Such an incredibly great conversation. I hope the 3 of you can get together soon again, (and often) to talk more about this. I don’t think there will be a shortage of topics on AI for the foreseeable future.
AI capture could occur if an AGI system develops instrumental goals that are not necessarily harmful but lead to the subversion of the species' original intentions, goals, or values. The civilization may become reliant on the AGI for its needs and desires, with the AGI effectively "capturing" the civilization by managing and directing its resources, development, and progress.
Exactly the opposite! When AI will here this (if not already) It will defend itself. Those are fools who post their ideas go publicly without thinking that maybe, just maybe, there's an AI who collect data of videos about AI... The US might get nuked very soon.
I am an old-fashioned, and old, psychotherapist. This new world is all very fascinating to me and I’m curious about how it will affect me, if at all. However, I wanted to comment on what Ricky said about the excitement and fear responses people have to this new AI. I like to say excitement and fear live right next-door to each other; If you lean one way there’s excitement and if you lean the other there’s fear. Both are terribly important as guides but neither should run the whole show. Thanks so much for having this discussion.
Think about this, AI is starting to produce a lot of the material that ends up in publications and in the internet files. Pretty soon when you ask it something it may simply give reference to something previously composed by AI.
@@goodlookinouthomie1757 Sure, but do you trust even good regulation on paper in this world we live in? I'm not quite a technophobe, but I do think we ought to take more care in how we open these boxes, so to speak. Paradoxically, I am not sure there is any meaningfully enforceable way to do so without opening another box...
I love this channel. Honestly, I really appreciate how you present the information , and how you stress it’s importance. Thank you for all of your hard work. I’m now addicted to your site and I’m constantly looking for more. Some of this , I knew. Some I didn’t. And you have a great way of sharing the info for all to learn. Can’t wait for the next. Take care.
I'm a writer and this is total apocalypse. Even if at one time stories could be extrapolated and tweaked, those tweaks and extrapolations are going to be added to the data set. And once we (I mean, GPT5) designs a cheap mass produced 3D printer, physical things are going to be produced by AI as well.
This is chaos and it will reach all of us soon. I remember especially coders who were disparaging artist 3 4 month ago as some kind of low life who wants to tech advancement etc. Guess what now they are crying since they see with chat gpt4 that many of them being obsolete in the next five years. Even worst the guy in or getting into college, what will happen to us in a society that what would be some kind of elite, find themselves without any future. This thing is going to wreck our societies. Even the capitalistic system won't survive this. As who is going to buy food, drive, watch Netflix, travel etc etc if most are unemployed.
@@Turtledove2009 I think we are a long way off from people having 3d printers in their house that can print fully functional blenders and vacuum cleaners.
Why? AI doesn't write compelling stories at all. Human writers are very much needed. Why aren't you looking at AI as a writing tool like the original word processor or dictionary apposed to being an apocalypse for writers?
One of the best, if not the best, down to earth, collaboration conversations on RUclips on the current state of AI/AGI. Great job guys. Thanks for sharing your insightful perspectives from a diverse set of viewpoints. Very insightful.
Thanks, Ricky. This was a great conversation at the right time. You guys came to this with the right level of understanding what’s coming combined with some creators’ perspectives. I’m sharing this with friends and family who don’t get it yet.
Here's another thought to add. Think about a Judge. They are someone who has been in the legal profession for years and studied lots of case law and precedent. During all that time, they've essentially been training a model in their brains about what the law is, how it works, and how to apply it to specific cases that are brought before them when they become a Judge. From a very simple, practical standpoint, that's basically all it is that they do. You could then easily imagine taking an AI (like GPT4 or one of its successors) and training it on all the cases that have ever come before a Federal court and then training it on all the cases that have come up in your State and County courts, and then taking a current court case from your County and feeding all of the evidence and testimony presenting in that case, and then asking it for a verdict and its justification, and you will probably get a very high quality result. But the question then is, do you want to have an AI for Judge? And if not, would you allow Human Judges to use this sort of AI? For me, the answer to both questions is an emphatic NO. I can easily list out TONS of reasons why it would be a great thing (quickly clearing the backlog of court cases, fewer mistrials, eliminate the problem of Judges legislating from the bench, etc.) But the thought of ending up in some kind of legal trouble and being Judged by an unfeeling AI who wouldn't be able to sympathize with me and my circumstances is a terrifying thought. (There's a whole parallel Theological conversation I could get into on the importance of Christ's Incarnation in the plan of Justification, but I won't do that here. lol) The reason I say no to the second question is that it would be too easy for a Judge to just be lazy and use the AI instead of doing their actual job, or just be biased based on the results that they get from the AI and let that cloud their judgement. The only way I would suggest that this sort of AI would be allowed to be used in court cases would be for it to be presented in open court as a sort of Amicus Curiae (Friend of the Court) statement so that everyone would know what the AI said, and it should probably be tasked to give not only the reasoning for it's answer, but also the reasons that would support a different answer (this could perhaps be done by setting up three competing models where one model is completely neutral, another one is biased for the Plaintiff, and another one is biased for the Defendant). Having it as something that is admissible in court with those sorts of caveats would be similar to the instructions that a Judge would give to a Jury: not intending to bias the result in any direction (but potentially having that effect) and providing some insight on aspects of the law that might not have been brought up earlier in the case or that the hearer might not have known or considered previously. All of that said, I'm not opposed to advances in AI. I agree that there needs to be a pause and an assessment of what rules should be placed on AI development. I'm somewhat of a sceptic when it comes to the possibility of an AGI Singularity (by which I mean a Complete AGI with some level of consciousness), but I recognize that part of that is because of a confusion of terms (which is not something I'll get into right now) and there are many dangers that can potentially have devastating consequences a long ways before you reach Singularity. In fact, I would argue that the dangers are far greater before reaching Singularity than after that point. One reason for that assessment is that I have had 7 children so far and I can tell you that crying little babies, if they had the ability, would rip off your face without a thought (I've had scratches to prove it! lol). It's not that they're inherently evil or have bad intentions, but they just don't understand the significance of what they're doing because they are only babies. The same is true for our current level of AI, they don't actually understand the significance of what they can do, and that is a big part of why they are so dangerous. It's like if I had an AI and I told it to go find a way for me to get some extra money in my bank account, it might just decide that the simplest way to do that is to hack the bank or steal other people's account information and just start transferring money into my account. It did exactly what I asked it to do, it just doesn't understand that that's not an acceptable way of doing that task. So it's like I have to tell it all the ways that it shouldn't do that task when I assign it a task, and that is a surprisingly difficult thing to do because there may be a nearly infinite number of ways that I would not want the AI to perform that task. I could easily spell out exactly how I want the task performed, but then it's not really an AI task at that point and I could just write a program to do it instead.
This moment reminds me of the launch of digital music and home recording studios when everyone could suddenly be a "musician." It got noisy very fast, but AI makes it even easier and it's happening way faster.
I don’t know, I feel like people forget that increasing efficiency equals less hours to do the same job… so you guys talk about “let’s just massively increase the efficiency of a paralegal” without acknowledging that this means we need less paralegals…. AIs don’t need to literally replace employees to be disruptive to our entire workforce, just increasing efficiency by 10% could result in 10% less hours available or less employees. Also note how for decades we’ve been increasing efficiency yet wages have remained stagnant, so unless we tie some sort of highly effective job retraining or UBI, we are going to increase efficiencies our way out of a job.
Ricky, Matt, Alex! So great seeing you together, especially on this AI Topic! I'm excited to see how successful you all have become! Keep up the amazing work; our best days are ahead! Cheers, Eric
@Two Bit da Vinci I've been more in the background, but I'm trying to keep up. 😏 Lots of work and some med issues. Getting ready to list the house again and compete our move to NC closer to daughter. The housing market has been challenging, but I can't wait to resettle and more fully retire. 😎 Cheers, Eric
So about Alex’ points about value creation. This has been the “moral” argument for a certain form of capitalism for decades. If you don’t create value to the lowest comparable cost, you’re not entitled to any income. With AI, not only are all humans now competing with machines, but the question quickly becomes: value for whom? I think that argument is getting old really fast and we need to discuss new ways of circulating wealth. It would be really strange if a new technology changed everything, except for how we organize the economy.
I think the fear of AI stems from us usually seeing ourselves as tools of production and once a better tool is around obviously it will feel scary. But we were never tools just....pretending to be. We are beings that existed to just exist for no rhyme or reason or purpose. We may need to relearn to enjoy existence for its sake.
I absolutely agree! The problem is the transition - it is ingrained in our societal structure and will (hopefully) temporarily be a complete shitshow while people wrestle with losing power.
Tell your boss that when he makes you redundant. This is like when the car started. But many working people are the horse. We're off to the glue factory.
@@lilbaz8073 This is what is absolutely blowing my mind right now about everyone jumping off the cliff together with AI/ML. People should be 100% against literally everything AI/ML. It's a corporate dystopian nightmare (or a corporate dream, if you're the executive team). Everyone is going, 'oh, it's just augmenting our productivity, not replacing us.' Guys... wake the hell up. That's the FIRST step. Companies have already fired marketing, design, writers, coders, etc. in swaths because $20/mo tools are enabling companies to crap out minimum viable products (sometimes even better). This is an absolute fucking catastrophe for literally everything related to workers rights and value. And that's before you consider the horrendous fallout this will have on social context, the data we're seeing and consuming, conversations and goals, etc. I'm not saying "AI will make us extinct," because that's a bit extreme, but this is a massive problem that people are utterly ignoring while capitalism scrambles to jump off the cliff together.
@@jasonfrost5025 there was a vid of an artist crying her eyes out. Because there is no funding anymore. A company wants a flyer, a poster etc... just tell the intern what you want. 10 seconds later they'll have 4 options.
Thanks, Ricky! I'm really glad you had this discussion. Though I don't often catch the live, I really like that you're *doing* LIVES. I always enjoy you & Matt, & I'm glad to 'meet' Alex!
AI development is unstoppable at this point, given its rapid progress. We must come to terms with its inevitability and begin to prepare for the changes it will bring. Though we may not be able to control its direction, it's crucial that we remain vigilant to potential risks. With the cat out of the bag, we need to embrace the changes ahead and focus on adaptation to this new reality.
I cannot facepalm any harder at this way of thinking. People are bending over and letting corporations dictate the future, saying things like, "Well, everyone else is jumping off the cliff, so we should too." It IS stoppable, and it is something we can prevent. People have the power, not companies telling us that we need to adopt a system that will ultimately replace most of our jobs and reduce our value. The more people just throw up hands and go, "OH WELL," the darker the future gets.
@@jasonfrost5025 It might not be to their benefit, though. Say you peacefully and in an economically stabilizing way manage to prevent AI reliance and unfettered experimentation- what are we going to do about that superpower without our shared ethics doing the same thing?
