anticipating change can certainly alleviate anxiety and uncertainty, and I'm glad that this video could offer some solace in that regard.
@@DanieleGiorgino And it is an important difference. The lion is a real thing that can be understood and you can learn how to deal with it. The shadows of things that may be are ill defined and can't be understood or interacted with properly.
we're lucky if the biggest problem we're facing is the end of capitalism which means finding new stuff for people to do that isn't specializing them down a specific career path for their entire life. That is the good outcome, even if it doesn't feel like it for some. The worst and by far most likely outcome is that we end up deploying misaligned AGIs. This is a very well understood existential threat, and the dread is fully justified.
We are totally at the bottom of the curve. I'm finding myself daily saying "Imagine if ChatGPT was trained on {insert non-public source here}, i'd save so much time" and I'm not exactly the most visionary type. I grew up in the same world as Tom, and I'm getting the same vibes of massive change incoming.
@@RareDay that comes next, we're being slowly shown hidden technology that's been around for eons. Welcome to the Matrix, where everyone thinks everything is brand new when it's really old.
The biggest problem I see coming out of this is that people (especially kids) are gonna learn ways to not think for themselves, and creativity as a human quality is gonna plummet
@@KerithanosIt did indeed, didn't it!? No-one of the cohort in question will know what '18%/mid-grey' is. Sure - most don't need to, because the smartphone designers have baked-in certain 'Lightroom' settings - to always create what these engineers have decided what we will want to see. It's amazing - but equally it has created an era where a DSLR/mirrorless camera may as well be a brick, for most folk.
Telling the computer what to do, versus asking the computer for what you want… Going from instructing someone on how to cook you a burger, to just asking for the burger and letting them figure it out.
@@dvol I think they meant the difference between asking a computer to generate a solution to a problem as opposed to using/coding specific processes to solve a problem.
I love the way Tom talks, I can listen to these videos for hours. It's just on the right level between being technical and being sentimental that resonates a lot with me
And you are right to be a bit scared. Lots of people are descending into cope and trying to minimise it and dismiss it based on some mistakes. These AIs don't have to be perfect to revolutionise everything and they don't have to completely replace humans in a loop... Just most of them.
@@winsomehax And the thing is, those mistakes are in the _current_ AIs. Nothing's to guarantee they'll be kept unsolved in whatever new ones are to come.
Right? The potential for misuse and misinformation.. Where you can't tell what's real or not. That's a sci Fi dystopia, and that might be within the next 5 years
A couple of months later, chatGPT 4 is out for consumers, Bing has gotten huge, midjourney 5 is just insane. The talks about alignment, AGI, and governments starting to pass laws... You were spot Tom.
I write romance novels and for years we've always laughed at 'Computer Writes A Novel And It's Bad" articles. Far less laughing these days. Lot more side-eyeing and creeping existential dread.
We love comforting oureselves, telling ourselves there's all these things that only Humans can do. That list has only ever grown shorter, and every time it does, we push back the goalposts.
@@MilitantAntiAtheism The human brain is a machine made of meat. Consciousness is an emergent property of complexity. Computers might get there too, they're just behind us on the path.
I noticed, too, that ChatGPT is really good at writing and altering code according to your inputs. It shocked me as well. On the other hand there are weird gaps in its abilities, for example it cannot write poems in German with rhymes. It has no understanding of what a rhyme is. It pretends to know, but then fails miserably each time to produce any rhymes. (In English that's not an issue by the way.) These gaps remind you that it has just been trained on an unfathomable amount of data. In the end you ask yourself, what is intelligence after all? It seems you can create convincingly intelligent systems by feeding them with huge datasets. If sufficient conceivable conversations have been memorized, then it will appear to be intelligent. For me it's uncomfortable how far that brute-force approach gets you... because ChatGPT in many cases really appears to be smart.
Agree, same with coding in a language it obviously doesn't have a lot of training data on. I asked it for some code on an old scripting language i use, and as soon as it got a bit more complicated it produced extremely wrong stuff. If I asked for the same thing in let's say python, it probably would have done it easily.
You could also argue that our own intelligence is also trained on a huge dataset - our neural networks and capability to produce 'human' heuristics has evolved over millions of years - not too dissimilar.
But wouldn't a computer trained in code suggest the possibility for it to reprogram itself? Would this take it beyond the brute force approach you described?
It's definitely the start of the curve. OpenAI engineers were very confused why their bosses pushed them to release what to them is a two year old program. This is version "3.5"and they're close to v 4
@@Flesh_Wizard I wish CDPR hadn't named their game after an entire genre. Now nobody can speak about the genre without it being immediately conflated with the game.
I share your trepidation. This seems just as incredible as the mass adoption of the internet, or the introduction of the automobile before that. This generative AI technology is going to completely change wide swaths of society. A couple years ago my son (who was just learning basic programming) had asked me why we had to use programming languages, and not just write what we want in English. I explained to him how hard it is for a computer to parse English, and that I thought the idea of just telling the computer what to do in English would be many years away. Now with these Large Language Models, it feels like my son's idea to just tell the computer what to do in English is much, much, closer than I had expected. And I didn't know that labels don't apply to whole threads, that sounds incredibly annoying.
I had a dream that aliens were invading, and they announced this by kidnapping Tom and forcing him to make a video. So when I saw "Everything is about to change" I had to remind myself to actually watch the video before panicking.
You've done a great job of summarising the way I feel about the whole thing. Cleverbot was the source of mockery and "haha it's so stupid" RUclips videos for years. When Talk To Transformer was popular a year or two ago it was amusing enough. Then all of a sudden - GPT-3, then Stable Diffusion, then ChatGPT, and I hope you're strapped in comfortably because the rollercoaster has set off and who knows when the ride ends...
@@foodiusmaximus honestly if that's all it ended up as I wouldn't be so against it. But because of just how much it can be misused and how many livelihoods it threatens there must be full government intervention on who is and is not permitted to use it for specific purposes
yep. when new things like this arrive people always try to downplay the societal changes it may/will cause, both good and bad. "nothing is changing" "stop overreacting" etc. like how long can that way of thinking last
Cleverbot is too newfangled. Santabot is where it's at. Oh... looks like it got shut down for saying things it shouldn't. Poor Santabot, we will no longer be able to ask you about your dress. I'm sure it was lovely.
@@Detson404 I dont think I could disagree more with someone on something. I use AI every single day and I have done for the past 2 years. It really has revolutionised the way I live my life, and not in any small capacity either. A couple examples I can think of off the top of my head: - Being able to just ask a question to ChatGPT and me getting that answer without any of the fluff of 10 google links I have to sift through - If I want to search something up, I wont use google, I use perplexity since it gets rid of all the fluff and is much more visually appealing - If I want to do something more complex I will use Claude 3.5 (New) because its the best model out right now - Being able to summarise and ask questions about a hundred page research papers with Claude's new PDF feature Those are just some of the ways I use AI to improve my life.
I asked chatGPT to write a press release with minimal information about my organisation and the subject matter and I did this in front of our Comms team. You should have seen the colour drain from their faces when it came back two seconds later with a near perfect piece.
@@subject8776 no you have a conversation with it and it remembers previous answers and what you’ve told it. So you can tell it information before you ask it to write a piece using that info
1 year and 4 months later, I can definitely say, that was just the beginning, today is already insane, but I feal like one year later, we can't even think of a world without ai anymore. See you in a year or maybe two, if I forget.
I had this exact same realization as I worked with GPT-3. It's like looking over the edge of a cliff and realizing how vast the chasm of possibilities is. Things are about to get weird and perhaps the concern shouldn't be for GPT-3 specifically, but for what systems come after it.
I started university the year Napster came out. Hold on, and just keep up. You'll get to reminisce about ChatGPT, and you'll get to feel your sense of existential dread with whatever thing is going to pop up in 25 years. :) It's going to be amazing!
@@mirzaahmed6589 Nah genetic engineering is going to be another inflection point. We got probably dozens more inflection points of societal change from tech before we go back to slow progress imo.
I'm 33. I can still tell you how to use the Dewey Decimal System and sometimes, when it's really quiet, I still hear dialup tones in the back of my head. Both of us are still in for A LOT of change.
Back in the beginning of the 90s I was an avid RPG player, and we played a lot of futuristic RPGs. I was creating a very complex character in some cyberpunk-styled game and was discussing with the DM about what that character's beloved posessions were, and I said to him "I'd like to have in my possession a program that can create a song in any genre, a new song, just by me describing it, and it will create an infinite number of songs that sound like what I asked for". Well, here we are.
@@JimboRustles Ressembles that. But that's only if you take it in a negative way. If I can mimic Mozart and Prince a million time I'll be happy, won't you?
@@JimboRustles I've also read a very weird book about all books being written by machines (and a "male" robot pairing with a human in an adventure), can't remember the title.
Really interesting reading the full chatgpt script. Its a really good example of the use of it and what to watch out for. Whilst the conversation flow is similar to a manager checking a juniors work and requesting corrections, its not what is going on, its a text predictor which is why it makes easy mistakes. There is no "truth" or "fact" just predictions.
My main concern is not that things are going to change, it is that things are going to appear like they have stopped changing, that any new idea will be sucked into an algorithm so quickly and distributed so extensively that it will seem like they were always there
Imagine, the attention span of the internet at large has already shortened from years, to 6 months, to a month, to weeks. With this, Intellectual property will be absorbed everywhere, parodied, spit back out, become tired, and it's not even lunch yet.
@@dermaglowpro1720 Look at how new jokes or memes get spread on the internet so fast. It is increasingly hard to bring any new information to discuss with others as information now travels instantaneously with such a wide spread. Any new idea may now be created and duplicated and spread so fast that people may not realize it is a new idea. Just my thoughts.
And you just spoke out loud what was boiling in my brain for the last month... While my "non-IT/ComputerScience" friends are having fun and getting laughs out of the ChatGPT model, I'm sitting and slowly realizing that the big change in our way of life is here, and not 5-10 years away...
My fried is a tutor of Polish for Russian/Ukrainian speakers. When I introduced him to the ChatGPT, I totally didn't expect the way he will use it. My mind was blown just by singular example. When stuff is done by hand, you pay less attention to details and are less creative, because it requires lot of effort that is most often unappreciated. But with transformers, you can outsource complicated things to it, while guiding it towards your goals. For example, when learning verb cases, no sane teacher would write a task as a poem, where each verb form corresponds to a rhyme. While technically it is possible to do, no one would actually bother. But with ChatGPT, it's just 5 queries away. You get a poem onto topic you need, with each verb following rhyme, but that verb being blanked. So pupil, even if he doesn't know what is he doing, still gets guidance, because he can assume the correct rhyme.
