AI can actually replace workers. The floorplan workers ? no, the managerial jobs that didn't even need to exist and don't do anything of value. That's easy to replace.
@@monad_tcp The problem is that the manager is the one taking decisions on who is replaced, they are not going to replace themselves. There is a reason why bureocracy generally grows faster than the rest of a company, is all about control and creating the conditions to mantain their own job through more bullshit that makes them "necessary."
@@jonasbaine3538 you didnt get the analogy at all... they arent complaining about having to pay for cheese, they are complainng there are people getting trapped, and now those people pay 500 dollars for the trap. it was bad free, its worse paid
I see here that you helped write the API we use at the core of our technology bu-u-u-u-ut, the thing is we're looking for someone with more experience. Our job listing *did* ask for at least six-hundred years of experience in this particular technology after all.
Not enough 1000+ is required for intern position, 2000+ for junior and 10k-50k years for middle level, 1M+ years for senior and 10B+ years for principal
Know what's more ironic, the fact that the entire managerial parasitic class could be actually replaced by ERP software, not even AI, but they have HR protecting themselves. And as their salary is higher than the pawns/wagies, it would be a huge cost saving for the share-holders. Why capitalism don't work ? its like there's two tiers, capitalism for you but not for the. I just stopped worrying and entered the grift of AI, you want AI, I want to make money, so I'm going to make AI for you, no questions asked. I have "wares" to sell.
Not even glorified. Google just convinced my mentally disturbed cousin to unalive my family. What happens if smy cousin does it? Is Google going to pay for the funerals?
Yesterday I watched a video on channel dedicated to AI and helping people to transit to AI era. In the video, they claimed that 4B jobs will be replaced, made obsolete, changed or in some other way "touched" by the AI in the next 5 years. Plot twist: There are 3.4B employed people in the entire world at the moment.
well, if you start replacing people with AI the amount of mistakes and bugs and things breaking will create 700 million new jobs. that's for sure. replacing the jobs ? maybe if they applied to managerial jobs it would work as they don't do much. A robot can perfectly emulate a human at not doing anything of value, just be there and be pretty like HR for example (heck, a plastic doll could do that job probably).
"Artisans won't complain about WLB" and a homeless person sitting next to the ad, this actually makes my blood boil, these CEOs need to chill rn dude or else they might encounter a really unhappy Mario someday
It's not everyone, and there's an open secret that the datacenters powering this are terrible for the environment, so perhaps policy should be applied.
Expect it to happen, a lot. It’s normal. Power gets consolidated too much, and eventually the slaves and peasants revolt. And not by voting for the political party that didn’t win the time before. By making consequences visceral. History is littered with this cycle.
So many people are looking at the AI replacing software engineers as basically “current system, but different supplier”. That’s NOT what will happen. let’s say that AI can replace software devs, so what happens then? We’ll 1 of 2 things, and both will make your job the least of your concerns: 1. The AI companies are now able to completely monopolize the modern information economy because if their tech is really that amazing why would they lease it out when they can just make a new google, new AWS, new TikTok etc. We see a consolidation of wealth and power in the hands of a few players like the world has never seen 2. The tech gets proliferated so that anyone can use it. Be prepared for a wave of malware like you have never seen, captchas will be even more useless, bespoke disinformation campaigns etc. The modern internet will be almost unusable. The only thing that’s stopping either of those now is the cost and limited availability of software engineers ready willing and able to do that stuff. You get rid of that and prepare for all hell to break loose. Jobs will be the least of our concerns at that point.
I'm legit hoping this will push things back into paperwork and brick and mortar stores and institutions. Or hell, maybe laws requiring computers to be slow and datacenters banned for their obvious danger to the environment and humanity.
One thing to consider with AI agents is, that the error rate compounds really quickly. Even if an agent is correct 99% of the time, its only correct 0.99^n correct for multi step tasks. That number goes to 0 really quick if you let devin run without supervision.
And there is no one there to correct it. Self-calling and self-checking isn't a solution there. Developers with experience can foresee a lot more than people think. And one of the main reasons is that experience in 99% isn't collected in the written or spoken form. It's also too abstract. Those experiences are so valid because people's dot-connecting skill of encapsulated, sometimes irrelevant knowledge.
@@Justashortcomment Of course humans also make mistakes. But we generally know when we are out of our depth, can stop to diagnose a problem or ask for help. LLMs aren't self aware like that. The agent can't stop when it doesn't know something, which is why you get hallucinations. And it only takes one of those to invalidate the whole output
@@JustashortcommentYes. Human has common senses and the ability to verify, so when multiple errors stacks together they start to notice something's off. At least competent human can.
Techbros like Sam Altman getting philosophical about labour like a 13 year old having an epiphany always gets me. They're so clueless and wear it proudly for other equally clueless techbros.
they’ve never had to struggle, they’ve never had to learn the struggles of a laborer like you and I. We’re truly being led and exploited by fools and idiots.
I’m a Sr ML Engineer at Oracle and the trend I’m seeing with the larger customers is that they’re shifting towards fine-tuning models and inference. In the next few years we suspect a pretty significant drop in large scale distributed training (or people building foundation models). They aren’t going away, but happening less. To me this means: 1. Companies find AI useful for specialized training or internal knowledge 2. Companies are leveraging special models as part of their product sell 3. MOST IMPORTANT: Companies have realized that there are limitations to what these models can achieve To summarize, it’s not going away, but it is curtailing
I see similar thing in my day job, almost every customer facing business is thinking about internal domain knowledge bots to support operations and learning. But what a lot of people are missing is that we've finally found a way to make natural language a useful human machine interface. Allowing users to interact with products via normal speaking language can be huge if utilized well.
I literally just had my two top executives at my company try to pull a hostile takeover because they think they can replace me and my PhD research with a chatbot. They failed, and the company is still mine.
I watched an interview on doctor mike with dr mike (yes, they have the same name). Its insane what some people think AI is and how well it will perform.
Dont slow down yall, keep learning, keep upskilling, and keep making projects. This AI bubble is going to burst one way or another, and when it does, YOU will be a prime candidate.
Anyone who’s actually used GPT to solve a coding problem that’s novel, understands that it cannot solve novel problems very well. Like even basics of, “I have lost A and List B. Here are their structures.” I need you to do (weird thing).” And it struggles hard. AI workers… sure.
Literally the only thing it's good at is explaining basic concepts. Give it a list of strings and tell it to ID common substrings and count how many times they appear in the list and it can't even do that. Can't even do undergrad python assignments. Complete waste of time
I've found AI to be very helpful in my basic python scripting. But yeah, there's some serious holes. And once it has a logical problem with something, it's stuck and you have to figure it out.
@@eazolan Exactly. For the stuff it konws, it can do really well. For stuff that it doesn't and actually has to deduce, it just falls flat even if the solution is obvious.
Yep. And it's painfully obvious when it tries to over generalize the problem and start applying basic beginner tutorial stuff to your problem because that's the majority of the content it was trained on. It's Like, "No, I don't want you to write a simple 2D knapsack solver. If I did then I would just tell you that chatbot. I want you to follow these exact steps for a problem I've already solved but I think you *might* be able to write faster.... ah forget it!" lol
6:14 I hate the word educate, such a trigger word for me. I always understand it as "we're going to gaslight/psyop you" into our marketing. I would very much prefer people use the word teaching, but that's not what he meant. We're going to "force behavior" (schuab), yea. Educate in a very Pavlovian way.