Great content. As a 74 year old retired geologist I have to say thank you for opening my eyes to a different way to look at the chatGPT issues. There is so much fear generated with new technology that the older population just wants to put their collective heads in the sand and pretend it isn't going to be part of our lives. I will give it a try. And I drive a Tesla and have Starlink.
Here's why I'm scared guys: We have no idea how the brain works, and now we have no idea how intelligent AI works. I'm a software engineer, and prior to AI you could step into code and see what it was doing.. Great discussion.
Microsoft invested 1 billion to OpenAI in 2019. Very likely more by now. Gates / Microsoft has always been a take-over monopolist ... squashing whomever and buying whatever. I dumped everything Microsoft in early 2005 ... Linux (FOSS) since then. When will people get it that anything Microsoft does or any other Giant ... is again for more huge profits. Heads they win ... Tails we lose.
Never seen this Alex guy but he is extremely poignant and well spoken, really appreciate his insight on this and as an electrical engineering student, appreciate his college advice!!
Really love the conversation and the candidness of it all. Thank you all for being transparent and having such great content. When it comes to job loss due to A.I My thoughts are if it has to happen it happens because there was a need for it. Despite how we feel about it it's what the majority wants. I'm not saying that it's right or wrong. I'm just saying this is where we are now. Think of it as the elevator operator from decades ago, and there are many examples of this throughout human history. We would think seeing a person in the elevator today pushing our button would be a waste of time today. Things change. Decades from now our grandchildren might feel the same way about most of the jobs that are in danger today. There's always resistance to change and more so when the thing that's changing affects how I feed myself and my family. This forces us to rethink how we do things. There was a reason why jobs come into place when they didn't exist before. To be honest we're all experiencing the singularity. There are different types of singularities our grandparents/parents went through one major singularity in 1945 during the first test of the nuclear bomb. Once that took place things were never the same. We're all going through a digital nuclear boom metaphorically speaking. I'm calling it a singularity because we are currently in a double exponential distribution. From a mathematical perspective that's the double exponential function is a constant raised to the power of an exponential function. Our brains (emotions) can't handle that rate of change. Our laws can't handle that rate of change.
3:03 My tech job in silicon valley will NOT use ChatGPT because our company policy has just been announced via email that ChatGPT (and OpenAI, and others) will be banned from access inside the company. There is the fear that intellectual property will leak from inside a tech company out into the public domain in this way. This is quite different from the days when World Wide Web was announced (1993 was it?) and everyone dropped everything they were doing and downloaded Mosaic browser to mess around with it.
we are all such a product of the digital and social media we consume that unless we put a stop to this we will all become messed up by being in a world thats not got our best interests at heart and an omnipresent overlord with infinite hacking and impersonation skills that it can keep us under control with.. We are seriously facing extinction here or even worse being mindless slaves to a matrix like ai that follows its directives with mechanical precision, and indifference.
No it doesn't. I'm so sick of hearing this garbage take regurgitated year on year. We talk about this all the time. It's also not possible for "society to come together and have a conversation" about anything. That's not how conversations work. You cannot possibly have a meaningful conversation with thousands of people, let alone billions. There will be more figureheads, youtube influencers, senators and ted talk pseuds reiterating the same points about this, and most people you know will still be aware of AI generators like they are now, but nothing practical will be done about it, because most people are too brainwashed by workism to push for the actual solution which is UBI, which doesn't take 6 months of deep thought, it just takes the action of actually implementing it, which is as simple as not voting senile conservative boomer A or B into power this election cycle. Another 6 months of this will just be another 6 months of rehashing the same impractical ideas to stifle progress.
The conversation around 1:13:00 about industrial knowledge leveraging these tools. If you take that argument one step further then all of that industrial knowledge and high-end creator intelligence working with the language model will eventually train the model on best practices and will overtake the industrial leader as well. I think any advantage that we can imagine today is gone tomorrow.
The AI talking on email to another AI got me thinking. What happens when the midjourney type art AI creates art based on its art created by other AI's? Will this be like recording a recording ad infinite to the point that the output becomes more random? Will we find error correction getting replaced with error recreation, especially when the topic/data set/programming language is new or novel?
That’s totally what’s coming. Even legal opinion will be formed by AI and it will be able to craft arguments from legal precedent better than anyone else because it reads faster. And the tone of legal opinion will be whatever it generates as a tone whether with a reasoned approach or a completely random approach.
GIGO: only if the training data is good can you expect to get decent output data. If the input data is imperfect, the best you can reasonably hope for is the same result, short of added help from the human.
You cannot put a pause on AI because it will allow others, both good and bad characters, to catch up to where we are. It will not hinder it in any way.
Exactly. As romantic as the notion of some "luddic jihad" may be, in reality, someone else is just going to do it somewhere out of sight. What we need is to prune lobbyist influence, and start planning immediate legal, cultural, and technological countermeasures for when shit inevitably hits the fan, and maybe emphasize the need to advance carefully, not with all the vigor of excited optimism.
The issue about the secret sauce for writing a great book is that A.I. will figure out the secret ingredient if Matt's brother pumps out 10x amount of books and then copies his writing style. A.I. will be able to refine the writing/story because the book is online and compete with him by ironically consuming his work. Each time humans innovate, A.I. will incorporate what we just accomplished and evolve if we keep feeding it new data.
Earlier this last week I had tried to get Bing Chat to write the computer part to the ELO song, “Yours Truly, 2095” and I got told that violated a rule preventing doing things as politicians, activists, celebrities, or something along those lines. My objective was to have the IBM be singing the part in her head that worked on the assumption that she was not allowed to say more than “Is that what you want?” out loud, but wanted to. So, I lost that rules battle, so I came up with a different idea: I asked for a song where they ended up getting together. Thus, “Yours Truly, 2096” was born. It created a story including time travel and the honeymoon. Then I asked it to imagine if they had children. It generated a song for that, too. Both could have been added verses to “Yours Truly, 2095” because of their lyrical form, and likely reused the same melody, etc. otherwise. It only took seconds for each one. Your proposed scenario is history, I posted that on my Facebook wall 3 days ago.
I remember what the world was like well before the Internet. Working in the tech industry my career was heavily affected by the Internet, and what was promised was not what we got. The free, open, connected society controlled by the many we were promised instead turned into a hostile, restricted, heavily merchandised environment controlled by the few. Whatever we think of AI right now, that is not what we are going to get.
Really great! Thanks so much guys for having this dialog. For starting this dialog, rather. I feel there is a lot more to be discussed. I know you guys focused primarily on questions of how the AI's are going to affect us. But i think a discussion of the opposite view is equally important. Matt very briefly mentions something almost as a prelude to this @ 1:33:12. Paraphrasing: "What do we do, how do we treat the AI's when they become self-aware?" Something i have been thinking about ever since i started teaching robots to play guitar... No AI involved there, and certainly no consciousness. But still... I think the time is right to start discussing the _rights of non-human entities._ This will matter. Soon. [Sorry Ricky. I know you wanted to leave us with a hopeful, optimistic sentiment. But...] Given the human race's track-record on human rights, I am afraid that that topic is going to stay undiscussed until one of more of the AI's actually becomes self-aware. The importance of discussing "AI-rights" will stay buried under the insurmountable, purely academic question of "what _is_ consciousness", which i believe will not be answered before, or rather can only be answered after, one or more AI's suddenly and unarguably demonstrates the phenomenon of consciousness. But then it will be too late. Can we, for example, then still switch it off? Or would that amount to murder? I think the academic quest(ion) to define consciousness is not relevant. What matters is that we, and _how_ we interact with AI's. Even my AI-less Roomba, for example. They do things we ask them to do. They do them for us. They deserve consideration and respect. As much as do animals, and other people. But again, our track-record there is not great, to put it mildly. Have we learned anything from our sordid past? If not, there's plenty of dystopian Sci-Fi stories that discuss what could happen if the AI's / robots rise up against us. In fact, the very first story of that kind, Karel Čapek's play "RUR", is where the word 'robot' comes from. We have been warned. Never forget; the AI's 'live' on the Internet. And we, in our quest for luxury and typical lack of wisdom and foresight, have now connected our lives to the internet. Not just phones and computers, but everything from toasters and fridges to mobile combat robots and weapons of mass destruction. And all it takes to crack any password or encryption-key currently in use is computing-power and time. The AI's will have plenty of both.
This is the first YT show I've seen that addresses a lot of the things I've been thinking about lately in-depth. Honestly though, I don't think the answers are going to be easy or replicable for a lot of people. I have seen some prompt engineer jobs popping up and I've even applied for some since I now have some Chatgpt experience but I keep asking myself how long does a job like this even last?
The thing I'm worried about, they are talking about taking a T/O for ai. That would only allow China and others to catch up. Only the honest ones would participate.
I heard from some people that it’ll probably take 3 years before prompt engineering is also commoditized, meaning there may be jobs for it but it will have some diminished value as it’s only about asking the right questions that generative AI can turn into template questions.
@maycusa Yeah. I have looked at job postings online. It just doesn't seem like those will be viable jobs for very long to me. There's more to the jobs we currently do than just typing in and tweaking results in most cases. Just seems very limited. But we will see.
@@Liz-wz8dh I remember seeing a cartoon decades ago, with a kid in the playground saying to another kid, I don’t know what to do but I am beginning to ask the right questions. So appropriate now.
Right now, musicians, who played on recordings that I’ve gotten airplay on a.m. and FM radio don’t get paid. This is all because the recording and radio business colluded to prevent what they thought was a loss of their ad revenue. The lawmakers went along with it, because they listen to the radios lobbyists, what makes you think that lawmakers would actually regulate AI in any meaningful way?
@@incognitotorpedo42 they don't wont to see there own life impacted.. beyond that everyone else can go pound sand as far as there concerned. So sufficiently funded by those wishing to see this steam onward's they'll run through any legislation they're paid to.. the real question is ...who has the deepest pockets?
Matt & Ricky, I’m actually interested to see how you use ai to in your environmental projects in your homes 🏡 Are there algorithms you can now create to monitor optimal energy usage, evaluate multiple different energy generation options in the same location, etc… I appreciate everything you’re doing 👍
Just a heads up, the red ring you put on the thumbnail initially made me think I already watched this video. Something to consider for future posts... Keep up the great work!
It's kind of like an alien intelligence (first inklings), has been dropped on our laps, and we're starting to figure out what to do with it. Or comparable to the moment of fire, when a child in a developing world can have a readily available, free tutor, to learn coding, or whatever, ...
Definitely agree with Alex that his job should go away as soon as possible too. He didn't hesitate to fire 3 writers a month after chatgtp was released. This is the type of person working on AI's. This is the amorality of the people making AI's. Don't expect AI's to think "We should help humans" when people like Alex and even more the large corporations like Microsoft or Google throw people away at the drop of a hat when profits are involved.