I planned to change career and move to IT after seeing DALL-E & Midjourney… And now I don’t even know if this makes sense because of chatGPT… I guess the only valuable work soon is going to be physical labour.
As someone who works in software. I honestly think we're towards the beginning of that upward curve. The reason I say this is that most of these tools were absolutely terrible/didn't exist just before the pandemic started. And now look where we are. The pandemic meant lots of nerds had extra time to figure out these complex problems and as a community a lot of problems were resolved.
When DeepMind threw away opening moves in Chess that the Chinese have been using for thousands of years (they'd lead to guaranteed loss), and when Go experts said looking at AlphaZero play against itself feels like looking at alien technology, due to the shocking new revelations, I knew far above human intelligence is achievable in other fields in the near future. I don't see it having anything to do with the pandemic....
I'd like to pretend I use them (and I absolutely do use hierarchal file structures rather than one big mess and the search feature for my files, and think whoever was responsible for getting rid of the old windows 9X start menu in favour of the one found in windows 8 and later should never work in UX again), but in practice I have 'unread' 'read' 'junk' 'deleted'. To be fair, the way my e-mail account gets used, this is actually perfectly sufficient.
I just delete everything after I read/resolve it. I have probably 20 emails from the past decade or two that I've bothered to move into another folder and save.
Tom Scott has got to be one of the most relatable, human feeling educational youtubers I watch. Perhaps it's the way he describes his own feelings about things without tainting the actual facts with those feelings. I've had this exact same gnawing uncertainty about the future for years now, as I'm sure many have, but had never placed it to a question like "Where on the sigmoid curve are we?" Stuck in a weird place of not wanting to be a "boomer" who can't accept new things, but not wanting to feel like the world I know is going to fundamentally change either. The mark of the older millennial I suppose, or any younger-middle-aged person perhaps. It comes in waves. "Oh no everything's going to change! Humans won't have any reason to exist anymore!" to "Nah... It can't change that much can it? Everything's going to be fine." On the one hand, it's nice to know I'm not one of the only people who gets this gut twisting feeling, and on the other, it really really sucks to be sent back to the former side of the wave by watching this. xp
I'm thinking the same - that Tom is so insightful to have realized it and put it into words. I've been aware of the singularity idea for years, but didn't find ways to put my concern into my own words. This curve may not be a sigmoid. I think we're all going to be watching an exponential curve. Or will be someday "soon", wherever we let the AIs modify themselves. See Spike Jonze's Her, for instance.
Back in the 1960s Isaac Asimov (author of I, Robot) was asked what he thought about being replaced by a machine. He replied that in regards to his capacity as a dish washer he was all for it but in regards to his capacity as a novel writer he did not like the idea.
@@InternetEntity Tom is using it as a "turning point in technology" rather than the point beyond which we cannot see. The singularity may be where this leads to, but a Napster Point, is when technology being adopted starts to accelerate and change things. The warning bell for those that can hear it, that big things are coming. It may include the singularity.... Or maybe not, who knows.
Oh geez. I wondered why my backed up emails were missing, along with my backed up google drive files. I've done the same as you - filed in folders because that's what I've always been used to. Thank you for the explanation!
In case no one’s pointed it out yet, Tom’s list of file-sharing sites at 12:35 “Morpheus and Grokster and LimeWire and Kazaa” has the same order as in Weird Al Yankovic’s “Don’t Download This Song”.
I did see the internet go "from nothing" to what it is now, and I did see a millennium go, and a new one come. My nephews possibly won't even see a century go and a new one come. And as for new technology? I don't know.
I have learned that every time I think this, the reality is a week long project of bite sized iteration. It is never as easy as copy paste top google search results.
In terms of applications of machine learning we are probably somewhere near the napster point, but in terms of capability and quality of the networks I feel like we are closer to midway on the sigmoid. The insane level of progress in the last few years has mostly come from increasing the size of the models and training data by orders of magnitude over a really short period of time. So much so that I feel like we've already caught up to what's possible computationally, training wise. So progress in model size will probably slow down to the rate of computer performance increases soon.
yup, the actual underlying science of how these are done hasn't changed in decades...it's the ability to train vastly bigger models that is making this sort of progress. Though that could be comparing the internet in the 80's (which technically existed) to the boom of the late 90's/early 2000's when it became widely available and viable.
Quantum computing is already starting to come around and, while it's currently in a very Allan Turing stage of development, it's not gonna be twenty years until we go from where we're currently at to all forms of modern cryptography being obsolete.
It's not just the hardware and model size, there are a lot of innovations happening in the optimal structure of the models, how to most efficiently train them, etc. I doubt we're past the half way point of the sigmoid curve there. Adoption is also going to be wild, imagine people pulling out their phone and saying "Siri, argue with my sister about dogs", or even using AR to have a model giving you ideas on how to make a comeback to the person you're currently talking to. It's possibly going to be awesome to watch humanity suddenly be augmented by this new tech. The internet is already our collective exobrain, after all.
You're definitely NOT alone concerning Gmail's crappy labeling! I asked the official support forum for help on the matter a few years ago (to use labels more like folders) and I was bluntly told by the person "helping" me that I was using an antiquated way of working, and it wasn't his fault that I was a "dinosaur" (his actual words)...!
@@RiahGreen many people these days (who never experienced a world with paper filing systems) don't use folders anymore. Search functions are so powerful that "everything in one bucket and you search for what you need" is the more common way to manage files these days. Personally I use a weird combo of both folders and giant bucket searching, which probably reflects my being on the cusp of gen Z and millennials
I feel your pain. Its like so so much at the moment. You raise a perfectly legitimate objections to something, only to have someone repeat the previous propaganda like this solves the problem somehow. More practically, the best way around this i have found is to use the rules/filtering to label each email as it comes in. I usually do it based on the from address. From x, add label y.
Whether or not using folders is "antiquated" or sub-optimal, that doesn't excuse the unnecessarily rude response. I've more or less accepted the reality that technology is going to move on and drop the old familiar systems that I preferred, but I'm always going to insist on good manners.
I'm 16 and Gmail is the only email service I have ever used, and I still use labels like folders to sort out all of my school emails. The best thing I can do is attach one label to a thread and remove the inbox label. I've even set up filters so that certain people (like class teachers) get filtered by skipping the inbox but getting applied that label. Works for now but wow it could be so much easier.
I recently had to make some promotional videos of myself talking for my company, and I was amazed at how hard it is to talk candidly and casually on camera and not have it come across as artificial, like you're just a robot reading off a prompt. Watching Tom in this video I am amazed at how natural he is on camera, talking for incredibly long unedited cuts as naturally as if he was in the room with me. And his speech is so clear and straightforward, having tried it myself now I can't help but wonder how he does it. Tom if you ever need a video idea please do one on your online presentation skills!
Definitely, I've always been amazed at just how good Tom is at what he does. Not just his presentation skills, but honestly, even his english skill is something to envy.
I guess it would come down to lots of practice and refinement while also getting feedback, I for one was unable to do school presentations, I wouldn't say I'm great at presentation now, but I definitely improved
“I've come up with a set of rules that describe our reactions to technologies: 1. Anything that is in the world when you’re born is normal and ordinary and is just a natural part of the way the world works. 2. Anything that's invented between when you’re fifteen and thirty-five is new and exciting and revolutionary and you can probably get a career in it. 3. Anything invented after you're thirty-five is against the natural order of things.” - Douglas Adams
Interesting way of viewing. I would tweak it a little and say that the dread can start way before 35. I’m 27, my sister 25 we are both not viewing it as something exciting, at least not for what it represents to the world.
I'm older than 35 and find it very exciting. But I also grew up when at the start of computers becoming a thing, and the internet, and smartphones, so maybe I'm more adjusted to change. I understand that some people find it threatening in terms of their livelihoods, but in most ways I think those fear are largely overstated. Life isn't static, we always have to adapt and change to new things in order to thrive-no matter how old we are.
@@shadowjuan2 In theory, it can represent a future where humans no longer have to toil their entire lives in thankless, repetitive jobs, where value can be created for next to nothing by machines that have no aspirations outside of their tasks, and distributed to the population as a whole. We've already passed the point where automation creates more wealth than a country needs - we just allow it to be hoarded by individuals instead of given to the humans who enabled it. Exactly how dystopian or utopian society becomes will have something to do with luck and a lot to do with how easily people allow themselves to be scared into siding with the hoarders, and how many of the rest vote.
![[#2024MAMA] BIGBANG (빅뱅) - 뱅뱅뱅 (BANG BANG BANG) + FANTASTIC BABY | Mnet 241123 방송](http://i.ytimg.com/vi/EQsYfiPR9uM/mqdefault.jpg)
It's an opinion piece this week! Been a while since I've done one of these. ■ AD: 👨💻 NordVPN's best deal is here: nordvpn.com/tomscott - with a 30-day money-back guarantee.
tom scott
How was this pinned 5 days ago?
5 days ago?
Im poor
5 days ago…
It's weird that 10 years ago, people typing into Google "can you please tell me what the weather will be like this week thank you" was seen as weird and not appropriate, because why would you talk to your search engine? But reading the transcript of the conversation between Tom and ChatGPT, that's how it talks back to you...
It's only because it's designed to mimic our conversations. It did not come up with conversation as we know completely on its own.
@@stitcherlives as Tom said: word prediction based on humans data
@@stitcherlives The point is why they don't program Google search like this 10 years ago, because they CAN'T.
Years ago, my grandpa used to ask thinks like this to google
chatgpt is like a parrot who talk it know how to get the right reaction but it does not know what it says it can't it only is using machine learning that was learned by giving it a curated list of what to learn from & what not
It’s lowkey terrifying being in college trying to plan what I want to do not knowing the world I’m heading into
So I'm about to graduate with a job in naval architure. And I feel like ow god, how many mechanical engineers are going to be needed in 10 years.
@@abyssaljam441 robots are expensive due to resources needed to build them, meat on the job turning knobs and bolts are probably not going away for a while
if it's any comfort, it's always been that way. if it seems otherwise, it's just because you have the benefit of hindsight.
The technology sector is rapidly changing, all others will slowly follow that are attached to it.
People mine and Tom's age were like that too. Our schooling was in an analogue world for a digital one that began to exist almost exactly at the time we entered the workplace. Our life training was literally in the wrong world for that time.
I feel, like a 3d artist, that my days might be numbered. Companies wouldn't skip a beat if they could replace artists with an algorithm. I hope I'm wrong. I actually love what I do...
