I am tired of CEOs and AI. Facebook has sucked for a while now. I don't want to listen to this over and over again. As they try to continue to inflate AI and hype there stock price. I am going to continue to work on my engineering skill. CEOs are not engineers
I think this is a PR exercise. They are peddling the narrative that AI is actually replacing engineers or rather increasing their productivity so much that they need fewer of them. Displacing might be the correct term. In order to free up funds for AI investment they need to reduce head count. Either that or they are using AI as a cover for what are likely bog standard layoffs and cost cutting.
No one knows how far companies will be able to go cutting humans out of the loop. But does anyone think society would find it acceptable to have somone like Elon Musk or Sam Altman and maybe a dozen investors all own a Trillion Dollar company with zero employees? Part of owning a business in a captialist society is to provide jobs. If you no longer do that, a new paradigm for purpose, pay and ownership must be determined. That AI should be seized and owned by the world or the country to benefit all. No a small set of lucky investors who forever get to live off the brilliance of a machine. I dont believe there are many functional societies that can exists where human work is not a factor.
@RohgishSun a universal basic income may be valid. But humans still need purpose. Arguably, it may not need to come from jobs. Not all jobs provide purpose. But if we are not at least partners in the continued growth of humanity, I think that will be spiritually detrimental. It may all simply be fear mongering that AI will be so perfect that human thinking becomes second class. But time will tell.
@@amesaswi dont know, spirituality has less to do with your work, look at buddha or jesus, they didnt do anything and they didnt need purpose either, they were fulfilled, maybe that kinda pursuit will become mainstream
@Ojas97 purpose is usually derived from personal growth and community. If human knowledge based work losses all value, as society perceives value it will leave a massive void.
It doesn't really matter whether we accept it or not the question is it going to happen we'll see because there's nothing Any of us could do about it except cry And wine a little bit
I still feel like people don't understand the basic problem. No - AI won't replace you completely (not yet). However, it will speed up work to such an extent that your company won't need 10 employees, but e.g. 2, so 8 will be unemployed. This is already happening in many industries. So when they say that "there will be no work for programmers" - nobody means that "programming will disappear", only that there won't be as much demand. I'd love to come back to this comment in 2-3 years and see if I was right.
@@perrym8048I partly agree with you, except that there's a distinct possibility that AI, unlike email, will also be able to do most of the new jobs it creates.
I love how people smoke crack cocaine and say shit like "AI will increase output." Thats not how things work. Thats not how anything works. Companies don't "maximize output." They maximize profits. Doing the same work as a 10 man team with 2? OH YEAH LETS KEEP EVERYONE ELSE ON AND WORK MORE. We are all so insanely fucked, and this rhetoric prevents us from addressing critcal needs such as UBI and workers protections.
The problem is far worse than that. AI is going to gradually (or not so gradually in some cases) reduce jobs to the equivalent of typing what you want and then following up to be sure that it's done to your specifications. This lowering of the bar is going to average salaries across the board. What's worse is that AI labor is going to be far cheaper than human labor, and it will only ever get cheaper and better, so human labor is going to go down from that as well. Your debt is going to stay the same though, so good luck paying off that $500k home when you're earning burger flipper salary or less. Banks will fail, the economy will collapse and we'll be in complete chaos. The wealth gap will be an insurmountable hurdle that will only increase as AI improves. This could all be avoided with some government planning, but they aren't even talking about it. All they want to talk about is using AI in warfare and other obvious, but much less likely problems. When anyone in govt. tries to bring it up, they are told "it's too early to talk about that for a tech. that's not even fully realized." By the time it's fully realized it will be too late to come up with a new system.
There’s quite a few contradictions in this assessment- how is stating in black and white that not requiring 10 employees but 2 so 8 will be unemployed not saying that AI is replacing humans?? The SE market is saturated as it is, if there’s less jobs this will be forcing people to seek other careers, often resulting to a huge pay cut. And only those 2 remaining SEs will be altering their jobs to code this AI. I’ve actually not seen a great deal in this space working as an SRE and the automation we do isn’t replacing many jobs but by the sounds of it, AI will and it is terrifying. Maybe it’ll take decades but it will hurt.
Really? I thought it's already way better than most average programmers, that, now a senior can effectively do without his team, than waste his time in a useless team and training them, and take care of th people issues.
Yes, now Ai can do a lot of work for you or instead you... actually, there are some days on my work when i just have a chat with the ai agent in the cursor and fix its generated code just a little
ASK an AI to visualize a Cube with OpenGL with 1000 instances and you will See the Code sucks. A Beginner trained with a Tutorial can solve that @@josephs4044
@@josephs4044 Sometimes it speeds up but one person can't have endless responsibilities. "way better" it still doesn't reflect or have new ideas except randomness, it just summarize historical data and it will never be up to date to trends and changes in languages and frameworks. If we can't agree why would we agree about AI result?
When emails came out in the 90’s they said, sending ng letters will be dead. Post offices will close. Postmen will loose their jobs. 30 years down the road, there are less letters sent. But there are still letters sent. Postmen had more job nowadays because of ecommerce. The coding might ending for some developers like UX/UI designers and developers but their will be more demand on a different scale for developers who will know how yo utilize AI. Job landscapes change. But they are still there.
@@4ka07_muhammadrizky I know. What I am saying is that the job landscapes will change. You need to go back and study history to understand what I am saying.
"Are gonna have", "once you have that", "over time we’ll get to the point where".. it’s all just empty talk. Devin, the AI that’s supposedly replacing software engineers, can’t even commit a single line of code without taking forever and still messing it up. It’s ridiculous. Stop hyping this nonsense like it’s revolutionary when it can’t even handle the basics! Software engineering is so much more than just slapping some code together. Even writing good code is hard. It’s not about spitting out lines that look fine, it’s about making them actually work in the real world. The current AIs are just throwing out code snippets and crossing their fingers. Replacing engineers? Give me a break. If that ever happened, everything else PMs, CEOs, managers..etc would vanish. Software engineers solve problems, build systems, and keep everything running. Let’s see these so-called ‘AIs’ try to take that on.
I have used Chatgpt to debug issues I find . It does a fair job for the most part but also has hallucinations and gives you solutions that simply don't work at all.
He said, AI will be writing a part of the codes, not the entire programme means we are far from getting full codes written by AI only but can get help or assistance. Technology is not our enemy.