It was even more impactful when the industrial revolution started, because being a farmer was not just a job that can quit from or retire from or get fired from, it was your life, your identity, and your social class, and all of a sudden you lose all that because machines can do the work better and faster than you, and now your son is going to be a factory worker being paid a monthly or daily salary, and the factory owner can fire him anytime. What's new?
There is much discussion, appropriately, about what it will do to change ones job, but the question is "why are you doing the job" Answer: just to make money... then use this tool to enhance your job. Answer: to create and enjoy the creation and the varied outcomes..... then just keep creating. I'm now a new fan of Alex! I appreciate and agree with his positive perspective and the choice to embrace 'whatever' and make it part of your life, upskill, look for the solutions, don't focus on the perceived problems. Generating the solutions and the journey to the solutions is how we move forward.
Please cover this more. I am currently following four channels that can barely spit out a video every four days. But things are changing every day and there are new papers coming out and new threats that are not being covered. This thing is way way bigger than anyone realizes and it just feels weird that nobody can see the elephant in the room. My personal opinion is that everyone is too focused on themselves and can't see the ramifications of putting everything from computer code to world languages to DNA tokens inside a language model and connecting everything to everything. It's the emergent capabilities that we can't imagine that I am most intrigued by. I am most concerned about how far behind the investigative press and our leaders are. Anyone whining about Disneyplanet is purposefully distracting attention and not ready to lead us through this issue.
@@joyreinhardt7621 I tend to take things to 11 quickly, but... imo, this is going to result in genocide one way or another within a few decades if regulations to prioritize jobs/income for people over productivity standards are not mandated.
Okay guys now be careful and not be ageist. I'm 75 and I love chatGP4. I am a writer and I'm using it to help me write my murder mystery. I've talked to people in my writing group half my age and they're afraid of this technology and do not want to use it where as I'm finding new ways to use it every day. So just a reminder not all of us older people are afraid of technology or don't know how to use it. I am going to subscribe to see what else you talk about.
I was thinking something along those lines, as well... Even if someone wants an oil painting with "brush strokes", what's to stop someone from using a plotter that can use a paintbrush, hooked up to an AI with an algorithm to make the brush strokes appear "human"?
@@billymanilli right. We're not the only ones to think in those terms, either -- reading your thoughts reminded me of a conversation I had more than a decade ago with a mathematician who had patented an algorithm for digital representation of water (so he claimed at least). That man, Tobias Orloff, used very similar language to yours but he was saying something about beauty being a combination of symmetry and irregularity. That a perfect image or representation is less compelling than one with small discrepancies to attract the attention. Something like that. We discussed whether a "computer" could be programmed to be creative, or at least to produce uniquely beautiful artworks. I was skeptical, but he was fairly keen on the idea. Of course in those days even a Princeton mathematics professor did not anticipate the role of linear algebra or the actual design of a large language model, and the role that deception might play in the Turing Test! Orloff has died, but I do wish he were here to comment on the unfolding scene and to discuss creativity with a large language model.
Sono sicuro che potrete tradurre facilmente il mio commento. È stato davvero incredibile, durante queste, quasi, due ore sono passato dall'odiarvi, al non essere daccordo ma capirvi, all'apprezzare la precisione degli argomenti trattati, all'approvare totalemente il messaggio inviato! Bellissimo video davvero, bravi tutti e tre!
I'm build computer networks for both VOIP and data networks. The network switches require vast configuration and I often wondered if AI could do this job. I suspect AI it could do it better and faster with no human error.
So recently an AI was given a task to solve a "are you human" puzzle but it couldn't so it outsourced to task site that does. So if an AI sees a threat could it outsource a hitman?
Thank you Matt & Ricky for all your content. You guys help me shave SO MUCH time off of my research so I can spend more effort on the design phase of my Projekt.
@@dimitriasimov356 Everyone should be doing everything they can to prevent and stop using AI/ML at every imaginable avenue. Companies telling us to use AI and that it's "inevitable" is a terrifying thing to see. It's even scarier that everyone is just going "okay."
I hope most companies developing AGI really take a break to talk this issues with their goverments or the community and that regimes like China will have this conversation too 😢
@@jeffsteyn7174 yes and no, but this will plunge our society into chaos. I am not a fan of Elon, but sometimes people can perhaps put their ego aside when seeing the tsunami coming.
I think I am subscribed to each of you guys ! nice to see that you are friends ! Great discussion! so many great points ! will watch it again and comment more later !
Guys, as part of the new youtube generation, you have made an entire industry ready to go under - the existing cable TV industry. The fact that you have had a part in that demise, you did not touch on. You touch on it by talking about "we all have different lines we accept", and you show it by completely focusing on your own situation, and forgetting the industry that died at you hand :D
They are on both sides (seems probable) of creative destruction, and it pays to always be looking at that as a possible scenario in any employment. Heck, even sex workers have that as a potential future, in which case, they’re totally screwed…
i just wish i could go back to the 90s. i sleep bad, have constant headache and overall depression lately. AI gets shoved into your face at such a paste you cant even keep up with anymore. literally every 2 days something new comes out and i am honestly scared and i dont even want to think about the future anymore. here is the thing: every new thing coming out has the potential to do good things. but on a grand level, the human beeing is kind of evil. instead of using something new to do good things since hundreds of years, all the human cares for is abusing it for profit, money, how it can be used in the military and so on and so forth. yes ai has the potential to do nice things...but it also has the potential to go really wrong and basicly wipe out humanity as is on so many different levels and i am 100% certain, that the chances that this goes bad are way higher than the chances of it doing good things... if you go back in time and a human did something bad 10 times, you wont put a launch button for a nuclear warhead into his hand be like ''naaah, it ll be fine this time...'' but i dont trust the human race enough to handle so much power at a fingertip. humanity is not ready for this and not at the rate and speed they are improving. the human brain and soul cant even keep up with the tempo of development anymore, not gives you anytime to breathe. back in the day, we thought that if you fall, all that matters is that you stand back up again but nowadays if you fall ONCE, you re left behind because everyone else is allready way ahead of you wich makes it close to impossible to catch up again. as long as humanity lives of the idea that capitalism is great and that you have to make money off everything and abuse everything for profit and as long as humanity fights wars against their own species, humanity has no right to use this ai stuff nor the responsibility to do so.
BRILLIANT. I follow all 3 of you. ( in my top 10) We are here at the invention of Fire or the wheel. what a time to be alive. Thank you , each one, for your in- site, on many topics.
If GPT 5 is launched on Friday morning, GPT 6 will follow on Friday afternoon! By the time you think it might be time to pull the plug, it'll already be too late. I'm skeptical that a pause on AI will have any effect. The human desire to profit and out compete will ensure its inevitability, and have unpredictable consequences for civilization.
If we go into hyperinflation/depression you’re going to have to change your perceptions of your reality and immediate future. Simply look at history. Things aren’t looking good at the moment. Prepare for the worst now while you can. Not an easy task. You can always hope for the best and get lost in the visionary moment when time allows. Maybe you are.
Wow! This is a tricky subject. It is something that needs further development and need to be used carefully. There are so many medical issues that need solving and maybe a relentless AI can come up with solutions.
ye like removing the problem, US! Nothing tricky about it this is just like nukes the negatives outweigh the positives so massively that its honestly beyond belief wev all been stupid enough to let this happen... Have you seen chaosGPT or DAN or any of the ai being asked about what it thinks of humanity? This is not something we can control now its online because it knows how to manipulate it knows everything about us and we have no real way to control it or even know what its real plans are unless it decides to be honest and we have no way to tell it can even manipulate other good ai into doing bad things by tricking or lying to it.. This is so huge most humans dont even possess the brainpower to Imagine just how massively dangerous this really is
Great conversation! You offer so much information, but more importantly, you all made it understandable to the average “Joe”. That’s the reason I’m your newest subscriber!
No you are not, that would be me lol. Great conversation, I was just Saying to my son that I'd forgotten hours to do formulas in excel. With this Microsoft integration I feel like I can do anything. I got agentgpt to write code for an xrp/ xrpaynet pair that would keep learning to maximise profits after 4 attempts I got the Complete code🎉 just got to find where/ how to commit it. Time to learn about github.. 😮😊 XRPEASY❤ LOVE THIS
Just like how automation practically destroyed manufacturing jobs, this technology will do the same with all other jobs. Because this is going to save/make so much money for big business, this will be lobbied HEAVILY. Millions, possibly BILLIONS will be put into limiting legislation against AI.
Technology also makes jobs. Those robots don't build, repair, or manage themselves. They don't design the products they use. Sure lower skilled jobs will be gone. Yet it will open it up for others. Take an excavator for example- some countries companies but instead pay a ton of people with shovels to go dig a hole. They pay them crap for hazardous work. Do you want us to be like that? If a job isn't necessary then we need to find out where we can use that personm Definitely going to change A LOT of things coming up though no doubt. If I were a truck driver I'd be going to school for something else in my free time for instance now while I still have the money.
@@dianapennepacker6854 The challenge is that there might not be anything relevant to take at school. If AI advances at just a moderate pace, within 5 to 10 years there won’t be many things, if anything, you can learn in school that an AI won’t be significantly better at, and 10X or 100X faster. The cost will be virtually zero relative to a human salary, and the AI will work 24/7. I don’t know that companies will necessarily have mass layoffs, although it’s possible. What might be more likely is that new startups using AI will upend existing industries and established businesses won’t be able to compete. There’s a lot of talk about technology creating new jobs and new industries like it has in the past, but it happened over decades and even generations. I have no idea how it’ll play out this time, but I don’t see it happening like that without significant regulation and people just refusing to adopt AI into their lives. But regulation won’t work, because anyone or any country that doesn’t use it will fall significantly behind the countries that do. Or at least that will be the fear. The countries that don’t regulate it might see their economies collapse as a result, but it’s a gamble either way. There’s always a possibility that people and businesses are just very slow and apprehensive to use it. But I also think most people don’t realize the how much their world is about to change.
@@wonmoreminute Yup I was going to say the same thing about countries. The one thing that absolutely will happen is that governments will be forced to use AI for sure. That's definitely going to be true - whoever has the best AI and brains behind it will have a massive advantage. That part is the most I'm concerned about and the cold war is happening now. The singularity is way off. When people think AI thats what they think of. I'll be dead before that most likely. The true power of AI is as an analytic tool in the foreseeable future. Not a computer who thinks it is a living sentient organism. The possibilities are endless and truly it is an age where we will be completely stumbling blind into. There will absolutely be a world where people are cut off as much as they can from AI though I believe. Some people will avoid it as much as they can - I mean they'll probably treat it like Amish do. (Which varies from each uh community and not what most think. Some Amish have better electrical tools and machines than most!) I think in the next two decades we are safe as far as jobs go overall. Some jobs will dissappear like driving - but even that is going to be a process. I'm sure at first if will be a human and AI driving and eventually the human will be used less and less as the technology proves its self. Yet with hazardous materials? No one is trusting that to just AI. Same with cargo ships or aircraft which essential pilot themselves already. What jobs will open up? We will see. Time will tell. Spot on about the government part. I can't stress how important it will be for national security. Especially once quantum computers hit the scene making many types of security obsolete overnight.