Yeah, this is happening right in front of me. When company leaders rely purely on AI, and the results don’t meet expectations, instead of understanding that AI isn’t as instant as advertised in app development, they end up blaming the people inputting the prompts. And those entering the prompts often have no idea how programming actually works. These leaders only see the advertisements or others using it successfully, without realizing that those successful cases often have a programming background. Meanwhile, they are lured in by hyperbolic marketing that suggests you can just type something, and your application is ready.
Middle management at companies are trying to adopt AI or hire cheap labor from India to replace their American software engineers. They tried this kind of thing 24 years ago, but the new management has to relearn why it didnt work the first time.
What companies will adopt AI over engineers? My exp is that lots of companies move incredibly slowly when adopting emerging technology. Isn’t it mainly startups and ventures that will give it a go, and either succeed or fail at using it
There are some areas where the AI products work great. But they also seem to be a lot of solutions in the search of a problem. Like are we going to put AI in charge of company bookkeeping? Or tax reporting? Or sales? Nope! I can see helping with marketing though -- provided someone bothers to proofread. But it isn't a solution to every problem and isn't something everyone should be adopting.
Everybody would be stuck on universal basic poverty and you won't be able to move out of your current class by up skilling and you do do nothing but sit around the house with you 1200 a month
People will just find new ways to compete and find meaning in their lives. Humans are hardwired to better themselves and gain resources, so they can increase their chances of finding a decent partner who they can propagate their genes with. The average person living in a first world country already has more conveniences and luxuries than wealthy people from medieval times. I think people will just redefine “class” and create new hierarchies to compete within. “Class” isn’t so much about wealth and resources as it is about competition and dominance.
That would be something that a government would decide. I don’t know of ONE PERSON who would wanna do nothing aall day. I know people that say that a lot, but then they take 2-3 weeks holidays and they’re like “omg I wanna do something, I wanna work on something”. We are naturally inclined into being social, building for one another and helping one another.
The internet has made most people’s life worst. A small amount of people who use it for education and work have prospered from it, not the people on social media and adult sites are worse off
what if most people lose their jobs, won't the general consumerism drop because of this, and thus companies using virtual workers won't run out of money because the sales of their products will drop because people simply won't have money to buy them?
Pretty much. The more you automate the economy the lose money flows in the economy. Total automation would mean a total collapse of the entire economic system as we know it. What would likely follow is an era of revolutions and wars.
Very true what Prime said about change benefitting people unequally. The vast majority of humans need to 1) have work to do, 2) be told mostly what to do, and 3) have extreme difficulty adapting to change past a certain age. A lot of underworked people will cause societal chaos and havoc for the productive ones.
All of these companies portray an ideal world where robots and AI do all the work. Are any of them going to explain how they will make money when nobody has money to buy these products they are so efficiently producing seeing as nobody has jobs in their world?
@@JoeJacksonMhmm it's private equity logic. None of that matters. What matters is you sell off anything you can right now and you die before the bill is due.
You've just discovered one of the many contradictions of capitalism. Congrats. You have a lot to catch up with but the first step is always the most important.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
I wonder if AI will crash the economy. Trading only works if both sides have something to offer. What happens if one side has everything (big corps) and the other side has nothing to offer? What will happen to the government, if their tax revenue falls off, because people can't find jobs?
The capitalist system is fundamentally built on the idea of infinite growth, which relies heavily on consumers. The more consumers there are, and the more they spend on buying goods, the greater the growth. This system requires an ever-increasing number of people spending more money and consuming more products, alongside continuous production. If you take away the consumer, the core driver of the system, it simply cannot function. Few people earning little isn’t going to work, I’m sorry.
There wont even be low paying jobs. You think manual labor companies are not gonna combine boston dynamics robot with AI reasoning? Too expensive? If a robot is 100k it will still be a one time payment instead of paying for actual human salary over the year. Economy will crash so hard its not even funny.
I'm not a programmer but I love your content because you consistently have such real, thoughtful takes on this stuff. Glad I found your channel this year.
I hate the idea of lf this because 5 years ago I was a single dad working 12 hrs 7 days a week driving forklifts couldn't get into tech bc I didn't have time.. I was addicted to vicodin so I could function working over nites and staying up with my kid.. fr there were so many nights I would sleep on the toilet in a warehouse because I only slept for 3 hrs in a 2 day period.. now I got my shit together and finishing my bachelors in computer science this spring at 34... so i really hate the thought of it because I am passionate about it and not to mention all my sacrifices like kiddo time and fiance time..
Don't let this ruin your passion. I am also striving to become a great software engineer. It's heartbreaking for me to realize 90% of this industry is corrupt and doesn't really benefit society at all. All those 100k+ FANG jobs are ruining society. I hope I'll find something in this field that will benefit people instead of these rich assholes taking life out of people who made them rich. Linux is a good example of nobel software project (as are countless of others). I've been a Microsoft user for majority of my life. I am switching to Linux completely and will lower my usage of Microsoft products more as time goes on. The point is there are projects which give power to people and there will be more if we as software engineers don't give up. We must stand against the "bad" guys and start our own projects (or contribute and support existing ones) which will oppose them and give power back to the people who deserve it.
Sam Altmann is not «the creator of the future». He is the guy with stock options when his company went « ok this actually works a lot better than we thought it would, let’s release it as a product»
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾
I do think AI, robotics, and full automation is the better future. The problem is there's no economic/social plan for the transition other than the usual "just die during the next winter please". What good is a promised paradise if you and your family will not survive the transition.
What's good about it? All these hopes about bright techno future is basically a communism, or Christianity 2.0 with heaven after life. It's beneficial to see the tech through how it changes the power balance, what does it make things less valuable/more valuable.
Looking at the trajectory of AI since the release of ChatGPT I think I can predict with some confidence that it will not necessarily do what Sam is saying. Its going to be just another tool to widen the inequality gap in the world. I mean which of these rich CEOs is gonna wake up one day and say, okay I have enough money now, lets slash all the prices! Never gonna happen
OpenAI is an inherently unprofitable company. They loose money on everything and the objective to hyperscale definitely doesn't help this very huge issue.
I just dont see how this differs compared to factory workers who have lost their jobs to automation since the beginning of industrialism, they wept and rioted similarly when their jobs were taken, but capitalism doesnt give two shits about the feelings of the people. Its all about the $$$, i dont think making emotional arguments as to why AI (software engineers) is bad does anything, because at the end of the day, big corpa will take the cheapest option available on the market.
There are other areas like para legals, people approving bank loans or insurance claims etc that are much more directly threatened than software engineers.
I'd replace HR departments first. I think they are the easiest to replace by AI. You only need AI for speech, because they only work by fixed rules predetermined by the company. The rest is standard sentences to throw out.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
The coding is the easiest part of software dev. To learn the CONTEXT of the whole company and the history of why a decision was made, why it was rejected, what is for later, what needs to be done now, where is tech debt that was deliberate, where is tech debt bc nobody knew better at the time .... all these things are a huge part of dev knowledge and they are stored in 1000+ miro boards which have been deprecated in favor of 1000+ notion pages of which 50% are deprecated, 1000+ folders in ms-cloud and google cloud, 100 Grafana boards and 50+ repos of which 30% are deprecated. Good luck figuring out what's going on .
Fine, we have a handful of people to guide the AI and everyone else is gone. Nobody thinks ai will replace everyone, but most people will find themselves out of a job in the next 10 to 20 years
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾
I 100% agree with the take take of this video. I also find it very interesting, that enhancements in tech is often associated with raising the standard of living and the generel well being of the population, while the AI hype does exact opposite and potentialy costs a lot of people their base of living, offering them no solution
I can already picture the CEO arguing with AI : No, no, no! I want the code to do this! What do you mean you don't understand? You're AI! Fix it! Noooo, not like that!