I recently used ChatGPT and Claude Sonnet to write some fairly simple shell scripts. It did an amazing job and ultimately produced a script that worked. I then found several edge cases that it didn't handle correctly, as well as spotting redundant code, security flaws that would come from input that hadn't been properly cleaned, etc. Without the years I've spent programming, I wouldn't have spotted all those problems. By the time I was done, I felt a weird empty sense of sort-of accomplishment. The actual task of writing the code from scratch would have taken a lot more time, but the process of cut/pasting into ChatGPT and telling it to fix things was downright boring. I don't think I would enjoy being a SE if I were using AI for much more than occasional guidance or as a time saver for actually typing long complex patterns.
Seriously people who believe this , I work for a big company , most senior staff can’t even use excel properly and you think they gonna code complex sql or python , not in a 100 years lol . Pretty sure he is selling a hype to justify billions they spent
@@silvershadowjshbut the main reason they're even doing all this is to get trust and funds from non coders. So, if non coder won't be able to use these AI to create their dream facebook, then what's the point? Why invest in it?
Senior engineers don't need to use excel? lol, I haven't touched in years. Can do basic things but if I don't use/need it why would you bother investing time in it
I burnt out on development, AI has brought me back and I can do much more. It still has its pains, but I am the type that likes to figure out hard things.
The Ai people are lying to the CEOs. If you’re into AI, and a software engineer, this sounds , for very obvious reasons, utterly stupid. Copilot and other ‘AI’ tools have been around, which are only code generators. Too much to go into and not worth the time. There will be dev jobs for decades. New language versions will be developed. This is all hype. AI needs to generate interest to support the MASSIVE funding. This is sales talk.
@@johnmadsen37your statement by itself is condescending instead of sounding like you’re giving logical and sound advice, which is far from that because, that statement alone is an argument that’s in the same level of a 2nd grade. Your raw intelligence doesn’t match your emotional intelligence, hence why you’re being immature and egotistical.
@@easylife6348 here is another one. Did you think bill gates was coding windows past the first couple versions? Maybe you think the google founders are still writing code for their search engine? It is totally insane to think so. Zuk did what was essentially a prototype not even called Facebook before he got funding and hired real developers. He manages. At a high level. Like most founders. This is comical to explain this.
@@johnmadsen37I didn’t said zuck does AI, use common sense please, I was implying that zuck most likely knows enough about AI to formulate his opinions about AI within the company that he built and manages. All you have to do was asked a question that’s related to my elaborated statement of what I’ve meant about zuck on my first comment instead of you being emotional because you refuse to use your critical thinking skills before ask a rational question about what I really meant by my initial statement.
Why there is the preconception that Zuckerberg has the quality to make such assumptions? He is just rich some kid, who got lucky with his idea, not with his engineering or technology.
There’s a strong economic argument against mass displacement of software jobs by AI. If millions of software engineers lose their jobs, their reduced spending power would cause ripple effects throughout the economy, negatively impacting demand for products and services across industries, including technology companies that rely on consumer spending. Technology companies thrive on robust consumer markets. If large-scale unemployment leads to decreased spending on products like smartphones(iPhones), electric vehicles(Tesla), travel, and subscription services, it directly undermines the revenue streams of the very companies driving AI adoption. This would also slow innovation and growth, as companies would struggle to justify further investments in new technologies if their customer base is shrinking. Moreover, layoffs at such a scale could destabilize economies, reduce tax revenues, and create social unrest, all of which would further harm businesses, including those in tech. In this scenario, companies might find that the short-term cost savings from AI automation are outweighed by the long-term economic consequences. Ultimately, it suggests that while automation can increase efficiency, the balance between leveraging AI and maintaining human employment is crucial. Companies may need to consider reskilling displaced workers, creating complementary roles, and investing in sustainable economic growth to avoid harming their own markets in the long run.
You're absolutely right. In fact, the ripple effects on ALL industries will be unprecedented.. which is why, there has been quite a bit of talk in the recent past about universal basic income.
just another argument why AI utopia will never happen... for me the most convincing argument is principle of responsibility: who will be held responsible when people are killed because AI surgeon malfunctions, sometimes? or AI written program caused train / car / plane crash... as AI pundits proudly states, they themselves don't know how LLM works, for example... if no one can be blamed (with justice) - how to ensure errors don't happen again?... there are other interesting arguments (what will happen if people are not learning skills, thus not being productive anymore - from where AI will take new learning material - as stagnation means game over in economy)... etc... my prediction is that AI bubble burst will be so powerful that world economy could simply break - as people will lose any faith in others, telling them smart words
Facebook's engineers are getting more efficient, but so is the competition. This is a temporary effect. If it could replace software engineers, they would have fired them all.
The later analogy is exactly what we see with transportation infrastructure: the wider we make highways, the more people we have on the highway. Increased efficiency increases use.
Why people are trying to replace SWE in the first place? I'd say it's the one of the hardest desk jobs to replace. It is way easier to replace positions like lawyers or recruiters, but nobody says anything about those jobs. Everybody is trying their best to replace SWE's for some reason.
It's actually easier to replace managers and CEOs. All they do is look at excel sheets to make some decision. This can easily be done by AI since AIs are good at number crunching and understanding trends and patterns.
@@realnapster1522 It can already be happening. I can imagine something like "Hello ChatGPT, I am Mark from Meta. What lies should I say to the investors, so that they won't fire me after the metaverse failure? ". And now we have this vid.
AI is useful for performing math it is programmed to use on the data but it still is up to a human to interpret the information to mean anything @realnapster1522
We are entering an era where the focus is shifting from how to build apps to deciding what apps to build. Let AI do the coding, you decide what to build. Basically you become an architect.
yeah, till the point the system breaks and someone has to get hands dirty fixing it... but without basic skills and knowledge - who would be able to do that? similar to AI driven cars - if everyone would use them, there wouldn't be a driver skilled enough to even turn the steering wheel properly... skills have to be honed day by day, you cannot jump from the throne of architect into sewers of implementation :)
This is like saying excavators made manually digging obsolete. What happens with tools like Ai is everyone is just able to accomplish more with the tools.
Excavators made much less need for manual digging, and excavators can't replace people in the process as someone drives it, and it is not good for little digging tasks. This isn't really true for an even better AI than today's.
@ Is there anything in your text that makes it clear ‘you’ are a person, not an LLM? Touring test etc… That guy was clever. There’s a reason for why the test was set up the way it was.
He is nothing more than a manipulator, an opportunist who always positions himself on the side that will bring him the most money. Claiming that AI will replace engineers is merely a tactic to instill fear among professionals, pressuring them to work harder to avoid losing their jobs. This approach is not only low but also outright unethical. However, it’s already expected from him and other self-important CEOs who present themselves as the ultimate authority on everything. Let me remind you all that it’s not CEOs who build AI or rockets, but highly skilled professionals who have dedicated years of study and effort to achieve their expertise. If anyone is qualified to discuss AI and its potential to replace us, it’s these professionals, not CEOs driven solely by profit.