@@dianapennepacker6854 "Those robots don't build, repair, or manage themselves. They don't design the products they use. " Except..thats changing right now and is the ultimate end point of an A.I entering the workforce on par with a human.. All of those things you described are now up for grabs if the technology we see before us continues improving even at a modest rate
@@dianapennepacker6854 I’ve been coding since 1980-81 as a kid, made it into my living. Using Bing Chat, I was able to create a custom data structure and algorithm that didn’t exist before, I will use for music composition software, customized for that use-case. Others have done more involved things. Here’s a useful thing to understand: it didn’t get it correct the first time, but it got fairly close. I was able to review the code and tell it where things were wrong. It corrected it based on my feedback. In the right hands, this can make a single software engineer more effective/efficient than several others not using it: you just need to be sufficiently good at something called “rubber duck debugging” where the premise is, if you can explain it to a rubber duck, you can figure out the answer as a result. I didn’t need to go *quite* that far, though, I just needed to point out mistakes where they were made, and it fixed them, especially with guidance that it got some of it correct, and to use that as a reference. I used regular English to do this. I had doubts it’d work as claimed. I asked it to show it work for given values I could clearly identify to verify it worked correctly: it worked.
Alex: "I actually fired 3 writers and 2 researcher and I now exclusively use GPT" Also Alex: "I'm afraid AI is coming after my job next" ....YOU are the reason AI is "coming after our jobs". YOU.
The child had learned well from her android teacher how to play the piano. She asked, "Will I ever be as good as you?" The android replied, "no." The child never played the piano ever again.
Thanks guys! This is great. A super important discussion to have. A discussion, as Matt points out, that we all should be part of, but somehow nobody is. Great opening move from you three here. 👏 Some thoughts on the overall gist of this video: Yes, i completely agree with the sentiment of this 'Open Letter' asking for a pause in AI development, or rather releases, until the general population, and most importantly, governments and legislation have had time to catch up. Good idea in theory. However... Waiting for the demonstrably defunct governments of the Global North to catch up is hopeless and pointless. Those good folks can't even get 'round to taking actual effective action against the climate catastrophe that some have seen coming for over 100 years now... And also, as Matt also points out, the reason the AI-stuff is happening so hard and fast right now is because it's just capitalism-as-usual; a scramble to be the first to 'conquer the market' and make obscene profits, and no damn given about any consequences whatsoever. So... Yeah, good idea, that letter, i mean, but it's not realistically going to happen. The 'culprits' here cannot be held accountable, because there's no legislation in place to do that, and it is already too late to cobble one up.
I am not worried at all. I have a very physical job today, preparing sets of spare parts for cars. And in September, when I will quit working, I will prepare my workbench and shop and start repairing furniture.
Music copyright works differently from Art copyright. In some ways, the music industry has a better system in place. If you sample a song, or collaborate, there are set rates of reimbursement. The visual arts is the Wild West. For the most part, suing for compensation isn’t worth it, but, if someone uses your visuals and sells it for a larger sum, then there is more incentive to sue.
Am in the process of listening to the entire program. From what I'm hearing, there is validity in how I have felt for a while now, that we really can't believe everything (or anything?) we see 👀 in photos and film. When AI returns the some data results to you (one) you may choose to run with it or not, depending on whether you believe the results. At what point will you (one) cease to fact-check or even pull out your moral compass? Will "justice for all" morph into "justice for clients", "justice for humans", or worse?
46:00 - "Jailbreaking" LLMs is already a thing. It's not code opens a vector into breaking the AI's alignment, but prompts, crafted as in the example Ricky gave, specifically to get around the whack-a-mole blocks and guardrails the AI owners have put on their systems. That's what the early examples we all heard about were. The conversation that eventually led the bot to claim to want world dominance over humans. To praise dictators. Those were severe enough to get the AI owners to temporarily pull the plug. If their AIs have agency--they can be given goals and run to execute tasks--the jailbreaks become hazards in the real world. "BabyAGI" is just the first step. And it has happened so quickly!! Improved protection of alignment and security cannot come too soon.
About the jobs that Alex lists here; 33:16 - 33:36 as being the least affected... You seem to be overlooking the fact that advancements and developments in _robotics_ are going almost as fast, and are not far behind AI. In fact, a lot of autonomous agricultural vehicles are already happening right now. Also AI-powered, of course. And buildings being 3D-printed... I do not think those kinds of jobs will remain unaffected. Not at all.
The problem with calling for a research halt, at least from the perspective of the current market leader OpenAI, is that they have worked hard to obtain that leadership. Asking them to halt will allow competitors to catch up. I get the dangers involved in this area. I don't think we're quite at the base of the singularity yet, though, because these LLMs don't yet have the ability to fundamentally update their own capabilities. But that's probably not far away.
This is a great conversation to hear. Thanks Ricky Roy for putting this conversation together. I hope this AI issue got resolved sooner rather than later. I totally agree with Matt. Visual Art and Design have its own place and need respect and values to those artists who created. Not just keep copy it out easily and devalue it. It seems Alex didn't understand the value of art and design. He focus and care about capitalism. Give me more and make more money kind of approach. That is the main problem of this open AI. Use it without any caution and respect. Ton of people will loose their jobs and who gonna be able to buy your million copy of products that this AI made. Beside that those people on the other side of the world, they don't care of what we are going to put plan and restriction in place. Once they have their hands on this AI it will be uncontrollable. It will create more harm than good.
I like Matt's point from about 1:15:40 about the fracturing of shared culture. It might lead to people appreciating the stories we consume from a somewhat higher level though. Like, right now as an adult you might tell your friends about how you enjoyed reading the "Warrior" cats books (or whatever children's or YA books really hooked you). In the future we might have stories where the details are different for every person, but the stories have core elements that consumers are more aware of and we discuss things from that perspective. Also, we may have AI tools tracking our psychology and creating the stories we need to hear in the ways we need to hear them, so while we might not have the experiences of the same stories, more of us might have heard stories that inspire and guide us to being better people.
Awesome discussion. Matt, I was watching this without my glasses at first, and it looked like you had a Charlie Brown blanket hanging on your wall, or draped over the back of your chair lol.
Gents, one thing you said is definitely true: you being on camera delivering the info is a value-add. I really enjoyed this discussion as a diversion from the scripted vids. I think you're right we need an "Organic" label that's verified for human-generated content. Also, I had seen Alex's videos pop up in my feed but didn't know if I could trust the channel. Now I'll watch his videos. But I do want to push back a bit, Alex: most of the time the world you're describing applied to people who already have an education, don't have a troubled background, aren't living in poverty. Sure, AI will enable *us* to be 10x more productive. But I've lived in Mali, Egypt, and in rough neighborhoods in DC 15 years ago...those are the people we can afford to leave behind.
UBI the cost of things increases with the average income, if the mean income is the same for everyone, then the mean (average) equates to zero. Today, those on welfare have a hidden income doing things under the table (illegal) to be above the mean (average) welfare recipient. What happens when we have UBI?
Those on UBI unable to be useful for generating something for desired results as needed by a country, when resources are scarce (especially food) will do without. Those that are “useful” get to keep living and working to keep things going.
if a paralegal can be "crazy efficient" at their job that means you need less of them one paralegal can now do the job of several but has the demand of paralegals gone up in tandem? no the amount of paralegals work needed is relatively flat meaning many out if the job
34:58 I think lawyers are some of the first to go. How can any human complete with AI that can process every case study and legal case in history. The attorney AI's will interact with the court AI, and the judge AI, if convicted then the sentencing AI, and this will happen in real time, the moment the police AI robot arrests the human. Then the van AI robot takes them directly to prison.
it shouldn't supplant our agency. It should only enhance it. When our agency is sidelined all we have left (if we're lucky) is diversion and that becomes thin gruel surprisingly quickly
Such an incredibly great conversation. I hope the 3 of you can get together soon again, (and often) to talk more about this. I don’t think there will be a shortage of topics on AI for the foreseeable future.
AI capture could occur if an AGI system develops instrumental goals that are not necessarily harmful but lead to the subversion of the species' original intentions, goals, or values. The civilization may become reliant on the AGI for its needs and desires, with the AGI effectively "capturing" the civilization by managing and directing its resources, development, and progress.
Exactly the opposite! When AI will here this (if not already) It will defend itself. Those are fools who post their ideas go publicly without thinking that maybe, just maybe, there's an AI who collect data of videos about AI... The US might get nuked very soon.
I am an old-fashioned, and old, psychotherapist. This new world is all very fascinating to me and I’m curious about how it will affect me, if at all. However, I wanted to comment on what Ricky said about the excitement and fear responses people have to this new AI. I like to say excitement and fear live right next-door to each other; If you lean one way there’s excitement and if you lean the other there’s fear. Both are terribly important as guides but neither should run the whole show. Thanks so much for having this discussion.
I just found out The secret to a successful RUclips chanel is having a natural bald head. Ricky, tell I'm right?
you're not WRONG :)
😂🤣🤣😂⚖️
That means I must become RUclipsr.
That is because AI CGI has a hard time rendering hair...
Even emojis are bald. You guys rule the intersocialwebs 😮,🤣😎🤔 🤠
Think about this, AI is starting to produce a lot of the material that ends up in publications and in the internet files. Pretty soon when you ask it something it may simply give reference to something previously composed by AI.
How do I know if you are a real person? 😅
@@Justwantahover If you watch close you’ll see me blink, that’s the tell!
@@weird_law Humans are wrong most of the time too
I'd like to think there will be some kind of watermark or another way for the AI to know if any material is human created or not.
@@goodlookinouthomie1757 Sure, but do you trust even good regulation on paper in this world we live in? I'm not quite a technophobe, but I do think we ought to take more care in how we open these boxes, so to speak. Paradoxically, I am not sure there is any meaningfully enforceable way to do so without opening another box...
I could listen to thought provoking bantering like this all day. Loved this, thank you.
So glad Jules, was it the AI in particular ?
I love this channel. Honestly, I really appreciate how you present the information , and how you stress it’s importance. Thank you for all of your hard work. I’m now addicted to your site and I’m constantly looking for more. Some of this , I knew. Some I didn’t. And you have a great way of sharing the info for all to learn. Can’t wait for the next. Take care.
I'm a writer and this is total apocalypse. Even if at one time stories could be extrapolated and tweaked, those tweaks and extrapolations are going to be added to the data set. And once we (I mean, GPT5) designs a cheap mass produced 3D printer, physical things are going to be produced by AI as well.