A lot of leaders have a hard time admitting mistakes.... even to themselves. So what is likely to happen is someone else will be tasked with making the AI do things. Then that someone else will get all the blame for the CEO's grand plan not working.
ceo: "It's a goddamn simple task! Just draw seven red lines all perpendicular to each other some with green ink and some with transparent! A monkey could do it!!"
I think a real interesting thing about the current LLM situation, is that we really just got a new fancy search engine. Even reviewing your CV or whatever it might be, is the equal of putting in part of your CV in google and ask if that is how it should look and what suggestions it has. Again, the same is true for the programming part. So it just removes that abstraction of us having to formulize a generic question, so that we can now just supply the direct question, and it hash out the difference to the generic question/questions we would have asked google. And of course it also condense it back into our original question. But it is not more, not less. Anyone in Software Engineering, Applied Math, Statistics and Data science(can we stop with the ML Engineer..) knows this. But I am excited that it potential open the way for traditional ML which has a lot of applications in decision support systems that has until now, been completely overlooked, even through we had the computing power for it for years.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
The problem with shorting is that a lot of badly run companies can have high stock prices for a long time. The stock may even go up quickly. Yes, eventually the market will catch on but that eventually can be quite long. But I'd stay away from getting a job at these companies, as your job won't last very long and would likely be a very unpleasant experience.
When cars came along, so many manure shovelers in cities lost their jobs. New technologies have gotten rid of jobs in the past. Tone dialing lost a lot of operators their jobs. Tape operators, lamp lighters, knocker ups. There's a long list.
It’s not just the stigma of being jobless, that’s part of the dynamic. A lot of human beings derive value and enjoyment from working. It’s not just this stigma of judgement. You sound like an “anti work” person from Reddit in which case you’re a minority because most human beings derive enjoyment out of working and being active whether they realize it or not
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
Prime quoting "The Abolition of Man" by CS Lewis I love it. Other great books on this subject include "The Managerial Revolution" by James Burnham and "On Power: The Natural History of its Growth" by Bertrand de Jouvenel.
7:20 You can acknowledge the incredible benefits the internet has granted us while at the same time recognizing that certain companies have made use of it in a way that is VERY detrimental to both individuals and society as a whole.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝🔝
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝🔝
I think an important sci-fi concept from Star Trek TNG that's been somewhat forgotten is the idea of people still choosing to work in whatever field they're passionate about. TNG had machines that could make you a good drink. It also still had bartenders.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
who needs customers, replace them with AI too ? heck, lets go even deeper, install an NVidia datacenter near the NYSE and make an entire virtual economy of AI so the people in Wall Street be happy with their fumbled numbers, meanwhile we the actual humans go create our barter economy outside the monetary system. If we need an actual stock exchange that's not corrupt we could use the economy of EvE online as an actual economy instead of NYSE, its less corrupt. Occupy WallStreet 2.0
Not all employees, just the good jobs that allow socioeconomic mobility. You can still work a minimum wage job and then give your money to the elite class.
@@zoeherriot but that "everything" must depend on the people who purchase that everything. And how are they going to make money, if there aren't people, who could purchase that "everything"?
If anything, these AI solutions could probably replace some junior level devs now, which in turn creates a very strange imbalance that may lead to a significant decline in senior devs later on. People seem to be vastly overestimating how much knowledge in general a junior dev actually has.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
Software Engineers will be even more needed. The combination with AI is great and will open a lot of posiibilities. I love to use Claude to help you organize your thinking process, and Copilot to complete code or write repetetive code (such as test scripts or code that is not easily to avoid a level of hardtype).
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
You earned a sub from me. Its refreshing to hear someone thats in touch with the population as a whole and not just the wealthy midddle and upper class.
In death stranding, everything becomes automated except for “porters”(delivery men), and buried in the lore it’s revealed that at one point it was automated but was a failure because people like to work so they went back to the porter system.
And it's kinda slow. But they made amazing progress in the last two years. I mean there have been humanoid robots showcased for decades but couldn't do a real, useful job until now. Boost the speed, add some basic intelligence and you have a physical agent ready to be deployed. If it can replace all the repetitive, boring manual labor then it's a great success already.
Same thing happened in the early 2000's when non-tech CEOs and CFOs though they could outsource dev to cheap labour pools overseas and fire all their high paid but experienced home grown devs...
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
Maybe since most companies decision makers and middle management can be replaced by AI, surely we will be working for AI and fetching them buckets of electrons from the well?
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
There's no doubt in my mind that before things will get better we will go through possibly years and years of dystopian poverty and a shift towards ultra poor and ultra rich with a wide chasm between two groups. There is no job which won't get automated with the use of robots. Then people will have nothing to do and they will lose their sense of self / belonging / purpose and we will see a huge spike in mental issues as well. It's also possible that things will never get better. Once the rich will automate everything what's the point in keeping the rest of humanity around ? They won't really need consumers anymore.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝🔝
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝🔝
Remember the Don't be Evil phrase from Google. While a company can say it wants to enhance your life, it likely will take you out of the equation the moment they get the opportunity. Try to stay in the opensource domain.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
AI replaces MANAGERS, not workers! Anybody in a productive job is ranked higher than anyone who is overhead in AI models. In an AI run company, everyone is on task and every task is verified, collected, part of your internal record of progress or lack thereof. But if you're sorting potato chips, yeah, replaced.
I remember a time when making AI meant people are free from doing the boring, dangerous, and monotonous tasks, and we're left to pursue arts and such. Literally, what is happening right now, is that art is being made by AI, and people have to do things the AI cannot do yet (physical/boring work).
The layman that pays for the ai to make a software, then the output software doesn't works correctly. Now the layman has no idea what is the problem and how to ask the AI to fix, or he doesn't even know the problem is there and it only seems to work. Great software.
People look at AI and think: "it's like industrial revolution but for the thinking jobs". Or "AI already got so much better than it used to be". But people forget - past does not determine nor predict the future. We can make some guesses at best
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝🔝
as someone with ADHD already, these suggested line completions (in my case, salesforce codegen) are pretty close to what i'm looking to do about 10% of the time, and the other 90% they're distracting and only serve to destroy my fragile train of thought.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
Man, this is a great speech! Even though we don't know if the Internet is good for society. So, we need to take with a grain of salt what these CEOs said
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
I dont know man. I think this is really going to happen. Not sure when, but the technology keeps improving. We are all biased as developers. It's def going to be interesting to see how it all plays out.
"It keeps improving" says who? There is absolutely no guarantee LLMs will magically get better forever. If you simply look at the billions spent and the fact they've already digested the entire internet without achieving AGI, you can see most models are capping out. Plus most of the good AIs. Are Actual Indians. Your coding request takes 5+min to process because Sajit has to pick up the ticket and do the work. Multiple companies have been busted for this. Generative AI does fine when all you want is slop art to sell products to uninformed consumers. It's an absolute travesty in any field that needs precision, reliability, auditability, legal responsibility...
I think UBI can help with the job thing because people can take more entrepreneurial risks opening small businesses and employing each other in small local businesses providing local services. Though crowd sharing apps might take this over.
The "LLMs are just next word predictors" argument doesn't really hold water. What if _we_ are just next word predictors? How do we know that we are not? _We_ can't produce more than one word simultaneously (for the most part).