Best case will be a drastic reduction in number of programmers needed…for sure less than half as many, likely only 10% or 20% needed. Worst case will eventually approach 0, or just a couple of senior programmers per division of large companies. There will always be exceptions, but the average numbers will definitely decline.
Without affordable tools with 100M context token, this will not happen Developing a complete system with interlocking modules you need a very large context for the tool to understand the whole system Also for legacy system, you need to reverse engineer the system to define the prompts or context of the system properly, if anyone look at Agile it does not encourage complete system design Yes, as I said to others before, AI tools have the potential to reduce staff, unfortunately no established or accepted way to utilize AI tools at the moment, some IT shop even discourage AI tools which for me a waste of valuable resource I really don't know how AI revolution will change IT setup, which is very entrenched in Agile which for me not designed to leverage AI tools User Stories will never be efficient as prompts because it lack technical details and independent against other User Stories, this must be replaced by some sort of system wide thinking Good luck to everyone because I see more chaos before things become useful and efficient 😉
first you need a person who will describe the requirements to the AI tool, so these requirements will not conflict with each other. 😅 What they call AI is just a parrot. Can a parrot replace an human? Of cause yes!! 😅🦜
Why does"The End Of Software Engineers" give me a vivid mental image of A Giant Zuckerberg head nested in the primordial soup whilst gazing at that stars
I am a software Engineer and I can really see in the past 1 year AI has really changed the way we work, unless there's an increase in demand for building more, there would be people who will definitely be no more required. All we can do in these times is up-skill yourself and provide true value which is not easily replaceable.
its not that we not gna be coding, its just gna be a higher level version of coding. same like how most developers don't code in computer language and these days do it in higher level software languages, its just moving the bar more up.
Just like anything in general, take music for example, thousands of people try to become musicians but only a few become popular and rich, take writers for example or scientists, coding will become just like that which was always gonna happen because its just too saturated
He didn't said that. He was just mentioning the use of AI agents. It's just a beginning process. And it will take time. Anyway, human engineers are always required.
I work for a company that uses generative AI. You have no idea how stupid AI is. It came a long way, but it's still stupid as hell. Try it yourself, use langchain and langgraph and find out how it can't do anything slightly complex.
As a professional developer i aleeady spent 1 year learn how to talk to claude. I talk about code part for 10-40 minutes before i let it implement. Then i check the generated code almost each line. If you just let the AI to implement without coding experience it generates bullshits. Any functionality can be implemented so many different ways. What fo you optimize for: cpu usage, disk space, network communication, etc. there are thousands of questions how you structure your data, what you emphasize on ui, how you modularize (scalability) security, etc. Only the not developers believe this what Zuckerberg says
Right. I feel like most of comment section is not SE. They just talk about this while never touched any advanced code in life. And AI is very weak on old technologies. Just few days before, I was arguing with ChatGPT what is better for refresh token in auth system, JWT or Opaque token and ChatGTP had idea of "combine best of both worldw" and give me solution that actually combined worst of both JWT and Opaque in every term, of networking, performance and storage space. Then I pointed it out and was like "Yeah, you are absolutely correct!...."
These guys should be ashamed of themselves. The only thing AI has been able to do is to act as a chatbot. It has not even been able to that up to 30%. Facebook that is full of security flaws even with humans programming it, is talking about replacing those humans with AI. Hahahaha...a little kid will crack down their website when that time comes. There are far more important things that the world needs AI to do for it (if AI actually have the power to do it), instead of taking people's job away from them. Zuckerberg should be ashamed of himself that at his age, he still talk like a kid. In fact all the CEOs thinking about replacing humans with AI. AI can't do shit. Problems like email hacking, business email scams, romance scams, health diseases, etc, has not been able to be solved by AI. Then you are thinking about taking people's job. And look at the way they are even saying it, saying it with so much pride. Common sense they say, is not common. Pathetic.
It's like saying "Don't learn a foreign language, just use a translator(which by the way was developed by programmers)". Now try to translate Dostoyevsky's "Crime and Punishment" with google translate. That would be something!! Coding will change but never disappear. Zuckerburg is the only one who lives in his Metaverse.
I started leaning Programming not so long before the ChatGPT and AI appeared. I started using them so I can understand some code and learn to get things to work .In my learning journey I am still half way through or maybe a lot more. I would say, I always got frustrated and discouraged to stay and keep learning but I trust some few people in this industry and they seem to imply what you just presented here. A good programmer will be always working and finding oppourtunities and I do believe that. Thanks! This video appears to be fair about the whole AI thing.
Once all the coding is automated I am so excited for the collapse - all of a sudden competent engineers will be in extreme demand- hacks will be prevalent, leaks, etc - also the quality of these products will tank drastically. It will work until the entire backends of these enterprises are legit spaghetti. Then these corporations will be ripe for upset.
AI will streamline projects, lower costs, and make technology more accessible to a wider range of companies and organizations(MORE PROJECTS). In reality, the growing reliance on AI will likely increase the demand for skilled engineers to meet the rising challenges. However, these claims often come across as exaggerated marketing tactics.Ironically, this type of AI narratives might discourage some people from pursuing careers in computer science due to misconceptions about AI replacing human expertise. We will face challenges in supporting such a high demand for projects. It’s also important to remember that AI is not a magical solution-it requires vast amounts of high-quality data to function effectively. As technology evolves, AI may struggle to keep pace with entirely new advancements.
Many people do not think of this problem in terms of economics. The principle of economics states that the lower the cost of developing software applications (lower barrier to entry). the more competitive the market will become. The big software companies will start facing enormous number of startups trying to compete with them. Those startups need to hire more software engineers to help them. As a result, fewer developers are needed per company, but more companies are there to need engineers.
Supply and demand. Price of intelligence goes down. Therefore, in general, it will become ever harder to get a job as a software engineer, amongst other things. Whether we will be witnessing this trend clearly due to AI in 2025 is another thing. Anyone who wants a job or wants to keep a job will need to be familiar with these systems.
Or, people expect software engineers to be replaced by AI, therefore they don't study engineering. A few years later AI didn't match the expectations and now you have a shortage of engineers and everyone gets paid more. It's all possible until it actually happens.
The speed of development will increase but the quality requirements will come along. Look at the quality of fb - hard to use it. I doubt that devs will be massively replaced!
While I hope you're right, I believe that's super wishful thinking. Salesforce just announced they will not be hiring ANY SWEs in 2025 and the City of San Francisco announced they will not be hiring again until June of 2025. I don't see any software company massively scaling up SWEs but there's evidence of many laying off or slowing down their hiring.