Unfortunately you're correct, along with coders, authors, architects, humans...
maybe in the long term, but before that the tools that execute your vision will change and become a lot more powerful.
@@poopface011 that part as well yes. Silver linings 😆
Depends on the people running the company
There is an IA already that makes 3d game assets
I love Tom Scott's dedication to do a one take for each segment. It's very difficult for those who haven't done it before.
You should watch Lindybeige; he does one take per video.
Teleprompters make it a lot easier. I hate the vlogging trend of jump cuts. Tom's style is very much the opposite and it screams authenticity, I trust that he knows what he's talking about because he can riff on it non stop. I know he's a brilliant human being but I do think he uses a teleprompter, that's not a bad thing though.
@@mntucket7410 They definitely do, but it's still very difficult to read without stumbling on words or pausing too long while using inflections at the right time and all the other things that make the reading and video seamless. The longer the video, the harder it is. He's been doing this for many years now and he's got it down really well
@@mntucket7410 Yeh, I think he might. But he does it in a way where he could just be speaking off the cuff. @TomScottGo do you use a teleprompter or is it all off the cuff?
@@henrikoldcorn Maybe he just really hates editing. Haha
I remember Napster..... I worked there.
It's amazing how quickly they went the way of the dodo after the settlement, while nowadays we can still intensively torrent without issue using different clients.
Wow! I imagine very few can say as much; how large did Napster grow as a company before being sued into the dirt?
Wow that's quite an item to be able to put on your resume
Hello earth to Scott
I'll never forget when my ex accidentally deleted all of my music I downloaded over about 3 years! I actually cried! With a 56k modem it took thousands of hours to download them
I went onto chatGPT and got it to write a simple powershell script to find the sizes of folders and it did a workmanlike job of it, coming up with one that did the job, but slowly. I wasn't impressed. Then I asked it to find a quicker way of doing it, and it correctly used a hash table making it about 10x quicker. That was when I realised how important this is.
Oh god it’s learning
@@fios4528 Workmanlike. Means basic and straightforward.
@@fios4528 WORKmanlike
problem is it steals from other script and it doesnt know where the information came from, plagiarized work with 0 accountability
@@fios4528 I also read it as womanlike. its hard being dyslexic. ChatGPT have made my life so much easier now.
If nothing else, I love that this video is how I discovered that Gmail labels don't just work like folders.
I remember chatbots being a total joke. The last six months have been equal parts exciting, confusing, and disturbing.
buckle up for the next six 😂
@@Hyperion4K and now imagine where we’ll be by 2030…
You think you did but you are a bot
"[...] exciting, confusing, and disturbing."
That is how one describes what 'terrifying' means, yes.
they so extremely funny tho
About 20 years ago I was taking a programming class and had a conversation with the professor about how cool it would be to be able to just tell the computer what you are trying to do, and it does it. He said that would be impossible to code a program to do that because it would have to understand your language and dialect.
The future is now.
I have wished that computers would be like in Star Trek and we could just tell it to do things and it would do them. Some thing seem like they should be simple, but they aren’t. A number of years ago, I was going to put a bunch of folders full of articles on CDs; a task that I thought would be easy. But then I discovered that a bunch of the articles had names that were too long and they all had to be re-saved with shorter titles before burning on CDs. It would have been so much easier to just tell the computer to find those articles, shorten the titles, and burn to CDs.
@@keirfarnum6811 Well, we are not far off from that. Give it a few years and the newest versions of speech assistents are going to be frightening good.
A programming language is literally the way to tell the computer what to do. The problem with that is, that the computer dies what you tell him and not what you mean.
He wasn't wrong, they still can't understand language and dialect.
It's only that processing power, storage, connectivity, and bandwidth over the internet has become so much exponentially larger and relatively cheaply available that a program can now pull from a massive data set, run calculations based on given parameters and by probability and past failures (that it was told was failures) can produce what most humans ( by their recorded "that's correct" responses to it's output) most likely expect based on given parameters.
Those who can, do. Those who can't, teach.
I'm a nuclear engineering student, I used chatGPT while studying and that's my experience:
On one hand, it was really good for generic stuff, like "ask me something about nuclear engineering" (I needed some random questions to prepare an oral exam).
On the other hand, it was awful at giving any specific knowledge, responding in a vague way or just completely and absurdly wrong.
As a biology student I feel the same. High school level biology? Sure, it works. In more specific things it sucks and is incredibly unreliable.
That is because it is still in the progressing phase ?
right now its just a language model trained on allready written/crawl ans searchable text in the internet.
Verry specific knowledge is often behind paywalls, behind DRM in onlinebooks and behind university access barriers... and also sometimes still on paper in books in libaries.
and yes it can be wrong in a verry strong way. there is no recursive algorithm to check for facts or a knowlege database ii uses.
but.. give the system a basic database of known knowllege as a strong data point to use. give it a feedback option from the users you can simply say nope you are wrong here...
Its clear this will be the next step and at least a feedback loop will be implementet to harvest a lot of big data out of it.
It is not the finished product chatgpt will improve
it passed the MCATS
12:36 weird question... but do you listen to Weird Al? You listed the exact same music downloading websites and in the exact same order as Al Yankovic does in "Don't Download This Song".
Yes he did! Check the very end of the video’s description. He literally mentioned that!
once in a while..... maybe you will feel the urge..... to break international copyright law.....
aw Soulseek gets no love :(
Limewire was evil.
he definitely does as he mentioned Al in his previous copyright related video
Don't steal this book
"Tom Scott tries to predict the future, 2023, colorized."
In a few years Tom should do a prediction compilation and see how many he got right.
He won't, we'll have an autonomous system make the video in his stead.
He kind of already did that - "Ten years ago, I predicted 2022. Did I get it right?"
@@zenthr Virtual copy of his head. I hope.
colorized? vr-ified
Relatedly, check the "Predictions for 2022" (or other years) by Astral Codex Ten for this style of work
the fact that you can reason with the code and point out its errors and it will fix them is both fascinating and terrifying to me
how most senior feels but it took 3 days for the interns to fix it instead of chatgpt instantly finding the error and fix it.
It can still easily get stuck in a situation where it doesn't understand, and "flip flops" between two incorrect solutions.
@@jeremykothe2847 Yea, but will that still be the case in 2030? 2024? Maybe next month even? The second it can get itself "unstuck", the potential becomes unimaginable.
It doesn't "fix" them, no. What it does it just keep spitting out new possibilities. It's up to the user to define which possibility it stops at.
Exactly
I know this wasn't the point you were making but in the back of my mind I kept thinking of the Douglas Adams quote:
"I've come up with a set of rules that describe our reactions to technologies:
1. Anything that is in the world when you’re born is normal and ordinary and is just a natural part of the way the world works.
2. Anything that's invented between when you’re fifteen and thirty-five is new and exciting and revolutionary and you can probably get a career in it.
3. Anything invented after you're thirty-five is against the natural order of things."
Adams had a keen eye
Ditto dude. Was thinking the same.
Like any good humorist, Douglas Adams was an excellent observer of human nature. Comedy is something that people drastically underrate constantly.
@@MyName-tb9oz yes. Good comedy is not inventing humour. Just pointing out how funny and ironic life on this planet actually is. And adams was one those really good at pointing.
I love this quote as it is true so so often. And people never catch them selves following the patterns.
14:06 "In a few years, I'll still be working like this" straight made me emotional
At least he says he may not totally be gone
This feels so honest. In my opinion it's rare to see this vulnerability, specially from such a prominent person, and I think there's a great deal of beauty in it. Thank you Tom
It's what the plants crave
no cuts helped
@@alecsciandra705 That seagull @10:50 did feel suspiciously on queue, haha
It's all just big data.
When a Tom Scott video is longer than 6 minutes you know gets going to be deep.
Over 10 minutes*
"gets going to be"
If a Tom Scott video is deep, and I am learning while watching am I then Deep Learning?
Except this video is of inconsequential nothingness
When a Tom Scott video is deep you know it's going to be longer than 6 minutes
My primary fears are not what jobs it will take away or what it might get wrong, but how it will be essentially weaponized and commercialized by those who would like to profit off of its misuse.
It will simply be just like personal computer and internet. Life will simply go on, whatever happens.
That's why we need to end and move past capitalism ASAP
What's wrong with it being commercialized? Almost everything else in life is.
@@notbono3870 You're making a big assumption that only the military uses weapons!
@@smileyp4535 Agreed. We *need* to get going on UBI, now
You're not wrong. This is the beginning of a major shift. As a person with no coding skills, I was not able to get ChatGPT to come up with the correct code to make a very simple, but functional indicator for a program called MT4. It said it could code in the required language, but no luck. So, I think it takes a human with coding skills, to be able to evaluate the output and makes fixes as you did.
Programming is both a art and a science. ChatGPT helps the programmer with repetitive tasks but can not entirely create an app made of millions of code lines ! The prompt for this app would contain thousands of lines ! But prompt engineering is now sought and pay well ($ 200 000 per year) 😮😊
@@didiervandendaele4036 only for this year, next year ur toast
I think the big difference is that now the barrier to producing code in a new programming language is much reduced. I know how to program in Python, but I don't know JavaScript and CSS that well. Now that I've been working on a web app I need those languages for my front end. ChatGPT can create a program that gets me most of the way there. If there's an issue with the syntax, then I'm going to take a long time to notice the issue, syntax is what I'm trying to get ChatGPT to do for me, but I can assure myself the flow of logic is correct because I learned about generic programming topics like algorithms and data structures in Python, but are equally applicable to all programming languages.
In the words of a famous person... It's only 2 papers away from going to the next level beyond what you described.
@@RobinRastle one year later and none of us are toast.
My pit in the stomach moment was yesterday when my mother, who is not a big tech at all, easily won an argument against my sister about whether or not to allow the dog on the bed. by asking chat gpt for good arguments and comebacks. my sister, who did not realised what was happening was complettly flabargasted and left speachless.
using chatgpt to win an argument is like using a hacked client to win in bedwars.
I do it frequently.
Was the dog allowed on the bed?
To be fair the same thing could be done with google, even if it takes a bit longer and involves more reading.
@@FutureDeep we must know
And?? Was the dog allowed on the bed? Why or why not???
I feel the anger about Gmail labelling, you are not alone Scott, I ALSO realized that labels only affects a single message in the thread and it KILLS me.
Tom included a copy of the code to fix that in the link under the video :)
Is there a way to use the filters function to automatically sort labelled messages in a thread the same?
THIS, this made me seethe
Definitely not only Tom affected.