Humans aren’t just “next word predictors.” The Beginning of Infinity by David Deutsch has a good explanation - great book. Long story short, the creation of explanatory knowledge requires conjecture and criticism, not prediction, and is directed at solving a problem, not prediction. If a problem hasn’t been solved before, how can there be appropriate training data to inform the prediction? I like the example of the invention of 0 as a placeholder in the Hindu place-value number system. Without a placeholder symbol representing “no quantity,” a place-value number system has ambiguity. Eg, the string “1 23” is ambiguous: does it represent “one and twenty three” or “one thousand and twenty three”? 0 seems to have been introduced to solve this issue, and in doing so unlocked the concept of 0 as a mathematical object that could be manipulated algebraicly which enabled further advances in mathematics, such as the idea of negative numbers (numbers less than 0). At some point, the zero symbol and the idea literally never existed. The Romans and Greeks didn’t have a 0 equivalent in their numeral systems. So when the first person did the “0 as a placeholder in overflow columns” trick, they had no past data were they “extrapolating” from to solve that problem. There are lots of examples in science too of people creating new ideas, imagining new entities, and creatively conjecturing alternative models of the world that didn’t exist in any data set before they solved the problem they were working on. Humans are not merely predicting next words; we are solving problems, many of which have not been solved before or demand solutions that are original, and thus can’t be encoded in data sets representing our existing knowledge.
Quality assurance with AI output is still a huge problem, which cannot be solved completely, because AIs can only be trained on existing labeled (in this context: validated) data. And for that you need humans to do the data validation so that an AI may learn it, so that it may determine whether the output is good. Right now all LLMs make so many mistakes. Therefore, LLMs are like a very good "sed" or "awk" command and I often use them to generate syntax based on templates (e.g. classes based on JSON structure). But they will fail horribly when the task is based on semantic implications rather than explicit syntactical tasks.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
Please add the Devin I wanna more money in my wallet when the companies hire me to fix it. IMAGE you have a fucking LEGACY in JAVA Struct lol it will be amazing I wanna see it
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
I think the impact of AI on developer jobs is more nuanced than a simple replacement. The current infrastructure isn't optimized for AI, which limits its immediate capabilities. However, as the infrastructure evolves, AI will undoubtedly automate many aspects of development, potentially making some current developer tasks redundant.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
Devin chaos-ing through the Prime Again’s infra instead of push to master was top comedy and could be the vindication of older GPT3 girlfriends people needed
It's another tool in humanity's toolbox. Tools can harm or could be put to good use, but fundamentally it's better to have more options as long as you are not confused about this options implications= ai/ml good, but could be harmful if misunderstood and 'AI replacement of engineer' great example of misleading to take advantage of misunderstanding.
I totally agree with prime on the fact that these tech bros are so certain about what will be the outcome of the AI revolution. There is definitely a shift in power and fewer of us individually will have a say in things. I’m not saying world was 100x better in the olden days, industrial revolution was critical for humans but still it’s an ongoing experiment of not clear outcomes.
One of the top leaders in the state government I work in does essentially nothing but go completely ham in meetings about AI. He said we have a “pro-AI governor” as if it’s smasmorshin or something. But in the same exact meetings where he breathlessly extols the unbelievable sea change this is ushering in, he says “we just need to learn how to consume it and what it can do for us” Hi Cart, it appears you have left horse behind.
Driving people towards despair is the worst thing about this stuff and the way it's being marketed. People want to actually do things and see the results of themselves doing things, but if the marketing around this stuff makes it seem like people are just obsolete then even if it's not true, plenty of people will fall to despair and do unmentionable things.
People ask me "aren't you afraid AI might take your job", i tell them "my jobs to automate your job, so if my jobs automated what's going to happen to your job?"
"AI will replace workers" said increasingly nervous AI salesman
AI can actually replace workers.
The floorplan workers ? no, the managerial jobs that didn't even need to exist and don't do anything of value. That's easy to replace.
@@monad_tcpfloor plan workers will be replaced eventually as well. Just like bank tellers were replaced by ATMs.
@@monad_tcpmany of those could be replaced by air
@@boneless-yeet-1127 that's currently the reason for grocery store theft globally
@@monad_tcp The problem is that the manager is the one taking decisions on who is replaced, they are not going to replace themselves. There is a reason why bureocracy generally grows faster than the rest of a company, is all about control and creating the conditions to mantain their own job through more bullshit that makes them "necessary."
NGL I want them to try rn so the companies will be afraid to try it again
there will be another wave of layoff to make that happen
@@anbiabohlam5468 of course not. They will phase them in. Why take the risk? Try a few AI employees, conclude that they suck and then that's it.
Based
Failure hasn't stopped them from trying offshoring again and again no matter how many times the results are disastrous.
😮 That's genius!
Cheese in the mousetrap used to be free; now it costs $500 a month.
Water used to be free too
@@jonasbaine3538 you didnt get the analogy at all... they arent complaining about having to pay for cheese, they are complainng there are people getting trapped, and now those people pay 500 dollars for the trap. it was bad free, its worse paid
@@jonasbaine3538 Now it comes in a handy-dandy plastic cover from Nestle, at a cost!
@@jonasbaine3538 water was never free, the land used to belong to royalty. in order to use it you had to become a serf or something.
@@khhnator my water is for free...comes from the ground and I take it daily with a bucket.
Time to add 10 years of experience in the resume doing "Fixing bugs created by AI Workers"
only 10 years? Are you a junior? Nah we need 50+ years of experience minimum then we will consider you.
150years (total ai compute time) experience fixing ai bugs
Dont worry yu will no need resume anymore
I see here that you helped write the API we use at the core of our technology bu-u-u-u-ut, the thing is we're looking for someone with more experience. Our job listing *did* ask for at least six-hundred years of experience in this particular technology after all.
Not enough 1000+ is required for intern position, 2000+ for junior and 10k-50k years for middle level, 1M+ years for senior and 10B+ years for principal
They released a glorified chatbot and the entire parasitic class decided it was time to start the depop, lmao.
Know what's more ironic, the fact that the entire managerial parasitic class could be actually replaced by ERP software, not even AI, but they have HR protecting themselves. And as their salary is higher than the pawns/wagies, it would be a huge cost saving for the share-holders. Why capitalism don't work ? its like there's two tiers, capitalism for you but not for the.
I just stopped worrying and entered the grift of AI, you want AI, I want to make money, so I'm going to make AI for you, no questions asked. I have "wares" to sell.
Didn’t work during Covid, the ww3 isn’t here, they’re desperate lol
Beautiful.
Not even glorified. Google just convinced my mentally disturbed cousin to unalive my family. What happens if smy cousin does it? Is Google going to pay for the funerals?
@@andrewhooper7603 RUclips deleting replies to this comment proves you right
Yesterday I watched a video on channel dedicated to AI and helping people to transit to AI era. In the video, they claimed that 4B jobs will be replaced, made obsolete, changed or in some other way "touched" by the AI in the next 5 years. Plot twist: There are 3.4B employed people in the entire world at the moment.
😂
yea bro, but they accounted for the job growth in the next 5 years.
@@tedchirvasiu So, job market growth and replacing billions of people with AI will be happening at the same time? Makes perfect sense!
Guess an ai came up with these statistics 😆
well, if you start replacing people with AI the amount of mistakes and bugs and things breaking will create 700 million new jobs. that's for sure. replacing the jobs ? maybe if they applied to managerial jobs it would work as they don't do much. A robot can perfectly emulate a human at not doing anything of value, just be there and be pretty like HR for example (heck, a plastic doll could do that job probably).
"Artisans won't complain about WLB" and a homeless person sitting next to the ad, this actually makes my blood boil, these CEOs need to chill rn dude or else they might encounter a really unhappy Mario someday
You mean Luigi
Luigi is already in prison
It's not everyone, and there's an open secret that the datacenters powering this are terrible for the environment, so perhaps policy should be applied.
Expect it to happen, a lot. It’s normal.
Power gets consolidated too much, and eventually the slaves and peasants revolt.