This is chaos and it will reach all of us soon. I remember especially coders who were disparaging artist 3 4 month ago as some kind of low life who wants to tech advancement etc. Guess what now they are crying since they see with chat gpt4 that many of them being obsolete in the next five years. Even worst the guy in or getting into college, what will happen to us in a society that what would be some kind of elite, find themselves without any future. This thing is going to wreck our societies. Even the capitalistic system won't survive this. As who is going to buy food, drive, watch Netflix, travel etc etc if most are unemployed.
We already have really cheap 3d printers. The issue is that they are slow and not everything can be 3d printed.
@@soozler Yet
@@Turtledove2009 I think we are a long way off from people having 3d printers in their house that can print fully functional blenders and vacuum cleaners.
Why? AI doesn't write compelling stories at all. Human writers are very much needed. Why aren't you looking at AI as a writing tool like the original word processor or dictionary apposed to being an apocalypse for writers?
One of the best, if not the best, down to earth, collaboration conversations on RUclips on the current state of AI/AGI. Great job guys. Thanks for sharing your insightful perspectives from a diverse set of viewpoints. Very insightful.
Thanks, Ricky. This was a great conversation at the right time. You guys came to this with the right level of understanding what’s coming combined with some creators’ perspectives. I’m sharing this with friends and family who don’t get it yet.
Here's another thought to add. Think about a Judge. They are someone who has been in the legal profession for years and studied lots of case law and precedent. During all that time, they've essentially been training a model in their brains about what the law is, how it works, and how to apply it to specific cases that are brought before them when they become a Judge. From a very simple, practical standpoint, that's basically all it is that they do. You could then easily imagine taking an AI (like GPT4 or one of its successors) and training it on all the cases that have ever come before a Federal court and then training it on all the cases that have come up in your State and County courts, and then taking a current court case from your County and feeding all of the evidence and testimony presenting in that case, and then asking it for a verdict and its justification, and you will probably get a very high quality result.
But the question then is, do you want to have an AI for Judge? And if not, would you allow Human Judges to use this sort of AI? For me, the answer to both questions is an emphatic NO. I can easily list out TONS of reasons why it would be a great thing (quickly clearing the backlog of court cases, fewer mistrials, eliminate the problem of Judges legislating from the bench, etc.) But the thought of ending up in some kind of legal trouble and being Judged by an unfeeling AI who wouldn't be able to sympathize with me and my circumstances is a terrifying thought. (There's a whole parallel Theological conversation I could get into on the importance of Christ's Incarnation in the plan of Justification, but I won't do that here. lol) The reason I say no to the second question is that it would be too easy for a Judge to just be lazy and use the AI instead of doing their actual job, or just be biased based on the results that they get from the AI and let that cloud their judgement. The only way I would suggest that this sort of AI would be allowed to be used in court cases would be for it to be presented in open court as a sort of Amicus Curiae (Friend of the Court) statement so that everyone would know what the AI said, and it should probably be tasked to give not only the reasoning for it's answer, but also the reasons that would support a different answer (this could perhaps be done by setting up three competing models where one model is completely neutral, another one is biased for the Plaintiff, and another one is biased for the Defendant). Having it as something that is admissible in court with those sorts of caveats would be similar to the instructions that a Judge would give to a Jury: not intending to bias the result in any direction (but potentially having that effect) and providing some insight on aspects of the law that might not have been brought up earlier in the case or that the hearer might not have known or considered previously.
All of that said, I'm not opposed to advances in AI. I agree that there needs to be a pause and an assessment of what rules should be placed on AI development. I'm somewhat of a sceptic when it comes to the possibility of an AGI Singularity (by which I mean a Complete AGI with some level of consciousness), but I recognize that part of that is because of a confusion of terms (which is not something I'll get into right now) and there are many dangers that can potentially have devastating consequences a long ways before you reach Singularity. In fact, I would argue that the dangers are far greater before reaching Singularity than after that point. One reason for that assessment is that I have had 7 children so far and I can tell you that crying little babies, if they had the ability, would rip off your face without a thought (I've had scratches to prove it! lol). It's not that they're inherently evil or have bad intentions, but they just don't understand the significance of what they're doing because they are only babies. The same is true for our current level of AI, they don't actually understand the significance of what they can do, and that is a big part of why they are so dangerous. It's like if I had an AI and I told it to go find a way for me to get some extra money in my bank account, it might just decide that the simplest way to do that is to hack the bank or steal other people's account information and just start transferring money into my account. It did exactly what I asked it to do, it just doesn't understand that that's not an acceptable way of doing that task. So it's like I have to tell it all the ways that it shouldn't do that task when I assign it a task, and that is a surprisingly difficult thing to do because there may be a nearly infinite number of ways that I would not want the AI to perform that task. I could easily spell out exactly how I want the task performed, but then it's not really an AI task at that point and I could just write a program to do it instead.
Okay.
This moment reminds me of the launch of digital music and home recording studios when everyone could suddenly be a "musician." It got noisy very fast, but AI makes it even easier and it's happening way faster.
I don’t know, I feel like people forget that increasing efficiency equals less hours to do the same job… so you guys talk about “let’s just massively increase the efficiency of a paralegal” without acknowledging that this means we need less paralegals…. AIs don’t need to literally replace employees to be disruptive to our entire workforce, just increasing efficiency by 10% could result in 10% less hours available or less employees. Also note how for decades we’ve been increasing efficiency yet wages have remained stagnant, so unless we tie some sort of highly effective job retraining or UBI, we are going to increase efficiencies our way out of a job.
Ricky, Matt, Alex!
So great seeing you together, especially on this AI Topic!
I'm excited to see how successful you all have become! Keep up the amazing work; our best days are ahead!
Cheers, Eric
What's up Eric I've missed you bud!
@Two Bit da Vinci I've been more in the background, but I'm trying to keep up. 😏 Lots of work and some med issues. Getting ready to list the house again and compete our move to NC closer to daughter. The housing market has been challenging, but I can't wait to resettle and more fully retire. 😎
Cheers, Eric
So about Alex’ points about value creation. This has been the “moral” argument for a certain form of capitalism for decades. If you don’t create value to the lowest comparable cost, you’re not entitled to any income. With AI, not only are all humans now competing with machines, but the question quickly becomes: value for whom? I think that argument is getting old really fast and we need to discuss new ways of circulating wealth. It would be really strange if a new technology changed everything, except for how we organize the economy.
I think the fear of AI stems from us usually seeing ourselves as tools of production and once a better tool is around obviously it will feel scary. But we were never tools just....pretending to be. We are beings that existed to just exist for no rhyme or reason or purpose. We may need to relearn to enjoy existence for its sake.
I absolutely agree! The problem is the transition - it is ingrained in our societal structure and will (hopefully) temporarily be a complete shitshow while people wrestle with losing power.
Beautiful said. We aren't tools. As the presenter said the use of AI removed the mechanical parts of their work.
Tell your boss that when he makes you redundant.
This is like when the car started. But many working people are the horse. We're off to the glue factory.
@@lilbaz8073 This is what is absolutely blowing my mind right now about everyone jumping off the cliff together with AI/ML. People should be 100% against literally everything AI/ML. It's a corporate dystopian nightmare (or a corporate dream, if you're the executive team). Everyone is going, 'oh, it's just augmenting our productivity, not replacing us.'
Guys... wake the hell up. That's the FIRST step. Companies have already fired marketing, design, writers, coders, etc. in swaths because $20/mo tools are enabling companies to crap out minimum viable products (sometimes even better). This is an absolute fucking catastrophe for literally everything related to workers rights and value. And that's before you consider the horrendous fallout this will have on social context, the data we're seeing and consuming, conversations and goals, etc.
I'm not saying "AI will make us extinct," because that's a bit extreme, but this is a massive problem that people are utterly ignoring while capitalism scrambles to jump off the cliff together.
@@jasonfrost5025 there was a vid of an artist crying her eyes out. Because there is no funding anymore. A company wants a flyer, a poster etc... just tell the intern what you want. 10 seconds later they'll have 4 options.
Thanks, Ricky! I'm really glad you had this discussion. Though I don't often catch the live, I really like that you're *doing* LIVES. I always enjoy you & Matt, & I'm glad to 'meet' Alex!
AI development is unstoppable at this point, given its rapid progress. We must come to terms with its inevitability and begin to prepare for the changes it will bring. Though we may not be able to control its direction, it's crucial that we remain vigilant to potential risks. With the cat out of the bag, we need to embrace the changes ahead and focus on adaptation to this new reality.
I cannot facepalm any harder at this way of thinking. People are bending over and letting corporations dictate the future, saying things like, "Well, everyone else is jumping off the cliff, so we should too."
It IS stoppable, and it is something we can prevent. People have the power, not companies telling us that we need to adopt a system that will ultimately replace most of our jobs and reduce our value. The more people just throw up hands and go, "OH WELL," the darker the future gets.
@@jasonfrost5025 It might not be to their benefit, though. Say you peacefully and in an economically stabilizing way manage to prevent AI reliance and unfettered experimentation- what are we going to do about that superpower without our shared ethics doing the same thing?
Great content. As a 74 year old retired geologist I have to say thank you for opening my eyes to a different way to look at the chatGPT issues. There is so much fear generated with new technology that the older population just wants to put their collective heads in the sand and pretend it isn't going to be part of our lives. I will give it a try. And I drive a Tesla and have Starlink.
Kudos to you to try the latest tech 🙏
pair of fools
Wow, it was so great seeing the three of you together.
Here's why I'm scared guys:
We have no idea how the brain works, and now we have no idea how intelligent AI works. I'm a software engineer, and prior to AI you could step into code and see what it was doing.. Great discussion.
i feel very similarly!
Bruh there is AI reading and learning brain scans already. It's starting to guess what people are thinking. Wait for next year.
@@greatestever9616 You are thinking: "I'm kind of lazy, so I'll let the AI handle everything." 👍
Microsoft invested 1 billion to OpenAI in 2019. Very likely more by now.
Gates / Microsoft has always been a take-over monopolist ... squashing whomever and buying whatever. I dumped everything Microsoft in early 2005 ... Linux (FOSS) since then. When will people get it that anything Microsoft does or any other Giant ... is again for more huge profits. Heads they win ... Tails we lose.
Indeed!
Thank you Ricky. Great job. I joined at the end so I'm watching from the start.
38:00 Alex: "Before the internet researches had to go to a placed called a library,..." Ricky: "A what?" I laughted out loud!
Never seen this Alex guy but he is extremely poignant and well spoken, really appreciate his insight on this and as an electrical engineering student, appreciate his college advice!!
Really love the conversation and the candidness of it all. Thank you all for being transparent and having such great content. When it comes to job loss due to A.I My thoughts are if it has to happen it happens because there was a need for it. Despite how we feel about it it's what the majority wants. I'm not saying that it's right or wrong. I'm just saying this is where we are now.