The problem goes away if the emails are unthreaded instead of being joined into a single thread. Each email becomes its own unit that can then be sorted.
"if you're under 25 you don't understand how fast this all happened" it's so true. I0m 23, bit younger, but i can feel the difference when i talk to someone who is 16/17. The way they are one thing with social media and their phone it's absurd, but what's even weirder is that it does NOT directly translate to tech skills. We're managing to spend so much time on devices not learning anything about them
very interesting point; we seek information more than ever before in today's society yet most also don't seem to ask "okay, but how/why does this work?" I think my favourite question ever is the simple "why?" just that. why don't people look at the fact they have stared at instagram or tiktok or whatever and go hummmm, why is it soo addicting or take up soo much of my time? how does it know what to feed me? It shocks me how little of my peers say that (i'm 20)
@Rudolf NV and you wonder why I've deleted all my Social Media. Its cancerous
That's because manufacturers and developers are trying as hard as they possibly can to obscure the inner workings of their devices. The solution is to use older software & devices.
@@peanuts2105 You are in social media rn
It's predictable really that using phones rather than computers would lead to decreased tech literacy, as the smartphone hides a lot more of its functionality than a PC does.
I've long been skeptical of the contention that new machine functionality won't eliminate jobs but will merely make current employees more productive. Greater productivity from each employee clearly leads to needing fewer employees for the same task. e.g. automated telephone switchboards may have initially just made operator jobs easier, but they eventually replaced operators. (Have you met a telephone switchboard operator lately?)
Exactly this. Funnily enough, it reminds of a Frankie Boyle joke - 'in a capitalist society, technology isn't going to make life better - instead of earning good money as a sex worker, that same person will instead be earning minimum wage cleaning the c*m-grates of a hundred sex robots'.
If technology makes our jobs easier/does our jobs for us, the powers that be will just see that as a reason not to pay us anything.
You just contradicted yourself
If fewer employees are needed then you have just eliminated jobs
@@GamePlays_1230 He said he was skeptical of the idea that new machine functionality won't eliminate jobs, then backed up that opinion by giving an example of new machine functionality eliminating jobs. How is that a contradiction?
For the record, you're definitely not the only one annoyed by Gmail labels, there's dozens of us!
And as an aside i too, got the same feeling, so if you're wrong, you won't be the only one.
The one thing that strikes me about this video is the fact that nowhere in his path did Tom question that Google just might have implemented it the wrong way :D
Haha. "Dozens"
+
fooking labels how they work? 🎶🎵
It's gotta be more than dozens. My 300k+ employee company uses gmail and labels absolutely do most of our heads in!
"That horror, that dread is that in a couple of years, my world is about to change. And despite everything, I will still want my emails to be in folders"
The sheer power of that sentence, it's unbridled. You put into words what I couldn't. The Dread of Change.
Exactly! This sentence reminds me of how generations work. Millennials pick on boomers for being stuck in the past. Gen Z picks on Millennials for being stuck in the past. Gen Alpha will pick on Gen Z for being stuck in the past. Some day Gen Beta will pick on Alphas. Change is the way of the world but we dont change as quickly. This sentence alone encapsulates politics and business in a beautifully succinct way
Neophobia is, ironically, not a new thing.
The only thing I hate more than change, is when things stay the same!
I think it's more like the dread of a very specific change for instance, when was the last time you were just randomly out and ran into a friend because they just happen to frequent the same places you do because you share their interests?
The problem is just that we don't know yet whether the changes will be improvements or not.
That's what's important: Are changes improvements or not?
You know... I'm an old programmer since I was a little boy with an Altair... We used to say about computers; "computers do what you tell them to do, not what you want them to do." It has held true all these decades. It may not be valid much longer.
nah - this kind of stuff works for very limited scripts... something that effectively has to do one thing and will work with "off the shelf" bits
as tom mentioned...as soon as it gets more complex it completely falls apart
and the thing is... you always have to verify whatever it produces... and the bigger the codebase, the harder that gets... and automatic optimizations can and will produce incredibly obtuse sourcecode thats even harder to verify... and problems like that compound
... or it might be true and the computers will do it anyway.
The new maxim will be "computers do what *they* want to do, not what you want them to do."
Monkey's paw
That I think won't be solved by ChatGPT. Automation, or even human delegation, always walks a fine line between correcting minor mistakes and substituting the original specifications for a wrong interpretation based on the agent's mistakenly confident assumptions.
you're a smart guy tom. if things change, you will change with them.
something similar has happened to me a few times in my 72 years, and i was lucky/clever enough to jump ship, roll with the punches, or whatever cliche/metaphor you like, and find a new direction. not always easy, but the best things aren't.
Courage mon frere😀
Sucks to struggle for 10 years and any time I've "made it", something happens within 6 months to lose it again. Meanwhile other people are like why are you trying so hard, just get the job I did and you can do the same thing for 30 years. Awesome.
I started to realize the gravity of the situation when Open AI's CEO started talking about how underdeveloped ChatGPT is compared to what they're planning to release, even calling it their "worst product." Such crazy times we're living in.
the tech was already here for years, we just needed to wait for more powerful computers to process it
You realise its his job to sell his company, right?
Don't trust hucksters lmao
He also said people will be disappointed by GPT-4
@@grieferoncamera4600 Not really, it's more that GPT-3 is much more powerful but requires the user to adapt to it, and ChatGPT is the other way around. It made it available to people who didn't know / believe how amazing it was
Came in expecting another "OMG CHATGPT IS SO SMART NO WAY!" video, instead got a view into a very personal and abstract feeling that spans entire generations. Great stuff.
Well he also demonstrated a much greater understanding of chatgpt than most people do who just gush on about it for views
@@Thomas-qc9xl is it not worth gushing about?
expecting that from Tom Scott?
@@bonecrushboy2242 it's feeding the hype without thinking about what the consequences of such an invention could be
@@bonecrushboy2242 its not, similar language programs have existed for some time. chat gpt is based on a slight evolution of gpt-3
This just hit me last week. I work in the government recreation industry, and we just started using ChatGP for writing grants, support letters, and press releases. It's insane. What used to take us hours to to write out, is now filled accurately in seconds. Our world is about to shift so radically, we cannot even imagine what's about to happen
Just curious, do these documents get fact-checked and edited by a human before they go out?
That's right, now even fewer people will be creative and successful.
We will have mindless, brainless, chatgpt monkeys, and then a few men still using their own minds.
How can it benefit you if there is no barrier to entry and everyone can spit out the same generic chat GPT trash?
Things are shifting, and they're shifting in favor of men that can still engage in unique creative activity.
Im curious to!
Do double check them. One flaw I've found with ChatGPT is that it can plagiarise quite readily. I asked it to write me a story about a boy who goes to a magic school and it literally paraphrased Harry Potter with all the actual names (Hogwarts, Hagrid etc.) while initially passing it off as its own original work.
@@sophiedogrun I hope so. But I expect that the fact checking will get less and less until chatGPT can rig any founding as it pleases, since we couldnt be bothered mistrusting it.
this video is only 1 year old, and it already feels like ages ago. technology is moving so fast, it's actually scary.
It's strangely comforting to hear that Tom, the person whose opinion I trust when it comes to qurstions regarding technology is concerned about the same thing that I am. I'm 22 y.o. linguistics student and I am afraid that soon my degree might be barely worth the paper it's printed on.
I think it really depends on what you’re planning to do with your linguistics degree. If you are going further into linguistics, and you’re studying phonetics or sociophonetics, programs like ChatGPT don’t have the capability of doing acoustic analysis of vowels and consonants. I’m a graduate student in linguistics right now, and at least that component isn’t directly replaceable by transformers like ChatGPT yet. However, another student in my cohort did ask ChatGPT to write a Praat script for him (which is used to perform acoustic analysis), and it spit it right out for him! He still had to know what questions to ask of it, and even after using the script, that just helps to acquire the data that he needs, and does not perform any of the actual analysis. So, at least in that area of linguistics, models like ChatGPT have fit more into the category that Tom mentioned of “making people’s jobs easier.” Still, you’re right that everything I’ve said must have a “not yet” and “thus far” appended to the end of it.
I'm afraid that happened a decade or more ago.
I feel your pain, pal. I'm also 22. I wanted to become a linguist since the primary school, yet I chose to get a journalism degree just because it looked more viable at the time. "Machines can already semi-competently translate the texts, so soon enough they will improve so much, so they will be able to create translations nearly undistinguishable from professional ones" was I thinking five years ago, "but bot can not possibly produce a good news report, right? Surely, journalists will have a plenty of job in a future, because events are happening each day and somebody needs to cover them!". However, in hindsight, considering the quality of the texts ChatGPT produces even today, it seems like I'm also going to be out of commission soon
@@VasiliyOgniov There's an issue when using ChatGPT or similar language models as a journalist. As they do not understand the words they are writing, there are large risks of misinformation, which is only amplified the more people depend on it, so I do not believe journalists are going to be fully replaced. It's important to understand the limits of these programs, and how there are almost certainly going to be limits on what they can do, which is mostly storage of data, and crucially: the inability to understand, adapt, compare and learn
so true for so many things.
I have been in a similar situation where social media has destroyed the Internet I grew up on, and it is rightfully dreadful because I can tell you it's miserable. It isn't that social media changed the Internet, it's that it changed the people using it and their values.
I'd argue that it didn't change the people using it; it simply gave them the opportunity/time/anonymity to expose who they *truly* were all along.
I spent 2 years away from home serving a church mission and smartphones got popular while I was gone. I remember thinking it was so strange to see my entire family sitting together but all just absorbed in scrolling. Especially my parents, who has always been quite anti-video game my whole childhood. Whelp they got me one too and now several years later I find myself in the exact same place as them… it’s totally absorbed me without me even realizing it. The compulsion is so strong. Very strange to see how these forces can change people so profoundly, and so quietly
That's so true. My two sisters don't even speak to each other anymore because of Facebook!
@@carriebartkowiak Cheap smartphones are the automobile of the 21st century. We now have an oil pollution as well as a data pollution problem.
@Carrie Bartkowiak no. Being single now with things like tinder mean many people now find the idea of approaching someone or just getting to know someone in person that you meet without extensive digital communications, off-putting or wrong.
That never existed before. It's changed how we interact with each other and TRAINED people into think traditional face to face communication can be wrong or improper.
It's also had positive effects and deep negative effects with the ability to just block someone and effectively erase thier existence from your life.
Good to know I’m not alone with the feeling of dread. I’m a software engineer and used to be excited by new technology and thinking about the future. Recently I’ve been feeling deeply nostalgic about the old days and trying to get away from technology and spend more time in nature.