As a business owner myself, I'm not hiring new SWE's. I'm looking for architects and communicators. Project managers with enough skills to be able to translate the client's demand for A.I. SWE's. Until the smarter clients instruct the A.I. themselves.
Coincidentally someone from Salesforce is on the Board of Trustees of the World Economic Forum, I guess that prediction of software developers being the #4 fastest growing job area is just for other companies.
claude is actually intelligent , i was prompting the other day to claude that resources are limited and it in response said that , that is like law of conservation of energy in physics , it can find patterns , also brainstorm well , if it gets better it will actually revolutionalize way we work in all walks of life
You got tricked by an excuse - this is like how dogs can train their owners. It can't complete the task to it tricks you into believing that lol. You got operant conditioned
@@bananaear23 do you think ai is not doing tasks humans throw at it , it is , it can generate images , edit images, and understands language already , also it can code problems that were not in its dataset
@@anandkanade9500 I will give you $10,000 if you can make an AI generate the image of an analog clock indicating 3:35. since it's so smart, you'd only need 1 or 2 prompts for it . good luck
I’m ngl it’s useful but it’s not that good lol. It makes tons of mistakes, but it’s like a good google search or tool to jump off and make your own documents and conclusions
Mark isn’t saying jobs will be replaced, he is just saying a lot of code will be written by sort of mid level AI software engineer, you still need a dev to tell the AI what needs to be done, what sort of approaches etc. If you’re in a faang or some corporate you understand that a lot of that code is already written by previous engineers and some engineers who manage infrastructure and us devs mostly focus on initiatives, delivery, features, processes, discovery etc I being a software engineer rely heavily on AI to do the labour work so i can focus more on delivery and quality. Listen to NVIDIA CEO, your job will not be replaced by AI but people who use AI. I agree this industry will only expand more. In my company we’re already using AI to provide more value for customers. i.e. more to do for us
Yeah. You will be leading an orchestra of AI- bots. For 3 years. If you belong to the best 20% who have the most experience. And then AI will do the same. And then you are completely gone unless you are one of the best 1000 in the world. And then you´re gone
@@KyleStangline Yup. Agreed. He also forgets here, that while there will be more software, which can be used for people, LESS actual workers or companies will be needed. You will get your software written on a mediocre level by your own AI. You will not go to a company which hires programmers. It is like when you can go to Gordon Ramsays restaurant and have the best meal available. Or you just cook at home and use your stove, blender and kitchen machines for that. Very many people will use AI at home. How many people will by AAA games for example, if they can let their AI programmer program them ps4- level games but exactly as they want them? And improve them. Adjust them. Their OWN games. And if they want another one, here it goes. And another one. And another one. The Ai just spews games out without an end. The amount of people buying AAA products from companies will go waaaay down.
@@Dynamite3783 no that is not the solution. Only part of that I agree with, unionizing is good if for the right reasons and with the right structure and leaders. Unions are not a miracle cure for these issues. We need to become a united coalition of workers who speak with one voice to end suffering using every possible tool including ai. For example, if the docks can be automated as we see in China, then the dock workers can work a different job in the meantime before that too is automated. If enough jobs get automated then we will all have less work to do and eventually we can reduce mandatory working hours for humans to 0.
Facebook and are creating their competition and downfall. If ai can replace mid level engineers 👨💻, all engineers will be able to exploit ai and build all kinds of applications.
Developers still use the terminal a lot, but what if someone build reliable RAG chatbots that can control the terminal? Suddently dummys like me are capable of everything. Today the terminal and coding tools are still a barrier but as soon as that disapears things will evolve very quickly.
This isn't a bombshell, it's expected. We've heard for a while that 2025 would be when they release agents, and we know that Meta has been trailing the pack, so this just confirms what we saw in Dec. when OpenAI showed they have master / grandmaster coder AI. Programmers will still be needed because AI still has problems with problem solving (until maybe r-Star Math is employed, then all bets may be off.) But programmers are going to more and more become just prompt "engineers" and I use the term "engineer" lightly because really it just means typing what you want and then when it comes out wrong, you tell it what you want to change. You know, super hard stuff. Burger flipper salary anyone? Time for a CS degree haha. I think you're making a huge assumption when you see the future jobs report show that Software Engineers are the 4th fastest growing jobs. You're assuming those will be human positions... If AI is faster, better, cheaper, why wouldn't they be AI? You basically proved my point when you showed Jevon's paradox. AI labor will be more efficient to use than human labor, therefore there will be more demand for AI labor over human labor. In the short-term I think you're right. The need for human developers will increase or stay steady, but once AI can do literally every task that a human developer can do (AGI in 2-5yrs) then there's no need for human developers at all, hence why the OpenAI dev said "You'll know when we've achieved AGI internally because our job board will be empty."
lol you sound illiterate, Seriously people who believe this , I work for a big company , most senior staff can’t even use excel properly and you think they gonna code complex sql or python , not in a 100 years lol . Pretty sure he is selling a hype to justify billions they spent
It's not a bombshell because it's utter marketing hype bullshit....and what are you talking about w/r to "this confirms what we saw with OpenAI" when we did in fact didn't see anything from OpenAI or any other of these grifting clowns w/r to them making good on the promises they keep making. Funny how people like you keep repeating this nonsense when every Softwar Engineer who actually knows what they are doing dumped Devin/Co-Pilot/ChapGPT, etc when they found out how absolutely USELESS they are for engineering non-trivial software that people are actually going to use.
@@perdify you're simply ignorant of what AI will be able to do. Do some research before you post. You're thinking of jobs as they are defined today, not as they will be in several years. Jobs in the future will for the most part consist of you telling the computer or android what you want it to do and it does it. If you don't believe me, search around for software creation tools like lovable, or music creation tools like suno to see what will be possible in the future. Yes, those tools aren't great yet, it's not the future yet. But it's a glimpse of where we are heading. Every job, even farmers will be completely transformed. The farmer of the future will be more like Luke Skywalker complaining to Uncle Owen that he has to babysit the moisture evaporators rather than complaining about having to till the field.
@@BruceWayne15325 Jesus , you are still typing essays . Like I said , ai has been here for ages . The only company making a profit from ai is nvidia , which is selling the hype . Stop being too online . Get a job . Ur family did not raise a lifeless nobody fantasizing about ai taking over the world lol
@@BruceWayne15325 lovable and suno are two good examples that they are not superior at all, its just more ways to do it with it's own problems with randomness and hard to get exactly what you want. prompting and natural language will never be exact. at the end of day it's the person using the tools are that are responsible for the outcome.