And not by voting for the political party that didn’t win the time before. By making consequences visceral.
History is littered with this cycle.
do it.
So many people are looking at the AI replacing software engineers as basically “current system, but different supplier”. That’s NOT what will happen.
let’s say that AI can replace software devs, so what happens then? We’ll 1 of 2 things, and both will make your job the least of your concerns:
1. The AI companies are now able to completely monopolize the modern information economy because if their tech is really that amazing why would they lease it out when they can just make a new google, new AWS, new TikTok etc. We see a consolidation of wealth and power in the hands of a few players like the world has never seen
2. The tech gets proliferated so that anyone can use it. Be prepared for a wave of malware like you have never seen, captchas will be even more useless, bespoke disinformation campaigns etc. The modern internet will be almost unusable.
The only thing that’s stopping either of those now is the cost and limited availability of software engineers ready willing and able to do that stuff. You get rid of that and prepare for all hell to break loose. Jobs will be the least of our concerns at that point.
a DataKrash ... yaaay...
I'm legit hoping this will push things back into paperwork and brick and mortar stores and institutions.
Or hell, maybe laws requiring computers to be slow and datacenters banned for their obvious danger to the environment and humanity.
@@krunkle5136 have you been reading Dune or something?
@@user-sl6gn1ss8p no, lol. Always wanted the time to get into it.
Facts, we'll have to wait for Ai first. this is just a glorified chatbot that gives you the 80% most common google answer.
"Hey , boss I want some free time to spend with my family"
"Fine. You are fired"
It is free time
One thing to consider with AI agents is, that the error rate compounds really quickly. Even if an agent is correct 99% of the time, its only correct 0.99^n correct for multi step tasks. That number goes to 0 really quick if you let devin run without supervision.
And there is no one there to correct it. Self-calling and self-checking isn't a solution there. Developers with experience can foresee a lot more than people think. And one of the main reasons is that experience in 99% isn't collected in the written or spoken form. It's also too abstract. Those experiences are so valid because people's dot-connecting skill of encapsulated, sometimes irrelevant knowledge.
Bless you. Can you convince the CEOs?
Are humans immune to this piece of math?
@@Justashortcomment Of course humans also make mistakes. But we generally know when we are out of our depth, can stop to diagnose a problem or ask for help. LLMs aren't self aware like that. The agent can't stop when it doesn't know something, which is why you get hallucinations. And it only takes one of those to invalidate the whole output
@@JustashortcommentYes. Human has common senses and the ability to verify, so when multiple errors stacks together they start to notice something's off. At least competent human can.
Techbros like Sam Altman getting philosophical about labour like a 13 year old having an epiphany always gets me. They're so clueless and wear it proudly for other equally clueless techbros.
They are just making money, and living in their own bubble.
they’ve never had to struggle, they’ve never had to learn the struggles of a laborer like you and I. We’re truly being led and exploited by fools and idiots.
I’m a Sr ML Engineer at Oracle and the trend I’m seeing with the larger customers is that they’re shifting towards fine-tuning models and inference. In the next few years we suspect a pretty significant drop in large scale distributed training (or people building foundation models). They aren’t going away, but happening less.
To me this means:
1. Companies find AI useful for specialized training or internal knowledge
2. Companies are leveraging special models as part of their product sell
3. MOST IMPORTANT: Companies have realized that there are limitations to what these models can achieve
To summarize, it’s not going away, but it is curtailing
Thanks for the reporting from inside!
Always hated Oracle as a company (same as Microsoft) but not the engineers there!
I see similar thing in my day job, almost every customer facing business is thinking about internal domain knowledge bots to support operations and learning. But what a lot of people are missing is that we've finally found a way to make natural language a useful human machine interface. Allowing users to interact with products via normal speaking language can be huge if utilized well.
@@Frater1337yeah this is where I see a lot of value too. I’m seeing companies recognize use cases like this and I think it’s promising
@@aravindpallippara1577appreciate that! The folks I work with are good peoples. The company is just a huge machine, but the work is interesting.
@@Frater1337 Like an interactive wiki?
I literally just had my two top executives at my company try to pull a hostile takeover because they think they can replace me and my PhD research with a chatbot. They failed, and the company is still mine.
Bro AI is gonna solve all of physics apparently. It can't even count repeated letters in a string, but it's cooking up its own Unified field theory 🤣
And they're on the path to replacement right?
how did you get them to back down?
I watched an interview on doctor mike with dr mike (yes, they have the same name). Its insane what some people think AI is and how well it will perform.
@@zbyniew AI has always been able to count repeated letters in a string from day 1.
"Write a program that counts "? in a string.
Dont slow down yall, keep learning, keep upskilling, and keep making projects. This AI bubble is going to burst one way or another, and when it does, YOU will be a prime candidate.
It takes time for bubble to burst. It may take years, maybe 10 years.
I agree with you, but this sounds like the hardest compium I've ever read lol
It won’t burst… it will normally evolve, slowly and steadily. Eventually it will help us… not replace. well maybe some simple jobs.
When the venture capital dries up, these companies have no real product. They are doomed to fail. Keep learning and coding.
I feel like this is a W take. Or will be for a while anyway
Anyone who’s actually used GPT to solve a coding problem that’s novel, understands that it cannot solve novel problems very well. Like even basics of, “I have lost A and List B. Here are their structures.” I need you to do (weird thing).” And it struggles hard. AI workers… sure.
It's a great tool for wasting your time by being in an illusion that you are getting things done
Literally the only thing it's good at is explaining basic concepts. Give it a list of strings and tell it to ID common substrings and count how many times they appear in the list and it can't even do that. Can't even do undergrad python assignments. Complete waste of time
I've found AI to be very helpful in my basic python scripting. But yeah, there's some serious holes. And once it has a logical problem with something, it's stuck and you have to figure it out.
@@eazolan Exactly. For the stuff it konws, it can do really well. For stuff that it doesn't and actually has to deduce, it just falls flat even if the solution is obvious.
Yep. And it's painfully obvious when it tries to over generalize the problem and start applying basic beginner tutorial stuff to your problem because that's the majority of the content it was trained on. It's Like, "No, I don't want you to write a simple 2D knapsack solver. If I did then I would just tell you that chatbot. I want you to follow these exact steps for a problem I've already solved but I think you *might* be able to write faster.... ah forget it!" lol
6:14 I hate the word educate, such a trigger word for me. I always understand it as "we're going to gaslight/psyop you" into our marketing. I would very much prefer people use the word teaching, but that's not what he meant. We're going to "force behavior" (schuab), yea. Educate in a very Pavlovian way.
What's "schuab"?
@@monad_tcp agreed. I wish educate was a neutral term, but corpos and activists ruined it.
@@krunkle5136 Almost as if people are responsible for their words? Thought you lot didnt like that?
that's exactly what is always meant by "education" when it comes from top-down instittutions
Time to add "AI clean up and decontamination" to your CV to be ready for the inevitable fallout.
Yeah, this is happening right in front of me. When company leaders rely purely on AI, and the results don’t meet expectations, instead of understanding that AI isn’t as instant as advertised in app development, they end up blaming the people inputting the prompts. And those entering the prompts often have no idea how programming actually works.
These leaders only see the advertisements or others using it successfully, without realizing that those successful cases often have a programming background. Meanwhile, they are lured in by hyperbolic marketing that suggests you can just type something, and your application is ready.
Middle management at companies are trying to adopt AI or hire cheap labor from India to replace their American software engineers. They tried this kind of thing 24 years ago, but the new management has to relearn why it didnt work the first time.
Right on.
Oh man. Don't get me started on cheap Indian engineers.