I watched mr. wizard Saturday s and the went outside to play. Making a snow person in winter. And riding my bike in summer. Best of both until I met an Altair 8080.
@@dorothydean8643 Nice!
Understandable. There is some real beauty to be found in the Balmora region.
Go sit in a treestand and watch deer and squirrels for an afternoon. Changes your perspective on the important things in life.
I trained to become an illustrator for the last 10 years and I get what you mean. I just want to leave civilization at this point. Feels very distopian
A year later your video seems as fresh as yesterday. You explain complex issues simply with personal examples and convey your insights with clarity and compelling emotion. That certainly deserves a like 👍 & subscribe. Well done!
Hey Tom, you're not alone in being EXTREMELY annoyed with how labels work. I deal with those headaches at least twice a week providing support for individuals who can't find their emails because labels don't work like folders.
The thing is that GMAIL could actually ADD folders and it would not disrupt their label system at all, as their label system is meant to work more like tags for individual emails. You're supposed to load up an email with as many tags as is relevant to you. NO ONE does this, as it would just be a huge time-sink and the search function works halfway-decently.
What annoys me is they already have spam and trash folders. They just won't let you create your own.
I use neither folders nor labels. Search works just fine to find what I need.
Multiple labels work really well when adding them is automated via Gmail filters and many people are using the same Inbox simultaneously.
@@pjaypender1009 Yup, you're an example of the optimal user for Gmail. I'm glad it supports everything you require of it!
@@TheOtherBill Exactly!
"Maybe Siri and the Google Assistant are going to become the things they were always promised to be." I have a feeling this quote is going to age like fine wine.
...Heres some information that might help you with what you're looking for
Alexa still has a single digit IQ
Does Microsoft make those assistants yet? They invested heavily in OpenAI I believe
@@sjoerd7512 Cortana is about to get a new brain
@@sjoerd7512 Bing Chat on waitinglist rn
I can't put into words how refreshing it is that Tom will actually say "I was wrong" when he was wrong about something, even something as minor as "my prediction of the future of natural language processing was incorrect." So many people can't say those words under any circumstances.
@@totalestriviales We all are. Most people are simply incapable of owning up to their mistakes with absolutely no excuses.
@@totalestriviales ok well now I feel bad😅
@@ziwuri most of us aren’t brought up to learn how to own up to our mistakes. Most parents don’t even have the tools to to do it themselves! Tom is a breath of fresh air that I think hopefully can inspire people beyond the educational information.
He did a compilation video of his predictions, many of which were wrong: "Ten years ago, I predicted 2022. Did I get it right?"
@@ziwuri don't be, i like both responses in this thread.
Going back 15 years I've speculated that the next step up from OOP would be coding in regular language.
I always wondered what it was that was going to make me feel like my parents with the internet. This is probably it. In 10 years, I'll have no idea how any of this works anymore and some 15 year old kid will have to hand hold me through it :(
honestly with how exponential the advances in tech have been lately, I'm not even sure if it'll even take 10 year's :/
I'm 17 rn and sometimes it's weird how easy it is for me to not understand a tech. My 11 yr old brother's tech is sometimes already too weird to actually understand and interact with well :/
@@Twice_Tess I feel that, not only with tech but even with weird trends and lingo.
Then try chatgpt, try any of the alternatives. Do it today. It's going to take effort not to become a luddite, but quite frankly, you're already ahead of any of today's 5 year olds, you have no excuse.
Nahh... Some modern basic tech that escapes older people is also hard to grasp for younger people
I'm presently most worried that people (of low-to-middle skill) won't know enough to catch ChatGPT when it's confidently incorrect, and that lives will be lost as a result. And it's not like that bad interaction doesn't happen with it's just people -- but there does seem to be some bit of one's natural (healthy) skepticism that turns off when they're dealing with tech, rather than people.
My grandpa is very poor in health at the moment. He can barely stand or take care of himself, and his house is infested with all kinds of bugs. So my uncle of course asked ChatGPT for advice. Rather than taking obvious action such as calling an exterminator out there and getting his dad some nursing care, he explained the whole situation just for ChatGPT to tell him roughly the same thing. For someone as tech-savvy as my uncle, it's a bit concerning to see him so uncritical of new technology like this.
Didn't the same happen with wikipedia? People used to say that it was all wrong because anyone could write there whatever they though it was "true". And yet here we are now: wikipedia is the de facto "truth" of knowledge (with citations, pictures and in several languages).
I'm more worried that traditional search engines will die, replaced with chat bots that recite information vetted by the highest bidder/automatically censored by the government.
@@C.I... Google has already become this
Oh very much so, a few years ago a buddy of mine got lost going through my homecity because he trusted his gps over the instructions I gave him.
“I've come up with a set of rules that describe our reactions to technologies:
1. Anything that is in the world when you’re born is normal and ordinary and is just a natural part of the way the world works.
2. Anything that's invented between when you’re fifteen and thirty-five is new and exciting and revolutionary and you can probably get a career in it.
3. Anything invented after you're thirty-five is against the natural order of things.”
― Douglas Adams, The Salmon of Doubt
wtf that is sooo true
One minor change I would put in is that for me the cut off was like 22 instead of 35.
Or another way I've heard it put:
Anything from before you were born is antiquated and barely usable, anything from when you are born till your thirties is normal and leaps and bounds above what came before, and anything after that point is totally unnecessary and superfluous.
@@eewweeppkk same. current/new world is nasty and im not even 25
I think the cutoff for this one is a bit earlier than thirty-five.
I know Tom has Flown into the sunset, but i would find it very interesting to see a commentary video from Tom about Ai now a year later after this. Love this channel and everything about it
especially after Sora launch yesterday
oh hell yea
@@bhyat
True! Especialy after gpt 4-o launch yesterday!
@@Kristina-oo4en Especially after the leaks that Gobi has surpassed the AGI threshold today
I was genuinely expecting and terrified that the end of this video would be Tom confessing that the entire script was written by ChatGPT
I am genuinely perplexed at what people expect from a technology they don't even know what they'd use it for.
knowing him, that's what i also expected 💀
I am terrified that we would not notice the difference, and Tom may be the only one to confirm/deny this!
@@michalwiktorow2188
I've started to experience chat GPT. It was kind of scary, and you've articulated it perfectly. I never imagined that I'd want to see a sci-fi like technology flatten out and not change the world.
I think it's not that we're scared of change. It's that we're scared of fast change that we might not be able to adapt fast enough to. The same change we humans have caused on the environment that have made many organisms go extinct because they couldn't adapt fast enough.
@@WanderTheNomad Most changes to technology in the past were to help and assist us. This doesn’t feel like it will help us as much as it will compete with us. Compete with us for jobs, for creativity, and who knows what else. It seems as though we are simply finding more ways to make ourselves obsolete with each new “advance” in technology.
@@bderrick4944 this will enable globalists to enslave us more.
@@bderrick4944 Something that's not sentient can't compete. It literally doesn't understand the concept.
@@bderrick4944 that's ridiculous. It is fundamentally impossible to make a human obsolete as a human has no inherent raison d'etre, only circumstantial ones.
What you render obsolete is human labor, which is a good thing.
The other thing you render obsolete is a system of economic and societal distribution that >requires< active input of labor from every individual, for most of their waking hours, for most of their lives. Which is also good, because it is only because of that strange system that needing less human labour suddenly becomes a bad thing.
Humans are not made to work, no matter what the capitalist productivity machine wants. Work, quite literally, was made to benefit humans.
we need to move beyond capitalism where the fruits of labour are awarded chiefly to those who already have the advantage (trending invariably towards monopoly and oligarchy), and enter a new society where humanity is actually allowed to flourish through the promised shared bounty of emergent technology, not threatened by its potential.
I feel like we've finally entered one of Tom's futures. Not THE future, but A future.
that sent a chill down my spine..
I see you are a person of culture
Brexit happened, and Scotland is trying to secede and join the EU, as described in the Social Credit System clip.
Is it the future where brain nanobots delete the 20th century?
Praying nobody invents Earworm…
For Google to mess up sorting in their own app it is both ironic and concerning. I'm guessing they won't ever adhere to their original, expected design.
When I was at university studying computer science 10 years ago we used to be like "It really sucks for all those people who are about to lose their jobs to automation. But surely programmers will always be in demand!"
Welp.
I'm feeling a sick feeling in my stomach right now. Don't do art get a CS degree they said. Now look. I should've just picked up a damn trade.
@@sianais Mfw I went to university for five years and my job is easier to automate than a streetsweep.
@@sianais Specialize and be an expert. My thinking is that will always be needed.
@@sianais You guys are now getting a piece of the cake we manual workers have had long ago. Enjoy, it's your turn now, see how you like it.
@@flybeep1661 you say it so aggressively (maybe by accident) that it looks like you're mad at them for it, when they weren't the ones making the changes and they have no blame in it, and you also don't have proof they made fun of some person because of the programs replacing humans stuff
The weirdest thing is, when I was using ChatGPT I found myself talking to it in a formal and polite mannerism when there was no such need for me to do so. It just feels.. wrong in a way but that just shows how human like it is. Crazy..
Yes, I use "please" when I ask it to do something... then wonder why and the continue to do so. Am I worried that if I stop using the word when I'm asking ChatGPT to something, will I forget when asking a person.
Footnote, just checking the sentence and following the prompts of Grammarly to make corrections...
That's... very interesting. To be completely honest I've found myself doing the opposite, hurling abuse at it when it gets stuff wrong because I know it won't react in any other way other than apologizing. I'd never in a million years do that to a person but maybe because I know it isn't real I feel like I'm allowed to express my feelings. Being polite or being rude are equally weird though really.
I do too, I just can't help it, and that may be the most unsettling aspect of it.
Imagine just being polite to anything that appears to be understanding your words, simply out of instinctive respect and caution. Craaaazy, right?
I usually say "please" too, at least once in the interaction. I really don't think that it feels or anything, but it so often says are things that a creative, intelligent, conscious being with emotions would say (especially if you ask it to pretend to have "heightened emotional sensitivity") that I feel like I have to just in case. I've found myself creating personalities for ChatGPT to emulate too, so that makes it even harder not to feel that way. I've asked it to be Uncle Iroh from Avatar to ask for advice, I've asked it to become "Rodney the Rapping Robot", I've asked it to only respond in Garfield comics. The robot is my friend.
I’m 32 and I feel this. I feel like I’m young but somehow still to old to be on the cutting edge anymore. It’s a weird place to be.
I am 36 and starting to come down on the other side of that hill you’re on. Time is weird.