Seen this for every occupation that has posited the "oh, everyone will be so much more productive" angle. First, WEF looks at current trends and projects into the future, they do not rely on potentially disruptive technology as they do not know when it will show up, and which area it will target, first. Open AI is hiring the best of the best as that is the bar to develop the ultimate ASI platform. That is not a broad trend, that is a specific use case. It would be like looking at Los Alamos while they are developing the A Bom and saying, "wow, there sure is a big demand for nuclear scientists". First, very few could compete at that level, and second, there is a very small pool of demand for that kind of skill. If average intelligence in society is around 110, (including many CEOs which I would say are mostly grouped around 115-120) then exactly what creativity is going to be unleashed by a super intelligence clocking in at around 400 or so? And remember once this is achieved it becomes replicated millions of times so as to be ubiquitous in society. It may take 5 - 10 years to see this happen, but humans are very bad at following exponential trend lines, and AI dev is on such a trajectory. I think it is putting your head in the sand to say "well, maybe it won't be that bad", or "well, I am really, really clever and smart so I won't be impacted". It is time for people to pivot to making demands on what a post labor society looks like before so much wealth and power accumulates into Silicon Valley that we will be living their version of utopia. And that version will have been built while we were still trying to figure out how successful they are at developing AI. It is coming, it is just a matter of time..
Seriously people who believe this , I work for a big company , most senior staff can’t even use excel properly and you think they gonna code complex sql or python , not in a 100 years lol . Pretty sure he is selling a hype to justify billions they spent
Not end of software engineers. End of coding as we were used to it. Software engineers will get a more high specter of projects they are working on. I am from financial tech industry.. And believe me.. For the next 15 years, you DON'T want AI to get your card data.
Will it be the software engineers orchestrating the AI agents or their project managers? In my experience, most engineers do not really understand how a product is meant to come together they only are able to complete tasks they are told to complete. The PMs always have the big picture in mind. Cannot wait for the day the SCRUM Master leading a sprint retrospective with a bunch of AI agents lol. Seriously though, I think the PMs may be the winners at the end of the day as the world of AI unfolds. There will be winners and losers, no way around that.
At first you will need project managers / architects. You need to know enough about software to be able to instruct A.I.'s. But at some point the client will talk directly to the A.I. You can already see that in say, translation projects. The clients directly talk to the A.I. to get the translations without the need of a human translator. And TBH, software engineers are really people who can translate the wishes of the client (or project manager for some teams) to the computer.
It'll be the product managers who translate user needs into requirements managing the AI software engineers. Software engineers who haven't developed product management skills should start now. The actual skill of software engineering is going to be less and less needed.
Possibly the worst take I've ever seen in my life 😂 Software Engineers defo don't just complete tasks they're told to do - they constantly have to think about the big picture and possible impacts and they help break things down into tasks (as a team). Good luck to these mighty PMs who undervalue their teams thinking they're the only ones holding everything together. Sounds more like an autocratic totalitarian than a leader that facilitates and enables the team. Also when AI is doing all the technical tasks - Agile probably isn't needed since business can have everything, all at once, right now so you can probably scrap the Scrum master role from asking what the AI did yesterday and what it's going to do today 😂😭
@@jonblack7046 only the more senior and more intelligent do. And in many projects, that don't have sufficient information to understand the big picture: they don't talk directly to the product owner and stakeholders. Most engineers have become engineers because they like to engineer and they have a natural resistance to architect and communicate. Systems like SCRUM have come about to encourage such things as normal part of engineering. It is like giving a bricklayer a blueprint of the building: they will start laying bricks, they don't question why the building is being built and if the spaces are big enough for the crowds, but the good ones do alert the architect that they think the foundations won't bear the load. But once the A.I. can lay bricks, they will lose their jobs unless they become architects.
@@jonblack7046 Maybe, but this has not been my experience. The most experience & senior developers might have that level of competence, but there are a LOT of devs out there who can solve problems, but never seem to be able to identify the problem/need on their own. That is why so many ERP and LOB applications suck for the end user. As far as SCRUM is concerned, I was mostly joking. I do see that level of PM changing significantly if not disappearing entirely; however, I do not see the Agile development routine going anywhere. Clients are still normally trying to figure out what they need/want throughout the entire development process and product requirements are always very ambiguous (especially early on). What I see happening is the lengths of sprints being reduce significantly (maybe 3-4 days at most if you factor in a QA period) and maybe future projects take on a more hybrid approach.
SWE should just focus on problem solving, and when lower level problems are solved by AI, it just means human can focus on solving the bigger ones... same as every other tech in human history.
Even now the developers are relying on Ai generated code. If u are relying on Ai generated code, you should know you will get replaced one day. Unlike other tools, Ai can do certain level of decision making and some level of learning process. If we previously need 10 developers to write and maintain the project, now we might only need one developer to review the ai generated code. Even now some Ai tool can review the code and fix the bug by themselves. The possibility is already there and it just need a bit more time to fully replace the human software developers in the future in some areas. But we can't reject Ai obviously. We just need to find a way to work with it.
A ton use it for boilerplate stuff, but that's grunt work. It would be soooo slow if you had a non developer continually copy/pasting the entire code base into some LLM whenever you wanted to do something. Not to mention how error prone that would be. You think this scenario is something multi billion or even trillion dollar companies where the smallest mistakes in code can result in millions of loss in profits are going to seriously consider anytime soon?
I only use it for iterative works and some unit tests but that’s it. It has improved my productivity by 2-30 percent when it comes to coding but AI generated code is still shitty.
Or the next generation SWEs will simply build AI based solutions and make themselves more productive using AI based tools. Nobody said they will need to work for Zucky whose goal is obviously to get rid of all of them. There is so much work ahead solely in making existing systems compatible with LLM based agents, designing and integrating new work flows with those agents, et cetera.
The other day, I read this on RUclips "Never take a CEO's word at face value. There is always an ulterior motive behind the things they say. It's not necessarily intentionally sinister, but the reality is that it's in their best interest to guide the masses (that's you and me) toward the kind of things that benefit them and their companies. I have nothing against individual CEOs personally, but they are the public-facing corporate voice of their companies. Even if a CEO made an absolutely factual statement, such as "the sun is hot," I would still have doubts about their motives"
This development is inevitable. Once you get to the point where you can simply ask a question to the ASI or Machine Hyper Intelligence in whatever language you happen to know and it gives the answer in the same. Then it is time to face the music...
Agreed. Reminds me of the old days, when there were "code generators", for certain types of apps. But AI will make this more flexible. SWE's as orchestrators of these resources. But for AI products, one now needs to be more than a SWE. It's simple enough to glue a bunch of things together through LangChain or similar. But the main issues regarding AI, esp. agents, workflows, is reliability and accuracy. One also needs to be a ML engineer. How do you set up test cases for proper evaluation? How do you tweak it in fine tuning, ICL, or some other technique, or specialized training regimes, to make sure the product is solid enough to be fielded? How do you utilize user feedback (which could be additional training data)? How does one manage iterative improvements and releases?