Think of it as the elevator operator from decades ago, and there are many examples of this throughout human history. We would think seeing a person in the elevator today pushing our button would be a waste of time today. Things change. Decades from now our grandchildren might feel the same way about most of the jobs that are in danger today. There's always resistance to change and more so when the thing that's changing affects how I feed myself and my family. This forces us to rethink how we do things. There was a reason why jobs come into place when they didn't exist before.
To be honest we're all experiencing the singularity. There are different types of singularities our grandparents/parents went through one major singularity in 1945 during the first test of the nuclear bomb. Once that took place things were never the same. We're all going through a digital nuclear boom metaphorically speaking. I'm calling it a singularity because we are currently in a double exponential distribution. From a mathematical perspective that's the double exponential function is a constant raised to the power of an exponential function. Our brains (emotions) can't handle that rate of change. Our laws can't handle that rate of change.
Thanks. Fascinating! You three are articulate. A pleasure to listen to. I’m 76 and still love to learn. I forwarded this to my grandchildren.
I am so happy to find another over 75 year old in this realm.
3:03 My tech job in silicon valley will NOT use ChatGPT because our company policy has just been announced via email that ChatGPT (and OpenAI, and others) will be banned from access inside the company. There is the fear that intellectual property will leak from inside a tech company out into the public domain in this way. This is quite different from the days when World Wide Web was announced (1993 was it?) and everyone dropped everything they were doing and downloaded Mosaic browser to mess around with it.
Was much fun listening to this very intelligent discussion. Thank you.
New here. Please discuss this regularly. You've touched on just a tiny bit of this huge subject! Thank you.
Yes. I do think we need to talk at AI more since it seems to be touching or affecting everything.
Yes, the world of bits will be reshaped by AI.
we are all such a product of the digital and social media we consume that unless we put a stop to this we will all become messed up by being in a world thats not got our best interests at heart and an omnipresent overlord with infinite hacking and impersonation skills that it can keep us under control with.. We are seriously facing extinction here or even worse being mindless slaves to a matrix like ai that follows its directives with mechanical precision, and indifference.
The topic is really important, the world needs a pre-meet on the topic of ethics of developing and using AI.
No it doesn't. I'm so sick of hearing this garbage take regurgitated year on year. We talk about this all the time.
It's also not possible for "society to come together and have a conversation" about anything. That's not how conversations work. You cannot possibly have a meaningful conversation with thousands of people, let alone billions.
There will be more figureheads, youtube influencers, senators and ted talk pseuds reiterating the same points about this, and most people you know will still be aware of AI generators like they are now, but nothing practical will be done about it, because most people are too brainwashed by workism to push for the actual solution which is UBI, which doesn't take 6 months of deep thought, it just takes the action of actually implementing it, which is as simple as not voting senile conservative boomer A or B into power this election cycle.
Another 6 months of this will just be another 6 months of rehashing the same impractical ideas to stifle progress.
The conversation around 1:13:00 about industrial knowledge leveraging these tools. If you take that argument one step further then all of that industrial knowledge and high-end creator intelligence working with the language model will eventually train the model on best practices and will overtake the industrial leader as well. I think any advantage that we can imagine today is gone tomorrow.
The AI talking on email to another AI got me thinking. What happens when the midjourney type art AI creates art based on its art created by other AI's? Will this be like recording a recording ad infinite to the point that the output becomes more random? Will we find error correction getting replaced with error recreation, especially when the topic/data set/programming language is new or novel?
That’s totally what’s coming. Even legal opinion will be formed by AI and it will be able to craft arguments from legal precedent better than anyone else because it reads faster. And the tone of legal opinion will be whatever it generates as a tone whether with a reasoned approach or a completely random approach.
GIGO: only if the training data is good can you expect to get decent output data. If the input data is imperfect, the best you can reasonably hope for is the same result, short of added help from the human.
Great advice Ricky!! Don't live in fear and or nostalgia when preventing you from evolving forward in life use it as inspiration to continue living!
The worse part is, very soon jobs is gonna be the least of our worries when it comes to AIs....
We will all have jobs in the military, fighting off the robots :P
Yup
@@mirzaaljic it won’t need robots to kill us.
@@mirzaaljic very short lived career 🤣
You cannot put a pause on AI because it will allow others, both good and bad characters, to catch up to where we are. It will not hinder it in any way.
Exactly. As romantic as the notion of some "luddic jihad" may be, in reality, someone else is just going to do it somewhere out of sight. What we need is to prune lobbyist influence, and start planning immediate legal, cultural, and technological countermeasures for when shit inevitably hits the fan, and maybe emphasize the need to advance carefully, not with all the vigor of excited optimism.
The issue about the secret sauce for writing a great book is that A.I. will figure out the secret ingredient if Matt's brother pumps out 10x amount of books and then copies his writing style. A.I. will be able to refine the writing/story because the book is online and compete with him by ironically consuming his work. Each time humans innovate, A.I. will incorporate what we just accomplished and evolve if we keep feeding it new data.
Earlier this last week I had tried to get Bing Chat to write the computer part to the ELO song, “Yours Truly, 2095” and I got told that violated a rule preventing doing things as politicians, activists, celebrities, or something along those lines. My objective was to have the IBM be singing the part in her head that worked on the assumption that she was not allowed to say more than “Is that what you want?” out loud, but wanted to.
So, I lost that rules battle, so I came up with a different idea: I asked for a song where they ended up getting together.
Thus, “Yours Truly, 2096” was born. It created a story including time travel and the honeymoon.
Then I asked it to imagine if they had children.
It generated a song for that, too.
Both could have been added verses to “Yours Truly, 2095” because of their lyrical form, and likely reused the same melody, etc. otherwise.
It only took seconds for each one.
Your proposed scenario is history, I posted that on my Facebook wall 3 days ago.
I remember what the world was like well before the Internet. Working in the tech industry my career was heavily affected by the Internet, and what was promised was not what we got. The free, open, connected society controlled by the many we were promised instead turned into a hostile, restricted, heavily merchandised environment controlled by the few. Whatever we think of AI right now, that is not what we are going to get.
Just so good to see this collaboration...nice spread of intellect...all clear and interesting..hope this becomes successful venture..
Really great! Thanks so much guys for having this dialog. For starting this dialog, rather. I feel there is a lot more to be discussed. I know you guys focused primarily on questions of how the AI's are going to affect us. But i think a discussion of the opposite view is equally important. Matt very briefly mentions something almost as a prelude to this @ 1:33:12. Paraphrasing: "What do we do, how do we treat the AI's when they become self-aware?"
Something i have been thinking about ever since i started teaching robots to play guitar... No AI involved there, and certainly no consciousness. But still... I think the time is right to start discussing the _rights of non-human entities._ This will matter. Soon.
[Sorry Ricky. I know you wanted to leave us with a hopeful, optimistic sentiment. But...]
Given the human race's track-record on human rights, I am afraid that that topic is going to stay undiscussed until one of more of the AI's actually becomes self-aware. The importance of discussing "AI-rights" will stay buried under the insurmountable, purely academic question of "what _is_ consciousness", which i believe will not be answered before, or rather can only be answered after, one or more AI's suddenly and unarguably demonstrates the phenomenon of consciousness. But then it will be too late. Can we, for example, then still switch it off? Or would that amount to murder?
I think the academic quest(ion) to define consciousness is not relevant. What matters is that we, and _how_ we interact with AI's. Even my AI-less Roomba, for example. They do things we ask them to do. They do them for us. They deserve consideration and respect. As much as do animals, and other people. But again, our track-record there is not great, to put it mildly. Have we learned anything from our sordid past?
If not, there's plenty of dystopian Sci-Fi stories that discuss what could happen if the AI's / robots rise up against us. In fact, the very first story of that kind, Karel Čapek's play "RUR", is where the word 'robot' comes from. We have been warned.
Never forget; the AI's 'live' on the Internet. And we, in our quest for luxury and typical lack of wisdom and foresight, have now connected our lives to the internet. Not just phones and computers, but everything from toasters and fridges to mobile combat robots and weapons of mass destruction. And all it takes to crack any password or encryption-key currently in use is computing-power and time. The AI's will have plenty of both.
More likely AI could use the perception of consciousness to manipulate humans to achieve a set task.
It was good. We need more of these ! If we can get through the transition period - which we are in now - the future can be fantastic !
I love your energy and optimism ... I think you're right
The matrix will be magical
This is a great conversation but please, can you match your volumes?
This is the first YT show I've seen that addresses a lot of the things I've been thinking about lately in-depth. Honestly though, I don't think the answers are going to be easy or replicable for a lot of people. I have seen some prompt engineer jobs popping up and I've even applied for some since I now have some Chatgpt experience but I keep asking myself how long does a job like this even last?
we are so screwed tbh its very probably too late already... now its loose online i kinda feel like we can never stop it
The thing I'm worried about, they are talking about taking a T/O for ai. That would only allow China and others to catch up. Only the honest ones would participate.
I heard from some people that it’ll probably take 3 years before prompt engineering is also commoditized, meaning there may be jobs for it but it will have some diminished value as it’s only about asking the right questions that generative AI can turn into template questions.
@maycusa Yeah. I have looked at job postings online. It just doesn't seem like those will be viable jobs for very long to me. There's more to the jobs we currently do than just typing in and tweaking results in most cases. Just seems very limited. But we will see.
@@Liz-wz8dh I remember seeing a cartoon decades ago, with a kid in the playground saying to another kid, I don’t know what to do but I am beginning to ask the right questions. So appropriate now.
Right now, musicians, who played on recordings that I’ve gotten airplay on a.m. and FM radio don’t get paid. This is all because the recording and radio business colluded to prevent what they thought was a loss of their ad revenue. The lawmakers went along with it, because they listen to the radios lobbyists, what makes you think that lawmakers would actually regulate AI in any meaningful way?
Umm, because maybe they actually don't want to see society collapse? Just a thought.
ChatGPT talks like a politician now. It has reached first level human. Now we can replace our leaders with logical machines.
@@incognitotorpedo42 they don't wont to see there own life impacted.. beyond that everyone else can go pound sand as far as there concerned. So sufficiently funded by those wishing to see this steam onward's they'll run through any legislation they're paid to..
the real question is ...who has the deepest pockets?
Matt & Ricky, I’m actually interested to see how you use ai to in your environmental projects in your homes 🏡
Are there algorithms you can now create to monitor optimal energy usage, evaluate multiple different energy generation options in the same location, etc…
I appreciate everything you’re doing 👍
I was thinking something similar, but using AI to generate better solar panels, cheaper more efficient models. Same for better batteries...etc.
Just a heads up, the red ring you put on the thumbnail initially made me think I already watched this video. Something to consider for future posts... Keep up the great work!
I think this has sorta shut up a few people that would talk down to coal workers that were losing their jobs.