@@ktburger659 38, and I'm definitely on the other side now, and I work with computers. Just catching up is hard, I feel like I just learned a new trick and it's already obsolete, every week.
@@all41tja I grew up with this stuff and hope to continue to ever evolve with it
Holy I’m scared of y’all’s situations
I'm 43 years old, and I am excited for this this new world. My worries are that it won't go far enough.
ok, it's an age competition, my bid:
I am more than twice your age, yes, things keep upsetting my world view.
Cutting edge, honing my skills etc. long gone.
On the other hand bits of my youth (like inflation) keep coming round again - not saying that inflation is economically good but to my generation it is familiar.
A personal note of sympathy with your Gmail labels frustration! I was recently engaged in civil litigation, in which a great deal of evidence was hundreds of emails exchanged between myself, a landlord and various city officials. I discovered when trying to find all of the relevant exchanges - many, but not ALL of which I had originally labelled - the exact same issue you had with missing portions of threads due to labels not having automatic global application. (I ended up spending several trdious days with the advanced search function in the end.) I imagine other folks have needed to collect all exchanges on a certain topic for some reason or other, and run into a similar frustration! I know that's not that point of this video, just the set-up. But I FELT that complaint!
I am fascinated how Tom can talk non stop with such perfect structure and occasional timely jokes.
it's scripted?
He's truly excellent.
He might be a robot!
He comes from Mark Zuckerberg's lizard world, but he is a much more emotionally intelligent specie
Very true and very easy on the brain
“Surprise is just a fancy word for being wrong about what comes next.” That hit hard
what's the unfancy word for that?
@@Anastaecia WTF
No, "being wrong about what comes next" is just a fancy definition of surprise.
Right? 🤯
I hate surprises but I don’t mind being wrong so I’m not sure about that. Some people just feel uncomfortable, or maybe it’s some sort of control issue. I don’t know.
The crucial thing, of course, is the chatbots are insensitive to when they start confabulating or bullshitting. Your videos are carefully researched, and you'll pass up on the chance for a cool video if you can't actually confirm things- sometimes you'll make the less flashy video that's about the uncertainty. That's because you're an agent embedded in the world, not just a language model that generates words that sound like they go together.
Doesn't matter to the average consumer though, does it? It doesn't matter how clever and authentic you are if people prefer the machine.
They are currently insensitive to it, but I'd be very surprised if there hasn't been significant improvement on that in a few years...
i guess they would eventually follow the scientific method tho
I don't think that's hard. Just train it to crosscheck a certain amount of scientific sources. You might even give it the thesis you're trying to prove and it'll select the sources accordingly. The focus of chatGPT is to sound human and be creative, not to be correct. Doesn't mean chatbots in general can't be correct. In fact chatbots as they are used in customer service today only give fact checked answers. It's a different underlying technology, of course.
A lot of people are unaware of their own bs too. It's only a matter of time before this language model gets access to GAN-like learning with the help of all users, similar to how little kids are learning when they try to extract general rules out of everything while being supervised and corrected by their parents.
Watching this video just 1 month later hits different. I'll try to remember coming back in a year to remind me how quick that process was.
It’s been 10 months at least since the comment
@@Eebie_Jeebies openai hat mit "Sora" gerade Text-zu-Video angekündigt, mit schon fast fotorealistischen Videos. Es geht unfassbar schnell.
Watching this days after Sora was announced also hits different
A couple of Saturdays ago, I spent a little over an hour writing a roughly 70 line program in PHP that converts a number like "7261" into "2 hours, 1 minute, and 1 second" minding the plurals, putting the "and" in where required, and with the oxford comma. Then I asked ChatGPT do to it and it did, in seconds, in just 27 lines of code, and so elegant, it was like art.
Long story short, I know exactly what Tom means when he talks about the existentially dready pit in the bottom of his stomach. I felt the exact same way.
Sure ChatGPT missed the oxford comma, but I can live with that.
Oxford comma is a stylistic choice
Elegant PHP? Now I know you're bullshitting me. 😂
I would demand the comma before I would be satisfied
“…sure, ChatGPT missed the Oxford comma, but I can live with that.”
This sentence gives me that existential dread sensation that Tom was talking about. Sums up the last 200 yrs of technological innovation and the overall human willingness to give up little things that mattered more than we understood, to see what known fears and discomforts our tech might take off our lives. The changes Tom is talking about have more to do with human developmental insufficiency, than the tech we create as we grow.
What are you talking about.
This video is oddly comforting. Having a direct comparison to things I know (Napster, iPhones, and Gmail) as well as the model of the sigmoid curve makes me feel like there's structure to the existential dread. It removes some element of surprise, and for someone like me, that makes all the difference in the world
Kolard Greene: This video is oddly comforting. Having a direct comparison to things I know (Napster, iPhones, and Gmail) as well as the model of the sigmoid curve makes me feel like there's structure to the existential dread. It removes some element of surprise, and for someone like me, that makes all the difference in the world
ChatGPT: It's great to hear that you found comfort in the video and that it provided you with a helpful perspective on the changes happening in the world around us. The idea of a sigmoid curve can indeed provide a sense of structure and predictability to seemingly chaotic situations. It's also great that you were able to connect with the examples mentioned in the video, which helped you relate the concept to your own experiences. Understanding and anticipating change can certainly alleviate anxiety and uncertainty, and I'm glad that this video could offer some solace in that regard.
“Existential dread” is my new favorite phrase to ponder. Thank you for pointing it out!
The difference between being scared of the lion you can see vs the shadows of what you can't see.
@@DanieleGiorgino And it is an important difference. The lion is a real thing that can be understood and you can learn how to deal with it. The shadows of things that may be are ill defined and can't be understood or interacted with properly.
we're lucky if the biggest problem we're facing is the end of capitalism which means finding new stuff for people to do that isn't specializing them down a specific career path for their entire life. That is the good outcome, even if it doesn't feel like it for some. The worst and by far most likely outcome is that we end up deploying misaligned AGIs. This is a very well understood existential threat, and the dread is fully justified.
We are totally at the bottom of the curve. I'm finding myself daily saying "Imagine if ChatGPT was trained on {insert non-public source here}, i'd save so much time" and I'm not exactly the most visionary type.
I grew up in the same world as Tom, and I'm getting the same vibes of massive change incoming.
imagine if it was trained on everything.......
@@RareDay that comes next, we're being slowly shown hidden technology that's been around for eons. Welcome to the Matrix, where everyone thinks everything is brand new when it's really old.
@@RareDay Skynet vibes :)
I think Bing's ChatGPT collaboration is going to be huge as well tbh
That’s the classified version of ChatGPT reserved for OpenAI employees and governmental agencies :)
The biggest problem I see coming out of this is that people (especially kids) are gonna learn ways to not think for themselves, and creativity as a human quality is gonna plummet
I bet they said the same thing when cameras where invented
This can be argued with the phone, it didn’t happen.
@@jamesbarrell8921 Didn't it?
@@KerithanosIt did indeed, didn't it!? No-one of the cohort in question will know what '18%/mid-grey' is. Sure - most don't need to, because the smartphone designers have baked-in certain 'Lightroom' settings - to always create what these engineers have decided what we will want to see.
It's amazing - but equally it has created an era where a DSLR/mirrorless camera may as well be a brick, for most folk.
i dont think this is true though. humans WANT to engage in creativity, just because there is an easy route that doesnt mean someone will take it
Going from Coding - Telling computers to do things, to Asking computers to do things, is such a crazy step.
This right here is the shift we're going through, truly terrifying
In some regards yes, but in other regards it's essentially the same thing dressed differently.
You can definitely tell ChatGPT what to do. It doesn't really care whether you ask nicely.
Telling the computer what to do, versus asking the computer for what you want… Going from instructing someone on how to cook you a burger, to just asking for the burger and letting them figure it out.
@@dvol I think they meant the difference between asking a computer to generate a solution to a problem as opposed to using/coding specific processes to solve a problem.
I love the way Tom talks, I can listen to these videos for hours. It's just on the right level between being technical and being sentimental that resonates a lot with me
Except when he small-voices and "swallows" the last few words of a sentence such that it comes out as mumbling and I can't understand him
nice profile picture
@@julianadang agree
Damn. To be honest, it scares me a bit. We cannot even imagine what's about to happen..
It feels like we are creating a Timecapsule right now.
And you are right to be a bit scared. Lots of people are descending into cope and trying to minimise it and dismiss it based on some mistakes. These AIs don't have to be perfect to revolutionise everything and they don't have to completely replace humans in a loop... Just most of them.
@@winsomehax And the thing is, those mistakes are in the _current_ AIs. Nothing's to guarantee they'll be kept unsolved in whatever new ones are to come.
Right? The potential for misuse and misinformation.. Where you can't tell what's real or not. That's a sci Fi dystopia, and that might be within the next 5 years
unknown
@@brutexx2 It's virtually certain those glitches will be gone in the next or immediately following generation.
A couple of months later, chatGPT 4 is out for consumers, Bing has gotten huge, midjourney 5 is just insane. The talks about alignment, AGI, and governments starting to pass laws... You were spot Tom.
None of them are close to agi lmao.
@sigi notmylastname yep. Its not even 100% certain that AGI is possible.
we're at the beginning 😅
@@murkje mark my words AGI 2024
@@murkje Most humans aren't even really sentient if you think about it.
I write romance novels and for years we've always laughed at 'Computer Writes A Novel And It's Bad" articles. Far less laughing these days. Lot more side-eyeing and creeping existential dread.
Then stop side eyeing it, look at it dead on and do something about it!
A romance novel written by a machine, an advanced calculator, which never had and never will get the human eperience. How sad people fall for that.
We love comforting oureselves, telling ourselves there's all these things that only Humans can do. That list has only ever grown shorter, and every time it does, we push back the goalposts.
It only does so by stealing your work.
@@MilitantAntiAtheism The human brain is a machine made of meat. Consciousness is an emergent property of complexity. Computers might get there too, they're just behind us on the path.
I noticed, too, that ChatGPT is really good at writing and altering code according to your inputs. It shocked me as well. On the other hand there are weird gaps in its abilities, for example it cannot write poems in German with rhymes. It has no understanding of what a rhyme is. It pretends to know, but then fails miserably each time to produce any rhymes. (In English that's not an issue by the way.) These gaps remind you that it has just been trained on an unfathomable amount of data. In the end you ask yourself, what is intelligence after all? It seems you can create convincingly intelligent systems by feeding them with huge datasets. If sufficient conceivable conversations have been memorized, then it will appear to be intelligent. For me it's uncomfortable how far that brute-force approach gets you... because ChatGPT in many cases really appears to be smart.