What companies will adopt AI over engineers?
My exp is that lots of companies move incredibly slowly when adopting emerging technology.
Isn’t it mainly startups and ventures that will give it a go, and either succeed or fail at using it
There are some areas where the AI products work great. But they also seem to be a lot of solutions in the search of a problem. Like are we going to put AI in charge of company bookkeeping? Or tax reporting? Or sales? Nope! I can see helping with marketing though -- provided someone bothers to proofread. But it isn't a solution to every problem and isn't something everyone should be adopting.
Everybody would be stuck on universal basic poverty and you won't be able to move out of your current class by up skilling and you do do nothing but sit around the house with you 1200 a month
I hate this future
Exactly, I want to leave the USA before its too late lol.
People will just find new ways to compete and find meaning in their lives. Humans are hardwired to better themselves and gain resources, so they can increase their chances of finding a decent partner who they can propagate their genes with. The average person living in a first world country already has more conveniences and luxuries than wealthy people from medieval times. I think people will just redefine “class” and create new hierarchies to compete within. “Class” isn’t so much about wealth and resources as it is about competition and dominance.
That would be something that a government would decide. I don’t know of ONE PERSON who would wanna do nothing aall day. I know people that say that a lot, but then they take 2-3 weeks holidays and they’re like “omg I wanna do something, I wanna work on something”. We are naturally inclined into being social, building for one another and helping one another.
That's the optimistic scenario IMO.
The internet has made most people’s life worst. A small amount of people who use it for education and work have prospered from it, not the people on social media and adult sites are worse off
Fr, your approach makes a lot of sense, i have seen a lot of good examples of that.
what if most people lose their jobs, won't the general consumerism drop because of this, and thus companies using virtual workers won't run out of money because the sales of their products will drop because people simply won't have money to buy them?
Pretty much. The more you automate the economy the lose money flows in the economy. Total automation would mean a total collapse of the entire economic system as we know it.
What would likely follow is an era of revolutions and wars.
Very true what Prime said about change benefitting people unequally. The vast majority of humans need to 1) have work to do, 2) be told mostly what to do, and 3) have extreme difficulty adapting to change past a certain age. A lot of underworked people will cause societal chaos and havoc for the productive ones.
All of these companies portray an ideal world where robots and AI do all the work. Are any of them going to explain how they will make money when nobody has money to buy these products they are so efficiently producing seeing as nobody has jobs in their world?
Exactly. If no one has work, how will they trade for goods? If no one trades for goods, how will new goods be created?
Easy, communism.
@@JoeJacksonMhmm it's private equity logic. None of that matters. What matters is you sell off anything you can right now and you die before the bill is due.
They will be the only products your UBI is allowed to purchase.
They will profit because they have a monopoly on work itself
You've just discovered one of the many contradictions of capitalism. Congrats. You have a lot to catch up with but the first step is always the most important.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
I wonder if AI will crash the economy. Trading only works if both sides have something to offer.
What happens if one side has everything (big corps) and the other side has nothing to offer? What will happen to the government, if their tax revenue falls off, because people can't find jobs?
You can push one side to give, without giving anything back.
@@dmitriyrasskazov8858 It's more likely that big corps will be gods among men, deciding who will earn lil money and who not
There would still be jobs, but only either the lowest paying jobs or top level executive jobs with no socioeconomic mobility.
The capitalist system is fundamentally built on the idea of infinite growth, which relies heavily on consumers. The more consumers there are, and the more they spend on buying goods, the greater the growth. This system requires an ever-increasing number of people spending more money and consuming more products, alongside continuous production. If you take away the consumer, the core driver of the system, it simply cannot function. Few people earning little isn’t going to work, I’m sorry.
There wont even be low paying jobs. You think manual labor companies are not gonna combine boston dynamics robot with AI reasoning? Too expensive? If a robot is 100k it will still be a one time payment instead of paying for actual human salary over the year. Economy will crash so hard its not even funny.
I'm not a programmer but I love your content because you consistently have such real, thoughtful takes on this stuff. Glad I found your channel this year.
Aww, you probably say that about all the people you agree with..
You are not a programmer - Like why do you even exist?!
@@krsma86 maybe he's just starting
9:20 Prime unintentionally describes the plot of Blade Runner 2049 and the Wallace Corporation
Wallace... Gromit... Corporation?
@@electrinity yea
I hate the idea of lf this because 5 years ago I was a single dad working 12 hrs 7 days a week driving forklifts couldn't get into tech bc I didn't have time.. I was addicted to vicodin so I could function working over nites and staying up with my kid.. fr there were so many nights I would sleep on the toilet in a warehouse because I only slept for 3 hrs in a 2 day period.. now I got my shit together and finishing my bachelors in computer science this spring at 34... so i really hate the thought of it because I am passionate about it and not to mention all my sacrifices like kiddo time and fiance time..
Don't let this ruin your passion. I am also striving to become a great software engineer. It's heartbreaking for me to realize 90% of this industry is corrupt and doesn't really benefit society at all. All those 100k+ FANG jobs are ruining society. I hope I'll find something in this field that will benefit people instead of these rich assholes taking life out of people who made them rich. Linux is a good example of nobel software project (as are countless of others). I've been a Microsoft user for majority of my life. I am switching to Linux completely and will lower my usage of Microsoft products more as time goes on. The point is there are projects which give power to people and there will be more if we as software engineers don't give up. We must stand against the "bad" guys and start our own projects (or contribute and support existing ones) which will oppose them and give power back to the people who deserve it.
Sam Altmann is not «the creator of the future». He is the guy with stock options when his company went « ok this actually works a lot better than we thought it would, let’s release it as a product»
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾
I do think AI, robotics, and full automation is the better future. The problem is there's no economic/social plan for the transition other than the usual "just die during the next winter please". What good is a promised paradise if you and your family will not survive the transition.
What's good about it? All these hopes about bright techno future is basically a communism, or Christianity 2.0 with heaven after life. It's beneficial to see the tech through how it changes the power balance, what does it make things less valuable/more valuable.
This 100%
Sam GPT saying “yeah we will [indirectly] kill a bunch of people, but this is gonna be a good thing” makes me uncomfortable.
Looking at the trajectory of AI since the release of ChatGPT I think I can predict with some confidence that it will not necessarily do what Sam is saying. Its going to be just another tool to widen the inequality gap in the world. I mean which of these rich CEOs is gonna wake up one day and say, okay I have enough money now, lets slash all the prices! Never gonna happen
OpenAI is an inherently unprofitable company. They loose money on everything and the objective to hyperscale definitely doesn't help this very huge issue.
I just dont see how this differs compared to factory workers who have lost their jobs to automation since the beginning of industrialism, they wept and rioted similarly when their jobs were taken, but capitalism doesnt give two shits about the feelings of the people. Its all about the $$$, i dont think making emotional arguments as to why AI (software engineers) is bad does anything, because at the end of the day, big corpa will take the cheapest option available on the market.
10:08 onward my man's cooking towards realizing that every war is a class war
There are other areas like para legals, people approving bank loans or insurance claims etc that are much more directly threatened than software engineers.
I'd replace HR departments first. I think they are the easiest to replace by AI. You only need AI for speech, because they only work by fixed rules predetermined by the company. The rest is standard sentences to throw out.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
The coding is the easiest part of software dev. To learn the CONTEXT of the whole company and the history of why a decision was made, why it was rejected, what is for later, what needs to be done now, where is tech debt that was deliberate, where is tech debt bc nobody knew better at the time ....
all these things are a huge part of dev knowledge and they are stored in 1000+ miro boards which have been deprecated in favor of 1000+ notion pages of which 50% are deprecated, 1000+ folders in ms-cloud and google cloud, 100 Grafana boards and 50+ repos of which 30% are deprecated.