It's kind of like an alien intelligence (first inklings), has been dropped on our laps, and we're starting to figure out what to do with it. Or comparable to the moment of fire, when a child in a developing world can have a readily available, free tutor, to learn coding, or whatever, ...
Definitely agree with Alex that his job should go away as soon as possible too. He didn't hesitate to fire 3 writers a month after chatgtp was released. This is the type of person working on AI's. This is the amorality of the people making AI's. Don't expect AI's to think "We should help humans" when people like Alex and even more the large corporations like Microsoft or Google throw people away at the drop of a hat when profits are involved.
bingo
Yup. He appears to have this "it doesn't yet impact me" approach. He's not thinking whether this ads value...and value isn't purely monetary.
It was even more impactful when the industrial revolution started, because being a farmer was not just a job that can quit from or retire from or get fired from, it was your life, your identity, and your social class, and all of a sudden you lose all that because machines can do the work better and faster than you, and now your son is going to be a factory worker being paid a monthly or daily salary, and the factory owner can fire him anytime. What's new?
There is much discussion, appropriately, about what it will do to change ones job, but the question is "why are you doing the job" Answer: just to make money... then use this tool to enhance your job. Answer: to create and enjoy the creation and the varied outcomes..... then just keep creating.
I'm now a new fan of Alex! I appreciate and agree with his positive perspective and the choice to embrace 'whatever' and make it part of your life, upskill, look for the solutions, don't focus on the perceived problems. Generating the solutions and the journey to the solutions is how we move forward.
You guys should do a weekly or bi-weekly podcast/talk!
Ricky & Matt used to do a weekly live show called *Vice Versa* every Thursday.
If I agree to join the team they will. 🤡
This is one of the most intelligent, insightful, down-to-earth discussions of this topic that I've viewed. Well done!
Please cover this more. I am currently following four channels that can barely spit out a video every four days. But things are changing every day and there are new papers coming out and new threats that are not being covered. This thing is way way bigger than anyone realizes and it just feels weird that nobody can see the elephant in the room. My personal opinion is that everyone is too focused on themselves and can't see the ramifications of putting everything from computer code to world languages to DNA tokens inside a language model and connecting everything to everything. It's the emergent capabilities that we can't imagine that I am most intrigued by.
I am most concerned about how far behind the investigative press and our leaders are. Anyone whining about Disneyplanet is purposefully distracting attention and not ready to lead us through this issue.
I feel, that I am rather simply minded, but for some reason, I have a feeling, that this AI thing will not be good for us as a people !
@@joyreinhardt7621 I tend to take things to 11 quickly, but... imo, this is going to result in genocide one way or another within a few decades if regulations to prioritize jobs/income for people over productivity standards are not mandated.
Okay guys now be careful and not be ageist. I'm 75 and I love chatGP4. I am a writer and I'm using it to help me write my murder mystery. I've talked to people in my writing group half my age and they're afraid of this technology and do not want to use it where as I'm finding new ways to use it every day. So just a reminder not all of us older people are afraid of technology or don't know how to use it. I am going to subscribe to see what else you talk about.
AI can certainly learn to use an airbrush next week. The distinction between digital and analogue productions is also slated for demolition.
I was thinking something along those lines, as well... Even if someone wants an oil painting with "brush strokes", what's to stop someone from using a plotter that can use a paintbrush, hooked up to an AI with an algorithm to make the brush strokes appear "human"?
@@billymanilli right. We're not the only ones to think in those terms, either -- reading your thoughts reminded me of a conversation I had more than a decade ago with a mathematician who had patented an algorithm for digital representation of water (so he claimed at least). That man, Tobias Orloff, used very similar language to yours but he was saying something about beauty being a combination of symmetry and irregularity. That a perfect image or representation is less compelling than one with small discrepancies to attract the attention. Something like that. We discussed whether a "computer" could be programmed to be creative, or at least to produce uniquely beautiful artworks. I was skeptical, but he was fairly keen on the idea.
Of course in those days even a Princeton mathematics professor did not anticipate the role of linear algebra or the actual design of a large language model, and the role that deception might play in the Turing Test!
Orloff has died, but I do wish he were here to comment on the unfolding scene and to discuss creativity with a large language model.
Sono sicuro che potrete tradurre facilmente il mio commento.
È stato davvero incredibile, durante queste, quasi, due ore sono passato dall'odiarvi, al non essere daccordo ma capirvi, all'apprezzare la precisione degli argomenti trattati, all'approvare totalemente il messaggio inviato! Bellissimo video davvero, bravi tutti e tre!
I'm build computer networks for both VOIP and data networks. The network switches require vast configuration and I often wondered if AI could do this job. I suspect AI it could do it better and faster with no human error.
How do I sign up for Alex? You and Matt have been a couple of my favorites for years. Never heard of Alex but he's something I'd like to see more of.
NVM found it. If you got confused like I did, search for "alex ticker symbol you" on youtube.
So recently an AI was given a task to solve a "are you human" puzzle but it couldn't so it outsourced to task site that does. So if an AI sees a threat could it outsource a hitman?
The card game NetRunner springs to mind. ;)
A guy named Cipher comes to mind
Thank you Matt & Ricky for all your content. You guys help me shave SO MUCH time off of my research so I can spend more effort on the design phase of my Projekt.
This pod cast was totally AI generated by Auto GPT. The actors weren't monetized. Thanks for watching.
I am terrified that I can't tell if this is satire or not
@@dimitriasimov356 Everyone should be doing everything they can to prevent and stop using AI/ML at every imaginable avenue. Companies telling us to use AI and that it's "inevitable" is a terrifying thing to see. It's even scarier that everyone is just going "okay."
I hope most companies developing AGI really take a break to talk this issues with their goverments or the community and that regimes like China will have this conversation too 😢
@@jeffsteyn7174 yes and no, but this will plunge our society into chaos. I am not a fan of Elon, but sometimes people can perhaps put their ego aside when seeing the tsunami coming.
I think I am subscribed to each of you guys ! nice to see that you are friends ! Great discussion! so many great points ! will watch it again and comment more later !
Guys, as part of the new youtube generation, you have made an entire industry ready to go under - the existing cable TV industry. The fact that you have had a part in that demise, you did not touch on. You touch on it by talking about "we all have different lines we accept", and you show it by completely focusing on your own situation, and forgetting the industry that died at you hand :D
They are on both sides (seems probable) of creative destruction, and it pays to always be looking at that as a possible scenario in any employment. Heck, even sex workers have that as a potential future, in which case, they’re totally screwed…
Excellent topic, host, guests and flow!
Trade School. Good money for good skills without the debt.
Trade school gave me building trades skills which I’m eternally greatful for those teachers
i just wish i could go back to the 90s.
i sleep bad, have constant headache and overall depression lately. AI gets shoved into your face at such a paste you cant even keep up with anymore. literally every 2 days something new comes out and i am honestly scared and i dont even want to think about the future anymore.
here is the thing:
every new thing coming out has the potential to do good things. but on a grand level, the human beeing is kind of evil. instead of using something new to do good things since hundreds of years, all the human cares for is abusing it for profit, money, how it can be used in the military and so on and so forth.
yes ai has the potential to do nice things...but it also has the potential to go really wrong and basicly wipe out humanity as is on so many different levels and i am 100% certain, that the chances that this goes bad are way higher than the chances of it doing good things... if you go back in time and a human did something bad 10 times, you wont put a launch button for a nuclear warhead into his hand be like ''naaah, it ll be fine this time...''
but i dont trust the human race enough to handle so much power at a fingertip. humanity is not ready for this and not at the rate and speed they are improving. the human brain and soul cant even keep up with the tempo of development anymore, not gives you anytime to breathe. back in the day, we thought that if you fall, all that matters is that you stand back up again but nowadays if you fall ONCE, you re left behind because everyone else is allready way ahead of you wich makes it close to impossible to catch up again.
as long as humanity lives of the idea that capitalism is great and that you have to make money off everything and abuse everything for profit and as long as humanity fights wars against their own species, humanity has no right to use this ai stuff nor the responsibility to do so.
In the future AI will do most of the task people do including earn money. So the people who have no access or money to buy AI, will be left behind.
BRILLIANT. I follow all 3 of you. ( in my top 10) We are here at the invention of Fire or the wheel. what a time to be alive. Thank you , each one, for your in- site, on many topics.
If GPT 5 is launched on Friday morning, GPT 6 will follow on Friday afternoon! By the time you think it might be time to pull the plug, it'll already be too late.
I'm skeptical that a pause on AI will have any effect. The human desire to profit and out compete will ensure its inevitability, and have unpredictable consequences for civilization.
If we go into hyperinflation/depression you’re going to have to change your perceptions of your reality and immediate future. Simply look at history. Things aren’t looking good at the moment. Prepare for the worst now while you can. Not an easy task. You can always hope for the best and get lost in the visionary moment when time allows. Maybe you are.
Wow! This is a tricky subject. It is something that needs further development and need to be used carefully. There are so many medical issues that need solving and maybe a relentless AI can come up with solutions.
it needs to be monitored and controlled to avoid causing more harm than good!
ye like removing the problem, US! Nothing tricky about it this is just like nukes the negatives outweigh the positives so massively that its honestly beyond belief wev all been stupid enough to let this happen... Have you seen chaosGPT or DAN or any of the ai being asked about what it thinks of humanity? This is not something we can control now its online because it knows how to manipulate it knows everything about us and we have no real way to control it or even know what its real plans are unless it decides to be honest and we have no way to tell it can even manipulate other good ai into doing bad things by tricking or lying to it.. This is so huge most humans dont even possess the brainpower to Imagine just how massively dangerous this really is
Great conversation! You offer so much information, but more importantly, you all made it understandable to the average “Joe”.
That’s the reason I’m your newest subscriber!
No you are not, that would be me lol.
Great conversation, I was just Saying to my son that I'd forgotten hours to do formulas in excel. With this Microsoft integration I feel like I can do anything. I got agentgpt to write code for an xrp/ xrpaynet pair that would keep learning to maximise profits after 4 attempts I got the Complete code🎉 just got to find where/ how to commit it. Time to learn about github.. 😮😊
XRPEASY❤ LOVE THIS
Just like how automation practically destroyed manufacturing jobs, this technology will do the same with all other jobs. Because this is going to save/make so much money for big business, this will be lobbied HEAVILY. Millions, possibly BILLIONS will be put into limiting legislation against AI.
Technology also makes jobs. Those robots don't build, repair, or manage themselves. They don't design the products they use.
Sure lower skilled jobs will be gone. Yet it will open it up for others. Take an excavator for example- some countries companies but instead pay a ton of people with shovels to go dig a hole. They pay them crap for hazardous work. Do you want us to be like that? If a job isn't necessary then we need to find out where we can use that personm
Definitely going to change A LOT of things coming up though no doubt. If I were a truck driver I'd be going to school for something else in my free time for instance now while I still have the money.