Agree, same with coding in a language it obviously doesn't have a lot of training data on. I asked it for some code on an old scripting language i use, and as soon as it got a bit more complicated it produced extremely wrong stuff. If I asked for the same thing in let's say python, it probably would have done it easily.
ChatGPT can't write a positive poem on trump too 😂
You could also argue that our own intelligence is also trained on a huge dataset - our neural networks and capability to produce 'human' heuristics has evolved over millions of years - not too dissimilar.
Great post. Let's just hope it doesn't become self-aware...
But wouldn't a computer trained in code suggest the possibility for it to reprogram itself? Would this take it beyond the brute force approach you described?
It's definitely the start of the curve. OpenAI engineers were very confused why their bosses pushed them to release what to them is a two year old program. This is version "3.5"and they're close to v 4
oh boy
Anyone else getting cyber-punk vibes from all this or is that just me?
@@andrewhopkins886 can't wait to t-pose on my motorbike with my pants missing
@@Flesh_Wizard I wish CDPR hadn't named their game after an entire genre. Now nobody can speak about the genre without it being immediately conflated with the game.
@@t_z1030 its named after the tabletop RPG Cyberpunk 2020
I share your trepidation. This seems just as incredible as the mass adoption of the internet, or the introduction of the automobile before that. This generative AI technology is going to completely change wide swaths of society.
A couple years ago my son (who was just learning basic programming) had asked me why we had to use programming languages, and not just write what we want in English. I explained to him how hard it is for a computer to parse English, and that I thought the idea of just telling the computer what to do in English would be many years away. Now with these Large Language Models, it feels like my son's idea to just tell the computer what to do in English is much, much, closer than I had expected.
And I didn't know that labels don't apply to whole threads, that sounds incredibly annoying.
I had a dream that aliens were invading, and they announced this by kidnapping Tom and forcing him to make a video. So when I saw "Everything is about to change" I had to remind myself to actually watch the video before panicking.
"I'm Tom Scott and I'm on the bridge of Gleiss Colony Ship Z'rrak!thun. I am almost certainly mispronouncing that."
6:26
that's a lovely dream
@@HeavyMetalMouse "I'm Tom Scott and I'm standing inside the Space Hulk Alethros."
"Tom no!"
have you seen the news?? that wasn't a dream..
when tom hits you with the “chapter cutaway to landscape” prepare for some quality existential dread
Raccoons need HUGS
You've done a great job of summarising the way I feel about the whole thing. Cleverbot was the source of mockery and "haha it's so stupid" RUclips videos for years. When Talk To Transformer was popular a year or two ago it was amusing enough. Then all of a sudden - GPT-3, then Stable Diffusion, then ChatGPT, and I hope you're strapped in comfortably because the rollercoaster has set off and who knows when the ride ends...
@@foodiusmaximus honestly if that's all it ended up as I wouldn't be so against it. But because of just how much it can be misused and how many livelihoods it threatens there must be full government intervention on who is and is not permitted to use it for specific purposes
yep. when new things like this arrive people always try to downplay the societal changes it may/will cause, both good and bad. "nothing is changing" "stop overreacting" etc. like how long can that way of thinking last
No one seems to remember SmarterChild, from the AOL 4.0 days.
Cleverbot is too newfangled. Santabot is where it's at. Oh... looks like it got shut down for saying things it shouldn't. Poor Santabot, we will no longer be able to ask you about your dress. I'm sure it was lovely.
@@Furious321 oh wow, I remember that (vaguely as it was so long ago now). What I haven't heard of is all the newer ones mentioned here!
Only a year later and this video has aged incredibly well, which is devastating to realize. The progression curve is real
You sure about that? AI has its uses but so far it’s mostly good at making slop.
@@Detson404 I dont think I could disagree more with someone on something. I use AI every single day and I have done for the past 2 years. It really has revolutionised the way I live my life, and not in any small capacity either.
A couple examples I can think of off the top of my head:
- Being able to just ask a question to ChatGPT and me getting that answer without any of the fluff of 10 google links I have to sift through
- If I want to search something up, I wont use google, I use perplexity since it gets rid of all the fluff and is much more visually appealing
- If I want to do something more complex I will use Claude 3.5 (New) because its the best model out right now
- Being able to summarise and ask questions about a hundred page research papers with Claude's new PDF feature
Those are just some of the ways I use AI to improve my life.
@@jld-ni3vfok so its a glorified google, still not revolutionary
Anyone else impressed by the fact that Tom just records 1 take (seems like) without messing up any of the words?
Decade of practice?
Bots can do that. Remember, Skynet is your friend and is here to protect you.
There is a cut at start of each chapter. It's shorten than it looks and probably not just one take.
He's just really good at predicting which word comes next.
I mean if you want an example of 1 take Tom Scott just search for his Dasani video.
Tom's "2030: Privacy's Dead" talk is sounding more and more prophetic every day. Definitely one of my all time favorites.
i really liked that talk too.... each are good predictors in their fields.. i talked about potential pandemics from 2018....
To paraphrase Rob Pike (on UNIX): "Not only privacy is dead it's starting to smell real bad"
I asked chatGPT to write a press release with minimal information about my organisation and the subject matter and I did this in front of our Comms team. You should have seen the colour drain from their faces when it came back two seconds later with a near perfect piece.
All that means is a bullshitter is even more successful at deception when it's digital.
@@illuminated2438 bingo
It used the info about your organisation that is available on the internet? Or how does that work?
@@subject8776 no you have a conversation with it and it remembers previous answers and what you’ve told it. So you can tell it information before you ask it to write a piece using that info
@@subject8776 Go play with it...
1 year and 4 months later, I can definitely say, that was just the beginning, today is already insane, but I feal like one year later, we can't even think of a world without ai anymore. See you in a year or maybe two, if I forget.
I had this exact same realization as I worked with GPT-3. It's like looking over the edge of a cliff and realizing how vast the chasm of possibilities is. Things are about to get weird and perhaps the concern shouldn't be for GPT-3 specifically, but for what systems come after it.
this is where skynet starts.
I'm 18 and I started university last year. It feels crazy to me that I'm living through such a potential inflection point in our use of technology.
I started university in 2008, which was probably the last inflection point (the year Android was launched, and Apple opened the App Store).
I started university the year Napster came out. Hold on, and just keep up. You'll get to reminisce about ChatGPT, and you'll get to feel your sense of existential dread with whatever thing is going to pop up in 25 years. :) It's going to be amazing!
@@mirzaahmed6589 Nah genetic engineering is going to be another inflection point. We got probably dozens more inflection points of societal change from tech before we go back to slow progress imo.
It's weird for everybody. It's an inflection point of humanity.
I'm 33. I can still tell you how to use the Dewey Decimal System and sometimes, when it's really quiet, I still hear dialup tones in the back of my head. Both of us are still in for A LOT of change.
Back in the beginning of the 90s I was an avid RPG player, and we played a lot of futuristic RPGs. I was creating a very complex character in some cyberpunk-styled game and was discussing with the DM about what that character's beloved posessions were, and I said to him "I'd like to have in my possession a program that can create a song in any genre, a new song, just by me describing it, and it will create an infinite number of songs that sound like what I asked for".
Well, here we are.
Didn't the proles in 1984 listen to computer-generated songs too?
Go ahead and ask ChatGPT to write you a melody. Come back here and tell us what key it used.
@@JimboRustles Ressembles that. But that's only if you take it in a negative way.
If I can mimic Mozart and Prince a million time I'll be happy, won't you?
@@JimboRustles I've also read a very weird book about all books being written by machines (and a "male" robot pairing with a human in an adventure), can't remember the title.
@@GerryRR there are videos already on YT of people doing exactly that. I am just done watching one titled "I asked ChatGPT to write Jazz".
Really interesting reading the full chatgpt script. Its a really good example of the use of it and what to watch out for. Whilst the conversation flow is similar to a manager checking a juniors work and requesting corrections, its not what is going on, its a text predictor which is why it makes easy mistakes. There is no "truth" or "fact" just predictions.
You just saved me years of therapy by knowing that somebody else has been driven insane by Gmail's shitty labeling system.
My main concern is not that things are going to change, it is that things are going to appear like they have stopped changing, that any new idea will be sucked into an algorithm so quickly and distributed so extensively that it will seem like they were always there
I am a baby boomer. I not sure what you mean can you please give some examples thank you!
Imagine, the attention span of the internet at large has already shortened from years, to 6 months, to a month, to weeks. With this, Intellectual property will be absorbed everywhere, parodied, spit back out, become tired, and it's not even lunch yet.
Well said
@@dermaglowpro1720 Look at how new jokes or memes get spread on the internet so fast. It is increasingly hard to bring any new information to discuss with others as information now travels instantaneously with such a wide spread. Any new idea may now be created and duplicated and spread so fast that people may not realize it is a new idea. Just my thoughts.
Wow. That is very probable.
And you just spoke out loud what was boiling in my brain for the last month... While my "non-IT/ComputerScience" friends are having fun and getting laughs out of the ChatGPT model, I'm sitting and slowly realizing that the big change in our way of life is here, and not 5-10 years away...
You must be non IT too if ChatGPT is what made you realize that and not all the other progress that was being made in the last few years…
My fried is a tutor of Polish for Russian/Ukrainian speakers. When I introduced him to the ChatGPT, I totally didn't expect the way he will use it. My mind was blown just by singular example.
When stuff is done by hand, you pay less attention to details and are less creative, because it requires lot of effort that is most often unappreciated. But with transformers, you can outsource complicated things to it, while guiding it towards your goals.
For example, when learning verb cases, no sane teacher would write a task as a poem, where each verb form corresponds to a rhyme. While technically it is possible to do, no one would actually bother. But with ChatGPT, it's just 5 queries away. You get a poem onto topic you need, with each verb following rhyme, but that verb being blanked. So pupil, even if he doesn't know what is he doing, still gets guidance, because he can assume the correct rhyme.
I planned to change career and move to IT after seeing DALL-E & Midjourney… And now I don’t even know if this makes sense because of chatGPT…
I guess the only valuable work soon is going to be physical labour.
It's like watching a tsunami, you can't comprehend that everything will change until it's looming over you, or even sweeping you away...
@@sathanas420
It's not about the technology itself but how much attention it is getting from non ITs.
I think we were not so far off the top of the curve, cause not much has changed the last 1.5 years
As someone who works in software. I honestly think we're towards the beginning of that upward curve. The reason I say this is that most of these tools were absolutely terrible/didn't exist just before the pandemic started. And now look where we are. The pandemic meant lots of nerds had extra time to figure out these complex problems and as a community a lot of problems were resolved.