This AI Software Engineers will be more costly than hiring human software engineers. This companies that are driving this forward are those that would benefit more in terms of profit. I remember when they said that cloud servers are cheaper but right now it actually cost so much more than owning physical servers.
If you have AI and robotics that you could rely upon, then...Yeah. End of almost every type of work. In the professional domain, it's even easier to say so. Because it's easy for AIs and their physical presence i.e. humanoid robotics to do so. Thanks.
If not, then...Who are you... hunh.? Nobody, if you only have the education. Not even Nothing, if you don't even have a quality education. And, you are something..If you have got money.
With out the knowledge of programming, you can’t even prompt the AI to code. I would say the programmers will be programming 100 times faster with AI. I use AI to code crap ton of stuff, but I review every response I get before I implement the code. I would say Ai get about 30% of the time get it right first time, but again it depends how complex of a task I submit. If I give you all the carpentry tools that will not make you a carpenter.
Let's not 'sugar-coat' this issue.. There will be far fewer SW engineers hired as AI becomes more efficient. In the scenarios described in this video, 1 human SW engineer can support 3-5 AI SW engineer. Take Zuckerberg at his word..
The way I see it is that in the future, developers will also be PMs as I'm sure there will be AI tools (if they not exist already) for this as well. Also, the company cost will be reduced (you are gonna get the same pay for doing 2 jobs) and communication with the client will improve, avoiding the back and forth between Dev - PM, because who better than the same person working on the project to communicate the real scope of it, which is an area that PMs with no Dev background struggle with. So, I agree with what's being said on the video: we devs have to upskill in order to orchestrate all these AI tools
14:13 - This sounds like BS to me since obviously the other type of AI's built to debug code (even long and complex code) would do that automatically, if not now or in the next few years then later on at some point, it's only a matter of time, the jobs of everyone would be gone eventually and the only main jobs that would remain are those required to push the AI field and even those would eventually get replaced by AGI/ASI later on.
As a Senior Software Engineer I use o1 for work. Without guidance it ain't shit. Of course it can resolve or give suggestions. But everything needs to be monitored, though hallucinations are less frequent in comparison with other models. It eases my job, less google search, making me more effective. I rarely write code without AI. It's a no brainer to not use it.
As an architect, my software development productivity has increased 5000% in 12 months. I can write fairly complex Django apps without being a wizard at Python or Django - just need to know enough to put small corrections into the design. Same is true for AWS and Flutter/Dart.
Opensource contributors are being taken advantage of by big companies. For example, Meta started the Llama as Open source, but who benefited from that? Do they give compensation to those contributors?
This is how things arr gonna roll . 1 massive job displacement, unemployment reaching to near 30 -40 percent in the next 5 years . 2. Money will be printed and UBI will be established . 3. The inflationary effect of money printing will be offset by the deflationary effects of massive unemployment and rapid advancements in automation which is inherently deflationary. 4. The standard of living will be greatly diminished as UBI will only provide the bare minimum to survive.5 . Overtime with mass adoption of AI , the standard living for all will be greatly improved
Its not a good thing to be ignorant to the truth. It wont help you this time around. What you do is not that complex or nuanced. Just like everyone elses jobs. You will be replaced soon along with almost everyone else.
@@Dinofx_jr Then why aren't you out there right now using AI to build the most advanced _insert whatever system here_ and raking in the cash? The one being ignorant of the truth is you. Will AI eventually replace a large part of what programmers do? Yes. But its not gonna be 2025. But you believe what you want to believe. Come back to this post in a year and see how you did.
@@IyamwhoIyam no you dont read very well. What you do is not nuanced. Ai code generating agents are all thats needed. If not replaced itll still be 1% of what exists now
@Dinofx_jr You don't read very well. And you don't know anything about the industry. I can tell by your attitude and response. Perhaps you need counseling. Get some help.
If you're saying I'm going to be managing AI agents instead of being down in the weeds coding, okay. But isn't that going to drive down my and other SWEs' salaries because now our jobs are doable by a wider group of people? In other words, could it be that we remain employed but not necessarily thriving?
You wont be employed. I am an SWE and i switched to trades… you wont be overseeing any AI. The SWE workforce will be 1% of what it is now (population wise). In no more than 10 years.
No, you won't be managing AI agents because other specialized AI agents will do that. And in the end, a single agent would do everything. Why do they even need you to manage the AI agents if the AI can do it themselves, or at least the managers and officers can just do it through natural language?
AI is a great tool for engineers but it never replaces real engineers. It may replace just coders who always need direction but not real architects. Engineering is 70% thinking and 30% coding.
What I don't like about these videos are the click baits I hate who gives a fuck what ceo said business people have no credibility what's ever stop being brainwashed from them.
I'm an automation engineer that my work is software-heavy, I dare any kind of AI to replace my work. Without us AE, nobody could make the system run steady on site.
Listen: AI won't kill software engineering - it'll create a new breed of super-developers. Here's what's actually happening.... AI is becoming the most powerful force multiplier creative industries have ever seen. One competent professional who masters AI tools can now do the work of entire junior teams. Think bigger: This isn't just about coding. Every creative field - engineering, marketing, law, design, etc - is experiencing the same shift. The winners won't be those who fear AI, but those who learn to leverage it to amplify their expertise. The real disruption? It's not about replacement - it's about radical efficiency. We're watching the complete restructuring of the talent pyramid, where one AI-equipped senior professional can deliver what used to require 5-10 junior and mid-level staff. Bottom line: The future belongs to those who adapt. Are you going to fear the wave or learn to surf it?
disgusting chatbots of Facebook , and RUclips needs more of software engineer than this video ever predicts. when I putt my effort to write down something messenger, or in general argument on various topics on Facebook main page some irrational word pastas get into existence and make the whole thing obstructed repeatedly. though RUclips has corrected itself a littlest but the problem still remains but the appreciating thing is that when you put some uncommon word in RUclips videos argument section and you are not pretty aware of the spelling of the word the residual spelling showed up in bleak form in the argument sector where you write your opinion down. google is pretty cleverer than meta but never recommends such sinister proclivities of shedding down the importance of job section that is revered by industry pretty often and students haveutlized their whole life of having achieved mastery in it.
Notice big players want mid-level engineers to be gone. The reason is simple: all engineers with less than 10 years of experience are currently in a sweet spot to really disrupt everything these big players have built by using AI. There were significantly more software engineers made in the last decade than in the entire history of the computer science domain. They will do their best to keep mid-level engineers out because they will dominate in the next few decades. Anybody who is using O1 and other SOTA coding models can attest to this: the moment you want the logic to be changed, even in LeetCode problems, models have a hard time explaining the changes, try to be creative around problem solving and see model struggling (forget about large code base f**kups). He just want to decimate small to medium players in the market.