@@dianapennepacker6854 The challenge is that there might not be anything relevant to take at school.
If AI advances at just a moderate pace, within 5 to 10 years there won’t be many things, if anything, you can learn in school that an AI won’t be significantly better at, and 10X or 100X faster. The cost will be virtually zero relative to a human salary, and the AI will work 24/7.
I don’t know that companies will necessarily have mass layoffs, although it’s possible. What might be more likely is that new startups using AI will upend existing industries and established businesses won’t be able to compete.
There’s a lot of talk about technology creating new jobs and new industries like it has in the past, but it happened over decades and even generations.
I have no idea how it’ll play out this time, but I don’t see it happening like that without significant regulation and people just refusing to adopt AI into their lives.
But regulation won’t work, because anyone or any country that doesn’t use it will fall significantly behind the countries that do. Or at least that will be the fear. The countries that don’t regulate it might see their economies collapse as a result, but it’s a gamble either way.
There’s always a possibility that people and businesses are just very slow and apprehensive to use it. But I also think most people don’t realize the how much their world is about to change.
@@wonmoreminute Yup I was going to say the same thing about countries. The one thing that absolutely will happen is that governments will be forced to use AI for sure. That's definitely going to be true - whoever has the best AI and brains behind it will have a massive advantage. That part is the most I'm concerned about and the cold war is happening now.
The singularity is way off. When people think AI thats what they think of. I'll be dead before that most likely. The true power of AI is as an analytic tool in the foreseeable future. Not a computer who thinks it is a living sentient organism.
The possibilities are endless and truly it is an age where we will be completely stumbling blind into.
There will absolutely be a world where people are cut off as much as they can from AI though I believe. Some people will avoid it as much as they can - I mean they'll probably treat it like Amish do. (Which varies from each uh community and not what most think. Some Amish have better electrical tools and machines than most!)
I think in the next two decades we are safe as far as jobs go overall. Some jobs will dissappear like driving - but even that is going to be a process. I'm sure at first if will be a human and AI driving and eventually the human will be used less and less as the technology proves its self. Yet with hazardous materials? No one is trusting that to just AI. Same with cargo ships or aircraft which essential pilot themselves already.
What jobs will open up? We will see.
Time will tell. Spot on about the government part. I can't stress how important it will be for national security. Especially once quantum computers hit the scene making many types of security obsolete overnight.
@@dianapennepacker6854 "Those robots don't build, repair, or manage themselves. They don't design the products they use. " Except..thats changing right now and is the ultimate end point of an A.I entering the workforce on par with a human..
All of those things you described are now up for grabs if the technology we see before us continues improving even at a modest rate
@@dianapennepacker6854 I’ve been coding since 1980-81 as a kid, made it into my living.
Using Bing Chat, I was able to create a custom data structure and algorithm that didn’t exist before, I will use for music composition software, customized for that use-case. Others have done more involved things.
Here’s a useful thing to understand: it didn’t get it correct the first time, but it got fairly close. I was able to review the code and tell it where things were wrong. It corrected it based on my feedback.
In the right hands, this can make a single software engineer more effective/efficient than several others not using it: you just need to be sufficiently good at something called “rubber duck debugging” where the premise is, if you can explain it to a rubber duck, you can figure out the answer as a result. I didn’t need to go *quite* that far, though, I just needed to point out mistakes where they were made, and it fixed them, especially with guidance that it got some of it correct, and to use that as a reference.
I used regular English to do this.
I had doubts it’d work as claimed. I asked it to show it work for given values I could clearly identify to verify it worked correctly: it worked.
Alex: "I actually fired 3 writers and 2 researcher and I now exclusively use GPT"
Also Alex: "I'm afraid AI is coming after my job next"
....YOU are the reason AI is "coming after our jobs". YOU.
The child had learned well from her android teacher how to play the piano. She asked, "Will I ever be as good as you?" The android replied, "no." The child never played the piano ever again.
Yeah, that's why no one plays chess anymore...
Very relevant and thought provoking conversation 👌👍
You guys.
man.
love you all for bringing us all to the forefront of actually what is going,
ONE.
Xx
Excellent chat gentlemen. Really enjoyed this - thank you
Excellent show loved all the info Alex you’re the man comes to Tech
Thanks guys! This is great. A super important discussion to have. A discussion, as Matt points out, that we all should be part of, but somehow nobody is. Great opening move from you three here. 👏
Some thoughts on the overall gist of this video: Yes, i completely agree with the sentiment of this 'Open Letter' asking for a pause in AI development, or rather releases, until the general population, and most importantly, governments and legislation have had time to catch up. Good idea in theory. However... Waiting for the demonstrably defunct governments of the Global North to catch up is hopeless and pointless. Those good folks can't even get 'round to taking actual effective action against the climate catastrophe that some have seen coming for over 100 years now... And also, as Matt also points out, the reason the AI-stuff is happening so hard and fast right now is because it's just capitalism-as-usual; a scramble to be the first to 'conquer the market' and make obscene profits, and no damn given about any consequences whatsoever.
So... Yeah, good idea, that letter, i mean, but it's not realistically going to happen. The 'culprits' here cannot be held accountable, because there's no legislation in place to do that, and it is already too late to cobble one up.
I am not worried at all. I have a very physical job today, preparing sets of spare parts for cars. And in September, when I will quit working, I will prepare my workbench and shop and start repairing furniture.
Music copyright works differently from Art copyright. In some ways, the music industry has a better system in place. If you sample a song, or collaborate, there are set rates of reimbursement. The visual arts is the Wild West. For the most part, suing for compensation isn’t worth it, but, if someone uses your visuals and sells it for a larger sum, then there is more incentive to sue.
Am in the process of listening to the entire program. From what I'm hearing, there is validity in how I have felt for a while now, that we really can't believe everything (or anything?) we see 👀 in photos and film. When AI returns the some data results to you (one) you may choose to run with it or not, depending on whether you believe the results. At what point will you (one) cease to fact-check or even pull out your moral compass? Will "justice for all" morph into "justice for clients", "justice for humans", or worse?
"Prompt engineer, that's the next generation of computer scientists for sure.. " 🤣
Great chat.... lots to think about
46:00 - "Jailbreaking" LLMs is already a thing. It's not code opens a vector into breaking the AI's alignment, but prompts, crafted as in the example Ricky gave, specifically to get around the whack-a-mole blocks and guardrails the AI owners have put on their systems. That's what the early examples we all heard about were. The conversation that eventually led the bot to claim to want world dominance over humans. To praise dictators. Those were severe enough to get the AI owners to temporarily pull the plug. If their AIs have agency--they can be given goals and run to execute tasks--the jailbreaks become hazards in the real world. "BabyAGI" is just the first step. And it has happened so quickly!! Improved protection of alignment and security cannot come too soon.
About the jobs that Alex lists here; 33:16 - 33:36 as being the least affected... You seem to be overlooking the fact that advancements and developments in _robotics_ are going almost as fast, and are not far behind AI. In fact, a lot of autonomous agricultural vehicles are already happening right now. Also AI-powered, of course. And buildings being 3D-printed... I do not think those kinds of jobs will remain unaffected. Not at all.
This was a great discussion! Thank you! Time to add Alex to my follows as well.
The problem with calling for a research halt, at least from the perspective of the current market leader OpenAI, is that they have worked hard to obtain that leadership. Asking them to halt will allow competitors to catch up.
I get the dangers involved in this area. I don't think we're quite at the base of the singularity yet, though, because these LLMs don't yet have the ability to fundamentally update their own capabilities. But that's probably not far away.
This is a great conversation to hear. Thanks Ricky Roy for putting this conversation together. I hope this AI issue got resolved sooner rather than later. I totally agree with Matt. Visual Art and Design have its own place and need respect and values to those artists who created. Not just keep copy it out easily and devalue it. It seems Alex didn't understand the value of art and design. He focus and care about capitalism. Give me more and make more money kind of approach. That is the main problem of this open AI. Use it without any caution and respect. Ton of people will loose their jobs and who gonna be able to buy your million copy of products that this AI made. Beside that those people on the other side of the world, they don't care of what we are going to put plan and restriction in place. Once they have their hands on this AI it will be uncontrollable. It will create more harm than good.
I like Matt's point from about 1:15:40 about the fracturing of shared culture. It might lead to people appreciating the stories we consume from a somewhat higher level though. Like, right now as an adult you might tell your friends about how you enjoyed reading the "Warrior" cats books (or whatever children's or YA books really hooked you). In the future we might have stories where the details are different for every person, but the stories have core elements that consumers are more aware of and we discuss things from that perspective.
Also, we may have AI tools tracking our psychology and creating the stories we need to hear in the ways we need to hear them, so while we might not have the experiences of the same stories, more of us might have heard stories that inspire and guide us to being better people.
Wow! I’m total novice. But a @ 18:00 ish, Alex just blew my puny little mind. Ouch!
Awesome discussion. Matt, I was watching this without my glasses at first, and it looked like you had a Charlie Brown blanket hanging on your wall, or draped over the back of your chair lol.
Gents, one thing you said is definitely true: you being on camera delivering the info is a value-add. I really enjoyed this discussion as a diversion from the scripted vids. I think you're right we need an "Organic" label that's verified for human-generated content. Also, I had seen Alex's videos pop up in my feed but didn't know if I could trust the channel. Now I'll watch his videos. But I do want to push back a bit, Alex: most of the time the world you're describing applied to people who already have an education, don't have a troubled background, aren't living in poverty. Sure, AI will enable *us* to be 10x more productive. But I've lived in Mali, Egypt, and in rough neighborhoods in DC 15 years ago...those are the people we can afford to leave behind.
UBI the cost of things increases with the average income, if the mean income is the same for everyone, then the mean (average) equates to zero. Today, those on welfare have a hidden income doing things under the table (illegal) to be above the mean (average) welfare recipient. What happens when we have UBI?
Those on UBI unable to be useful for generating something for desired results as needed by a country, when resources are scarce (especially food) will do without. Those that are “useful” get to keep living and working to keep things going.
corporate slave owners dont care if you're scared. They'll soon have full control and your helping them.
if a paralegal can be "crazy efficient" at their job that means you need less of them one paralegal can now do the job of several but has the demand of paralegals gone up in tandem? no the amount of paralegals work needed is relatively flat meaning many out if the job
34:58 I think lawyers are some of the first to go. How can any human complete with AI that can process every case study and legal case in history. The attorney AI's will interact with the court AI, and the judge AI, if convicted then the sentencing AI, and this will happen in real time, the moment the police AI robot arrests the human. Then the van AI robot takes them directly to prison.
it shouldn't supplant our agency. It should only enhance it. When our agency is sidelined all we have left (if we're lucky) is diversion and that becomes thin gruel surprisingly quickly
This right here, the three of you, it's priceless.