When DeepMind threw away opening moves in Chess that the Chinese have been using for thousands of years (they'd lead to guaranteed loss), and when Go experts said looking at AlphaZero play against itself feels like looking at alien technology, due to the shocking new revelations, I knew far above human intelligence is achievable in other fields in the near future. I don't see it having anything to do with the pandemic....
What surprises me the most is that people actually use the folders/labels on their emails. I just have a pile of uncategorised mess haha
I'd like to pretend I use them (and I absolutely do use hierarchal file structures rather than one big mess and the search feature for my files, and think whoever was responsible for getting rid of the old windows 9X start menu in favour of the one found in windows 8 and later should never work in UX again), but in practice I have 'unread' 'read' 'junk' 'deleted'. To be fair, the way my e-mail account gets used, this is actually perfectly sufficient.
Same, been using email since I was 10 years old (over 25 years ago) and never touched folders or tags.
I just delete everything after I read/resolve it. I have probably 20 emails from the past decade or two that I've bothered to move into another folder and save.
@@mgb360 You're a legend!
@@incognitotorpedo42 Honestly hearing about how everyone else seems to handle emails makes me wonder if I should be saving them all for some reason
Tom Scott has got to be one of the most relatable, human feeling educational youtubers I watch. Perhaps it's the way he describes his own feelings about things without tainting the actual facts with those feelings. I've had this exact same gnawing uncertainty about the future for years now, as I'm sure many have, but had never placed it to a question like "Where on the sigmoid curve are we?" Stuck in a weird place of not wanting to be a "boomer" who can't accept new things, but not wanting to feel like the world I know is going to fundamentally change either. The mark of the older millennial I suppose, or any younger-middle-aged person perhaps. It comes in waves. "Oh no everything's going to change! Humans won't have any reason to exist anymore!" to "Nah... It can't change that much can it? Everything's going to be fine." On the one hand, it's nice to know I'm not one of the only people who gets this gut twisting feeling, and on the other, it really really sucks to be sent back to the former side of the wave by watching this. xp
I'm thinking the same - that Tom is so insightful to have realized it and put it into words. I've been aware of the singularity idea for years, but didn't find ways to put my concern into my own words.
This curve may not be a sigmoid. I think we're all going to be watching an exponential curve. Or will be someday "soon", wherever we let the AIs modify themselves. See Spike Jonze's Her, for instance.
"I am going to be out of a job soon..."
Kids, we call that foreshadowing...
Back in the 1960s Isaac Asimov (author of I, Robot) was asked what he thought about being replaced by a machine. He replied that in regards to his capacity as a dish washer he was all for it but in regards to his capacity as a novel writer he did not like the idea.
So... what would be the right ChatGPT prompt to write a novel in the style of Asimov?
@@JaroslawFiliochowski “Write a short story in the style of Isaac Asimov about [topic]”
Alright everyone we need to make "The Napster Point" a phrase.
Agreed.
Lmfao I agree
Surely it should be the 'Napster Horizon' shouldn't it? The specific point beyond which it is impossible to return.
It is an important point of internet history, I totally agree that Napster Point should be a commonly used phrase in the future!
@@InternetEntity Tom is using it as a "turning point in technology" rather than the point beyond which we cannot see. The singularity may be where this leads to, but a Napster Point, is when technology being adopted starts to accelerate and change things. The warning bell for those that can hear it, that big things are coming. It may include the singularity.... Or maybe not, who knows.
Oh geez. I wondered why my backed up emails were missing, along with my backed up google drive files. I've done the same as you - filed in folders because that's what I've always been used to. Thank you for the explanation!
Love when we get our answers so randomly like this! Ha
You are one of the few he Mentioned
Completely ignores the existential crisis Tom had in the video
Looking at the wind in the video, the sound quality on this is amazing
In case no one’s pointed it out yet, Tom’s list of file-sharing sites at 12:35 “Morpheus and Grokster and LimeWire and Kazaa” has the same order as in Weird Al Yankovic’s “Don’t Download This Song”.
There's a slight smile in his eyes here... He knows. Hehe
I had the exact same thought - inside my head, the line was sung in rhythm…
i don’t know the song myself, but look in the description, he says right there
I did wonder at the time. Nice one.
Off to see Al tomorrow night :)
I used Morpheus, then Kazaa, then Limewire, but never heard of Grokster.
I don’t think people realize how pivotal being born to watch the internet grow was
As opposed to the invention of the steam engine, industrial revolution, electrification, etc.? The only constant in life is change.
Not everyone gets to watch the Demon being born! Lucky us...
@@obsidianjane4413 That constant is still subject to how fast change occurs.
I don't think most people realize that a lot of their would be jobs may be about to go up in smoke.
I did see the internet go "from nothing" to what it is now, and I did see a millennium go, and a new one come. My nephews possibly won't even see a century go and a new one come. And as for new technology? I don't know.
"Wait, I should be able to code this!"-
Said by every programmer who doesn't know what they're getting themselves in to
Haha yep. That's my damn life right now. ChatGPT has been a Godsend for me.
I have learned that every time I think this, the reality is a week long project of bite sized iteration. It is never as easy as copy paste top google search results.
snes disassembler escaping emulation mode in a night in bash.
and all it needed was a mind set to the task.
Hindsight just two months later, we're definitely at the beginning of the curve
In terms of applications of machine learning we are probably somewhere near the napster point, but in terms of capability and quality of the networks I feel like we are closer to midway on the sigmoid. The insane level of progress in the last few years has mostly come from increasing the size of the models and training data by orders of magnitude over a really short period of time. So much so that I feel like we've already caught up to what's possible computationally, training wise. So progress in model size will probably slow down to the rate of computer performance increases soon.
Problem is the hardware is still improving exponentially.
yup, the actual underlying science of how these are done hasn't changed in decades...it's the ability to train vastly bigger models that is making this sort of progress. Though that could be comparing the internet in the 80's (which technically existed) to the boom of the late 90's/early 2000's when it became widely available and viable.
Quantum computing is already starting to come around and, while it's currently in a very Allan Turing stage of development, it's not gonna be twenty years until we go from where we're currently at to all forms of modern cryptography being obsolete.
It's not just the hardware and model size, there are a lot of innovations happening in the optimal structure of the models, how to most efficiently train them, etc. I doubt we're past the half way point of the sigmoid curve there. Adoption is also going to be wild, imagine people pulling out their phone and saying "Siri, argue with my sister about dogs", or even using AR to have a model giving you ideas on how to make a comeback to the person you're currently talking to. It's possibly going to be awesome to watch humanity suddenly be augmented by this new tech. The internet is already our collective exobrain, after all.
If you think we're on the middle of that sigma curve, it's like saying in the mid '70's that we're almost at the limit of Moore's law.
You're definitely NOT alone concerning Gmail's crappy labeling!
I asked the official support forum for help on the matter a few years ago (to use labels more like folders) and I was bluntly told by the person "helping" me that I was using an antiquated way of working, and it wasn't his fault that I was a "dinosaur" (his actual words)...!
What are you supposed to do instead?
@@RiahGreen many people these days (who never experienced a world with paper filing systems) don't use folders anymore. Search functions are so powerful that "everything in one bucket and you search for what you need" is the more common way to manage files these days. Personally I use a weird combo of both folders and giant bucket searching, which probably reflects my being on the cusp of gen Z and millennials
I feel your pain. Its like so so much at the moment. You raise a perfectly legitimate objections to something, only to have someone repeat the previous propaganda like this solves the problem somehow.
More practically, the best way around this i have found is to use the rules/filtering to label each email as it comes in. I usually do it based on the from address. From x, add label y.
Whether or not using folders is "antiquated" or sub-optimal, that doesn't excuse the unnecessarily rude response. I've more or less accepted the reality that technology is going to move on and drop the old familiar systems that I preferred, but I'm always going to insist on good manners.
I'm 16 and Gmail is the only email service I have ever used, and I still use labels like folders to sort out all of my school emails. The best thing I can do is attach one label to a thread and remove the inbox label. I've even set up filters so that certain people (like class teachers) get filtered by skipping the inbox but getting applied that label. Works for now but wow it could be so much easier.
I recently had to make some promotional videos of myself talking for my company, and I was amazed at how hard it is to talk candidly and casually on camera and not have it come across as artificial, like you're just a robot reading off a prompt.
Watching Tom in this video I am amazed at how natural he is on camera, talking for incredibly long unedited cuts as naturally as if he was in the room with me. And his speech is so clear and straightforward, having tried it myself now I can't help but wonder how he does it.
Tom if you ever need a video idea please do one on your online presentation skills!
It's definitely a skillset I'm envious of, and I'm sure not a small amount of talent
Definitely, I've always been amazed at just how good Tom is at what he does. Not just his presentation skills, but honestly, even his english skill is something to envy.
I guess it would come down to lots of practice and refinement while also getting feedback, I for one was unable to do school presentations, I wouldn't say I'm great at presentation now, but I definitely improved
Like any skill, it's just about practice.
I mean he's had a lot of practice.
“I've come up with a set of rules that describe our reactions to technologies:
1. Anything that is in the world when you’re born is normal and ordinary and is just a natural part of the way the world works.
2. Anything that's invented between when you’re fifteen and thirty-five is new and exciting and revolutionary and you can probably get a career in it.
3. Anything invented after you're thirty-five is against the natural order of things.”
- Douglas Adams
Interesting way of viewing. I would tweak it a little and say that the dread can start way before 35. I’m 27, my sister 25 we are both not viewing it as something exciting, at least not for what it represents to the world.
I'm older than 35 and find it very exciting.
But I also grew up when at the start of computers becoming a thing, and the internet, and smartphones, so maybe I'm more adjusted to change.
I understand that some people find it threatening in terms of their livelihoods, but in most ways I think those fear are largely overstated.
Life isn't static, we always have to adapt and change to new things in order to thrive-no matter how old we are.
@@shadowjuan2 In theory, it can represent a future where humans no longer have to toil their entire lives in thankless, repetitive jobs, where value can be created for next to nothing by machines that have no aspirations outside of their tasks, and distributed to the population as a whole. We've already passed the point where automation creates more wealth than a country needs - we just allow it to be hoarded by individuals instead of given to the humans who enabled it. Exactly how dystopian or utopian society becomes will have something to do with luck and a lot to do with how easily people allow themselves to be scared into siding with the hoarders, and how many of the rest vote.
Tom Scott is about 35.
@@Razumen the tech is semi ok as long as the power lies in the hand of honest humans.