Good luck figuring out what's going on .
Fine, we have a handful of people to guide the AI and everyone else is gone.
Nobody thinks ai will replace everyone, but most people will find themselves out of a job in the next 10 to 20 years
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾
I 100% agree with the take take of this video. I also find it very interesting, that enhancements in tech is often associated with raising the standard of living and the generel well being of the population, while the AI hype does exact opposite and potentialy costs a lot of people their base of living, offering them no solution
I can already picture the CEO arguing with AI : No, no, no! I want the code to do this! What do you mean you don't understand? You're AI! Fix it! Noooo, not like that!
🤣 I can see it too. “dO yOu EvEn KNoW wHo I aM!?!?”
A lot of leaders have a hard time admitting mistakes.... even to themselves. So what is likely to happen is someone else will be tasked with making the AI do things. Then that someone else will get all the blame for the CEO's grand plan not working.
ceo: "It's a goddamn simple task! Just draw seven red lines all perpendicular to each other some with green ink and some with transparent! A monkey could do it!!"
@@attribute-4677 "I'mma fire you😠"
I think a real interesting thing about the current LLM situation, is that we really just got a new fancy search engine. Even reviewing your CV or whatever it might be, is the equal of putting in part of your CV in google and ask if that is how it should look and what suggestions it has. Again, the same is true for the programming part. So it just removes that abstraction of us having to formulize a generic question, so that we can now just supply the direct question, and it hash out the difference to the generic question/questions we would have asked google. And of course it also condense it back into our original question. But it is not more, not less. Anyone in Software Engineering, Applied Math, Statistics and Data science(can we stop with the ML Engineer..) knows this.
But I am excited that it potential open the way for traditional ML which has a lot of applications in decision support systems that has until now, been completely overlooked, even through we had the computing power for it for years.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
We need a list of companies hiring "AI Employees" so what we can short them. Should be the short of the century.
The problem with shorting is that a lot of badly run companies can have high stock prices for a long time. The stock may even go up quickly. Yes, eventually the market will catch on but that eventually can be quite long. But I'd stay away from getting a job at these companies, as your job won't last very long and would likely be a very unpleasant experience.
When cars came along, so many manure shovelers in cities lost their jobs. New technologies have gotten rid of jobs in the past. Tone dialing lost a lot of operators their jobs. Tape operators, lamp lighters, knocker ups. There's a long list.
People stop working all the time. It is called pension. It is the stigma about being jobless that causes mental health issues not the lack of jobs.
It’s not just the stigma of being jobless, that’s part of the dynamic. A lot of human beings derive value and enjoyment from working. It’s not just this stigma of judgement. You sound like an “anti work” person from Reddit in which case you’re a minority because most human beings derive enjoyment out of working and being active whether they realize it or not
@@hovat When you admit that you're a Reddit lurker yourself all credibility goes out the window
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾
Nothing is being replaced. It's the old "buy my shovels to mine for gold" all over again.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
Prime quoting "The Abolition of Man" by CS Lewis I love it. Other great books on this subject include "The Managerial Revolution" by James Burnham and "On Power: The Natural History of its Growth" by Bertrand de Jouvenel.
7:20
You can acknowledge the incredible benefits the internet has granted us while at the same time recognizing that certain companies have made use of it in a way that is VERY detrimental to both individuals and society as a whole.
"We don't care about the machines and factories making stuff." This guy needs to leave the silicon valley bubble for a bit.
Then, there's going to be a high demand for engineers JUST to fix all the errors Ai added to the code
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
Hilarious how many people resist AI taking our 💩 jobs 😂 Cannot wait for the day we work a 2 day week and AI does it all!
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝🔝
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝🔝
I think an important sci-fi concept from Star Trek TNG that's been somewhat forgotten is the idea of people still choosing to work in whatever field they're passionate about. TNG had machines that could make you a good drink. It also still had bartenders.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
Pretty interesting, ask the CEo who will buy their product(s) if all employees are fired and replaced with AI.
who needs customers, replace them with AI too ?
heck, lets go even deeper, install an NVidia datacenter near the NYSE and make an entire virtual economy of AI so the people in Wall Street be happy with their fumbled numbers, meanwhile we the actual humans go create our barter economy outside the monetary system.
If we need an actual stock exchange that's not corrupt we could use the economy of EvE online as an actual economy instead of NYSE, its less corrupt.
Occupy WallStreet 2.0
They don’t care - they will own everything they need by the time it reaches that point.
Not all employees, just the good jobs that allow socioeconomic mobility. You can still work a minimum wage job and then give your money to the elite class.
@@monad_tcp How are they gonna make money, if there aren't people, who would buy the goods?
@@zoeherriot but that "everything" must depend on the people who purchase that everything. And how are they going to make money, if there aren't people, who could purchase that "everything"?
Sammy boy is pumping AI for last 4 years. It is getting boring.
If anything, these AI solutions could probably replace some junior level devs now, which in turn creates a very strange imbalance that may lead to a significant decline in senior devs later on. People seem to be vastly overestimating how much knowledge in general a junior dev actually has.
Nothing is being replaced. It's the old "selling shovels is more profitable than mining for gold" all over again.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
Software Engineers will be even more needed. The combination with AI is great and will open a lot of posiibilities. I love to use Claude to help you organize your thinking process, and Copilot to complete code or write repetetive code (such as test scripts or code that is not easily to avoid a level of hardtype).
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
So... I dont have to start farming?
You don’t own land
Still won't hurt to have it as a retirement plan, tbh dude being a retired and growing your own food from the source is pretty sick ngl
@@m0rt_s3c24 a retirement plan? Bro farming is hard and expensive work and you need land to do it.
@@NihongoWakannai I said do it for yourself not some Capitalist bs, just grow your own food and chill. Plus you'll be healthy and all too
all the land is owned by corporations to build strip malls or mega farming conglomerates. welcome to the fall of rome
You earned a sub from me. Its refreshing to hear someone thats in touch with the population as a whole and not just the wealthy midddle and upper class.
TECHNO FEUDALISM!!!
In death stranding, everything becomes automated except for “porters”(delivery men), and buried in the lore it’s revealed that at one point it was automated but was a failure because people like to work so they went back to the porter system.
Meanwhile Elons billion dollar robot barely picks up an egg...
And it's kinda slow. But they made amazing progress in the last two years. I mean there have been humanoid robots showcased for decades but couldn't do a real, useful job until now. Boost the speed, add some basic intelligence and you have a physical agent ready to be deployed. If it can replace all the repetitive, boring manual labor then it's a great success already.
Same thing happened in the early 2000's when non-tech CEOs and CFOs though they could outsource dev to cheap labour pools overseas and fire all their high paid but experienced home grown devs...
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
2nd amendment fixes all this.
Maybe since most companies decision makers and middle management can be replaced by AI, surely we will be working for AI and fetching them buckets of electrons from the well?
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
There's no doubt in my mind that before things will get better we will go through possibly years and years of dystopian poverty and a shift towards ultra poor and ultra rich with a wide chasm between two groups. There is no job which won't get automated with the use of robots. Then people will have nothing to do and they will lose their sense of self / belonging / purpose and we will see a huge spike in mental issues as well. It's also possible that things will never get better. Once the rich will automate everything what's the point in keeping the rest of humanity around ? They won't really need consumers anymore.
I can comfort you by saying that it will all end once the oil is scarce enough
We are ALREADY living that reality man
love your takes micheal scotts
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝🔝
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝🔝
Remember the Don't be Evil phrase from Google. While a company can say it wants to enhance your life, it likely will take you out of the equation the moment they get the opportunity. Try to stay in the opensource domain.