Software engineer of six years here, living in Midwest USA. I have mainly been working for startups and doing my own thing freelancing. But I have yet to make over 100,000 a year. I think the number to hit is less than 100,000 a year not 250 like you mentioned about halfway through the video
Maybe Mr Zucker here doesn't understand how complex writing even a simple program in code is. Software engineering is not going anyway. It will change in some ways but mark my words, coding is not going anyway anytime soon.
Let's not forget that Mark Zuckerburg is the same guy who believed that we'd be living in a Metaverse virtual reality at this point in time
😂😂😂
Some sort of it.
$40b down the tubes so far
How easily some nations can sell themselves.
Next year AI replaces CEO and Mark will start cleaning toilets for living
I am tired of CEOs and AI. Facebook has sucked for a while now. I don't want to listen to this over and over again. As they try to continue to inflate AI and hype there stock price. I am going to continue to work on my engineering skill. CEOs are not engineers
Same brother
Elon Musk is as well as the CEO for Rocket Lab
Yes because they said since 1990 😂. Programming skills will be more needed even in 2040.
AI is a TOOL.
I think this is a PR exercise. They are peddling the narrative that AI is actually replacing engineers or rather increasing their productivity so much that they need fewer of them. Displacing might be the correct term. In order to free up funds for AI investment they need to reduce head count. Either that or they are using AI as a cover for what are likely bog standard layoffs and cost cutting.
Once we have these software engineers working for us, we can do away with Facebook etc and create our own community social media apps.
great take!
I never thought about this
Because honestly
Open source would literally mean easy open source
Distributed ai
Creating social media is not difficult at all... nobody using it is the problem.
Thank God iam a low-level engineer
😂😂😂
😂😂😂
😂😂😂😂😂
😂😂😂😂😂😂
LOL
Says a guy who said we will all be living in metaverse by now . Wake me up when Apple says this
We're in it right now
No one knows how far companies will be able to go cutting humans out of the loop. But does anyone think society would find it acceptable to have somone like Elon Musk or Sam Altman and maybe a dozen investors all own a Trillion Dollar company with zero employees?
Part of owning a business in a captialist society is to provide jobs. If you no longer do that, a new paradigm for purpose, pay and ownership must be determined.
That AI should be seized and owned by the world or the country to benefit all. No a small set of lucky investors who forever get to live off the brilliance of a machine.
I dont believe there are many functional societies that can exists where human work is not a factor.
Yanis Varoufakis might have something say to your comment.
@RohgishSun a universal basic income may be valid. But humans still need purpose. Arguably, it may not need to come from jobs. Not all jobs provide purpose. But if we are not at least partners in the continued growth of humanity, I think that will be spiritually detrimental.
It may all simply be fear mongering that AI will be so perfect that human thinking becomes second class. But time will tell.
@@amesaswi dont know, spirituality has less to do with your work, look at buddha or jesus, they didnt do anything and they didnt need purpose either, they were fulfilled, maybe that kinda pursuit will become mainstream
@Ojas97 purpose is usually derived from personal growth and community. If human knowledge based work losses all value, as society perceives value it will leave a massive void.
It doesn't really matter whether we accept it or not the question is it going to happen we'll see because there's nothing Any of us could do about it except cry And wine a little bit
I still feel like people don't understand the basic problem.
No - AI won't replace you completely (not yet).
However, it will speed up work to such an extent that your company won't need 10 employees, but e.g. 2, so 8 will be unemployed. This is already happening in many industries.
So when they say that "there will be no work for programmers" - nobody means that "programming will disappear", only that there won't be as much demand.
I'd love to come back to this comment in 2-3 years and see if I was right.
@@perrym8048I partly agree with you, except that there's a distinct possibility that AI, unlike email, will also be able to do most of the new jobs it creates.
I love how people smoke crack cocaine and say shit like "AI will increase output."
Thats not how things work. Thats not how anything works. Companies don't "maximize output."
They maximize profits. Doing the same work as a 10 man team with 2? OH YEAH LETS KEEP EVERYONE ELSE ON AND WORK MORE.
We are all so insanely fucked, and this rhetoric prevents us from addressing critcal needs such as UBI and workers protections.
The problem is far worse than that. AI is going to gradually (or not so gradually in some cases) reduce jobs to the equivalent of typing what you want and then following up to be sure that it's done to your specifications. This lowering of the bar is going to average salaries across the board. What's worse is that AI labor is going to be far cheaper than human labor, and it will only ever get cheaper and better, so human labor is going to go down from that as well.
Your debt is going to stay the same though, so good luck paying off that $500k home when you're earning burger flipper salary or less. Banks will fail, the economy will collapse and we'll be in complete chaos. The wealth gap will be an insurmountable hurdle that will only increase as AI improves. This could all be avoided with some government planning, but they aren't even talking about it. All they want to talk about is using AI in warfare and other obvious, but much less likely problems. When anyone in govt. tries to bring it up, they are told "it's too early to talk about that for a tech. that's not even fully realized." By the time it's fully realized it will be too late to come up with a new system.
@@Larry_Dean those companies will be quickly left behind
There’s quite a few contradictions in this assessment- how is stating in black and white that not requiring 10 employees but 2 so 8 will be unemployed not saying that AI is replacing humans?? The SE market is saturated as it is, if there’s less jobs this will be forcing people to seek other careers, often resulting to a huge pay cut. And only those 2 remaining SEs will be altering their jobs to code this AI. I’ve actually not seen a great deal in this space working as an SRE and the automation we do isn’t replacing many jobs but by the sounds of it, AI will and it is terrifying. Maybe it’ll take decades but it will hurt.
Experienced devs understand very well that AI is nowhere near replacing SE.
Really? I thought it's already way better than most average programmers, that, now a senior can effectively do without his team, than waste his time in a useless team and training them, and take care of th people issues.
Yes, now Ai can do a lot of work for you or instead you... actually, there are some days on my work when i just have a chat with the ai agent in the cursor and fix its generated code just a little
ASK an AI to visualize a Cube with OpenGL with 1000 instances and you will See the Code sucks. A Beginner trained with a Tutorial can solve that @@josephs4044
@@josephs4044 Sometimes it speeds up but one person can't have endless responsibilities. "way better" it still doesn't reflect or have new ideas except randomness, it just summarize historical data and it will never be up to date to trends and changes in languages and frameworks. If we can't agree why would we agree about AI result?