On the topic of “the ability to work being important”, y’all should read Player Piano by Kurt Vonnegut. It’s realllly interesting
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
The first one that must be replaced is Sam Altman himself. I vote to send him to Mars 💯
Automate CEOs. Everything gets better for society at large.
AI replaces MANAGERS, not workers! Anybody in a productive job is ranked higher than anyone who is overhead in AI models. In an AI run company, everyone is on task and every task is verified, collected, part of your internal record of progress or lack thereof. But if you're sorting potato chips, yeah, replaced.
I remember a time when making AI meant people are free from doing the boring, dangerous, and monotonous tasks, and we're left to pursue arts and such. Literally, what is happening right now, is that art is being made by AI, and people have to do things the AI cannot do yet (physical/boring work).
The layman that pays for the ai to make a software, then the output software doesn't works correctly. Now the layman has no idea what is the problem and how to ask the AI to fix, or he doesn't even know the problem is there and it only seems to work. Great software.
People look at AI and think: "it's like industrial revolution but for the thinking jobs". Or "AI already got so much better than it used to be".
But people forget - past does not determine nor predict the future. We can make some guesses at best
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝🔝
We're all being Turing tested here and seems a lot people are failing! 😂
as someone with ADHD already, these suggested line completions (in my case, salesforce codegen) are pretty close to what i'm looking to do about 10% of the time, and the other 90% they're distracting and only serve to destroy my fragile train of thought.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
Someone should call Luigi
Man, this is a great speech! Even though we don't know if the Internet is good for society. So, we need to take with a grain of salt what these CEOs said
Loved this dialogue. Especially the latter half is what it seems is becoming more of a reality.
ok ok, now i need a full prime philosophy talk
hope everyone working or future working for him asks 4x more after the "fire humans hire AI as employee" fails. This is not detroit become human mf
The concept at the end there is what caused the Butlerian Jihad in Dune. A few people controlled the AIs which controlled everybody else.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
I dont know man. I think this is really going to happen.
Not sure when, but the technology keeps improving. We are all biased as developers.
It's def going to be interesting to see how it all plays out.
"It keeps improving" says who? There is absolutely no guarantee LLMs will magically get better forever. If you simply look at the billions spent and the fact they've already digested the entire internet without achieving AGI, you can see most models are capping out. Plus most of the good AIs. Are Actual Indians. Your coding request takes 5+min to process because Sajit has to pick up the ticket and do the work. Multiple companies have been busted for this. Generative AI does fine when all you want is slop art to sell products to uninformed consumers. It's an absolute travesty in any field that needs precision, reliability, auditability, legal responsibility...
I think UBI can help with the job thing because people can take more entrepreneurial risks opening small businesses and employing each other in small local businesses providing local services. Though crowd sharing apps might take this over.
We need an LLM which will release a new JS framework every week
Human: Html for food
Devin: Html for 500
The "LLMs are just next word predictors" argument doesn't really hold water. What if _we_ are just next word predictors?
How do we know that we are not? _We_ can't produce more than one word simultaneously (for the most part).
Humans aren’t just “next word predictors.” The Beginning of Infinity by David Deutsch has a good explanation - great book. Long story short, the creation of explanatory knowledge requires conjecture and criticism, not prediction, and is directed at solving a problem, not prediction. If a problem hasn’t been solved before, how can there be appropriate training data to inform the prediction? I like the example of the invention of 0 as a placeholder in the Hindu place-value number system. Without a placeholder symbol representing “no quantity,” a place-value number system has ambiguity. Eg, the string “1 23” is ambiguous: does it represent “one and twenty three” or “one thousand and twenty three”? 0 seems to have been introduced to solve this issue, and in doing so unlocked the concept of 0 as a mathematical object that could be manipulated algebraicly which enabled further advances in mathematics, such as the idea of negative numbers (numbers less than 0). At some point, the zero symbol and the idea literally never existed. The Romans and Greeks didn’t have a 0 equivalent in their numeral systems. So when the first person did the “0 as a placeholder in overflow columns” trick, they had no past data were they “extrapolating” from to solve that problem. There are lots of examples in science too of people creating new ideas, imagining new entities, and creatively conjecturing alternative models of the world that didn’t exist in any data set before they solved the problem they were working on. Humans are not merely predicting next words; we are solving problems, many of which have not been solved before or demand solutions that are original, and thus can’t be encoded in data sets representing our existing knowledge.
Quality assurance with AI output is still a huge problem, which cannot be solved completely, because AIs can only be trained on existing labeled (in this context: validated) data. And for that you need humans to do the data validation so that an AI may learn it, so that it may determine whether the output is good.
Right now all LLMs make so many mistakes. Therefore, LLMs are like a very good "sed" or "awk" command and I often use them to generate syntax based on templates (e.g. classes based on JSON structure). But they will fail horribly when the task is based on semantic implications rather than explicit syntactical tasks.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
Please add the Devin I wanna more money in my wallet when the companies hire me to fix it.
IMAGE you have a fucking LEGACY in JAVA Struct lol it will be amazing I wanna see it
Kudos to Flip! putting just enough effort at the right place 2:14
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
I think the impact of AI on developer jobs is more nuanced than a simple replacement. The current infrastructure isn't optimized for AI, which limits its immediate capabilities.
However, as the infrastructure evolves, AI will undoubtedly automate many aspects of development, potentially making some current developer tasks redundant.
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝🔝
Emily Chang: “You have an incredible amount of power at this moment. Why should we trust you?”
Sam Altman: “You shouldn’t”
Good will surely finish what he stated in your life, for everything works together for good, if you love the podcast, to know more about this podcast 🤍🤍sᴇɴᴅ🤍🤍ᴍᴇ🤍🤍ᴀ🤍🤍ᴅɪʀᴇᴄᴛ🤍🤍ᴛᴇxᴛ🤍±𝟷𝟺𝟺𝟾𝟾𝟼𝟽𝟾𝟼𝟷𝟾🔝
Devin chaos-ing through the Prime Again’s infra instead of push to master was top comedy and could be the vindication of older GPT3 girlfriends people needed
It's another tool in humanity's toolbox. Tools can harm or could be put to good use, but fundamentally it's better to have more options as long as you are not confused about this options implications= ai/ml good, but could be harmful if misunderstood and 'AI replacement of engineer' great example of misleading to take advantage of misunderstanding.
I totally agree with prime on the fact that these tech bros are so certain about what will be the outcome of the AI revolution. There is definitely a shift in power and fewer of us individually will have a say in things. I’m not saying world was 100x better in the olden days, industrial revolution was critical for humans but still it’s an ongoing experiment of not clear outcomes.
Wdym by critical?
One of the top leaders in the state government I work in does essentially nothing but go completely ham in meetings about AI. He said we have a “pro-AI governor” as if it’s smasmorshin or something. But in the same exact meetings where he breathlessly extols the unbelievable sea change this is ushering in, he says “we just need to learn how to consume it and what it can do for us”
Hi Cart, it appears you have left horse behind.
The only appropriate size of context window is lifetime. That's the best mimicking of human thinking
Driving people towards despair is the worst thing about this stuff and the way it's being marketed. People want to actually do things and see the results of themselves doing things, but if the marketing around this stuff makes it seem like people are just obsolete then even if it's not true, plenty of people will fall to despair and do unmentionable things.
People ask me "aren't you afraid AI might take your job", i tell them "my jobs to automate your job, so if my jobs automated what's going to happen to your job?"
Why is there no AI replacement for middle managers? Surely that would be more effective...