@@josephs4044 one day your senior developer was also a useless, untrained, average programmer
When emails came out in the 90’s they said, sending ng letters will be dead. Post offices will close. Postmen will loose their jobs. 30 years down the road, there are less letters sent. But there are still letters sent. Postmen had more job nowadays because of ecommerce. The coding might ending for some developers like UX/UI designers and developers but their will be more demand on a different scale for developers who will know how yo utilize AI. Job landscapes change. But they are still there.
AI and emails are to different things
only fans is booming too even though old way of prostituting isnt over
Volume in the post office is waaaay down. I don’t get anything I actually need in the mail anymore. Just junk mail and flyers I use for kindling.
@@4ka07_muhammadrizky I know. What I am saying is that the job landscapes will change. You need to go back and study history to understand what I am saying.
"Are gonna have", "once you have that", "over time we’ll get to the point where".. it’s all just empty talk. Devin, the AI that’s supposedly replacing software engineers, can’t even commit a single line of code without taking forever and still messing it up. It’s ridiculous. Stop hyping this nonsense like it’s revolutionary when it can’t even handle the basics!
Software engineering is so much more than just slapping some code together. Even writing good code is hard. It’s not about spitting out lines that look fine, it’s about making them actually work in the real world. The current AIs are just throwing out code snippets and crossing their fingers. Replacing engineers? Give me a break. If that ever happened, everything else PMs, CEOs, managers..etc would vanish. Software engineers solve problems, build systems, and keep everything running. Let’s see these so-called ‘AIs’ try to take that on.
Is it really that bad or maybe you are just that good?
I have used Chatgpt to debug issues I find . It does a fair job for the most part but also has hallucinations and gives you solutions that simply don't work at all.
He said, AI will be writing a part of the codes, not the entire programme means we are far from getting full codes written by AI only but can get help or assistance. Technology is not our enemy.
I couid see doctors getting replaced not entirely though or robots hired as we have shortage of doctors in this country .
I recently used ChatGPT and Claude Sonnet to write some fairly simple shell scripts. It did an amazing job and ultimately produced a script that worked. I then found several edge cases that it didn't handle correctly, as well as spotting redundant code, security flaws that would come from input that hadn't been properly cleaned, etc. Without the years I've spent programming, I wouldn't have spotted all those problems. By the time I was done, I felt a weird empty sense of sort-of accomplishment. The actual task of writing the code from scratch would have taken a lot more time, but the process of cut/pasting into ChatGPT and telling it to fix things was downright boring. I don't think I would enjoy being a SE if I were using AI for much more than occasional guidance or as a time saver for actually typing long complex patterns.
Seriously people who believe this , I work for a big company , most senior staff can’t even use excel properly and you think they gonna code complex sql or python , not in a 100 years lol . Pretty sure he is selling a hype to justify billions they spent
They’re not referring to business roles though, this is about reducing expensive SEs??
@@silvershadowjshbut the main reason they're even doing all this is to get trust and funds from non coders. So, if non coder won't be able to use these AI to create their dream facebook, then what's the point? Why invest in it?
Also fools grow old.
Senior engineers don't need to use excel? lol, I haven't touched in years. Can do basic things but if I don't use/need it why would you bother investing time in it
@@danielxbox28it just shows your lack of grip on technology
I burnt out on development, AI has brought me back and I can do much more. It still has its pains, but I am the type that likes to figure out hard things.
enterprise-ai AI fixes this. Zuckerberg on AI coders future.
The Ai people are lying to the CEOs. If you’re into AI, and a software engineer, this sounds , for very obvious reasons, utterly stupid.
Copilot and other ‘AI’ tools have been around, which are only code generators. Too much to go into and not worth the time.
There will be dev jobs for decades. New language versions will be developed. This is all hype. AI needs to generate interest to support the MASSIVE funding. This is sales talk.
Zuckerberg is an AI guy too lol. He the one built Facebook from his barehands as a software developer/engineer.
@ you don’t know what he did. You’re pretty much ignorant. You should probably stick to anime.
@@johnmadsen37your statement by itself is condescending instead of sounding like you’re giving logical and sound advice, which is far from that because, that statement alone is an argument that’s in the same level of a 2nd grade. Your raw intelligence doesn’t match your emotional intelligence, hence why you’re being immature and egotistical.
@@easylife6348 here is another one. Did you think bill gates was coding windows past the first couple versions? Maybe you think the google founders are still writing code for their search engine? It is totally insane to think so. Zuk did what was essentially a prototype not even called Facebook before he got funding and hired real developers. He manages. At a high level. Like most founders. This is comical to explain this.
@@johnmadsen37I didn’t said zuck does AI, use common sense please, I was implying that zuck most likely knows enough about AI to formulate his opinions about AI within the company that he built and manages. All you have to do was asked a question that’s related to my elaborated statement of what I’ve meant about zuck on my first comment instead of you being emotional because you refuse to use your critical thinking skills before ask a rational question about what I really meant by my initial statement.
Why there is the preconception that Zuckerberg has the quality to make such assumptions? He is just rich some kid, who got lucky with his idea, not with his engineering or technology.
Not even his idea #Winklevoss
@ good correction, my bad
Copied idea
There’s a strong economic argument against mass displacement of software jobs by AI. If millions of software engineers lose their jobs, their reduced spending power would cause ripple effects throughout the economy, negatively impacting demand for products and services across industries, including technology companies that rely on consumer spending.
Technology companies thrive on robust consumer markets. If large-scale unemployment leads to decreased spending on products like smartphones(iPhones), electric vehicles(Tesla), travel, and subscription services, it directly undermines the revenue streams of the very companies driving AI adoption. This would also slow innovation and growth, as companies would struggle to justify further investments in new technologies if their customer base is shrinking.
Moreover, layoffs at such a scale could destabilize economies, reduce tax revenues, and create social unrest, all of which would further harm businesses, including those in tech. In this scenario, companies might find that the short-term cost savings from AI automation are outweighed by the long-term economic consequences.
Ultimately, it suggests that while automation can increase efficiency, the balance between leveraging AI and maintaining human employment is crucial. Companies may need to consider reskilling displaced workers, creating complementary roles, and investing in sustainable economic growth to avoid harming their own markets in the long run.
Someone gets it…..
You're absolutely right. In fact, the ripple effects on ALL industries will be unprecedented.. which is why, there has been quite a bit of talk in the recent past about universal basic income.
just another argument why AI utopia will never happen... for me the most convincing argument is principle of responsibility: who will be held responsible when people are killed because AI surgeon malfunctions, sometimes? or AI written program caused train / car / plane crash... as AI pundits proudly states, they themselves don't know how LLM works, for example... if no one can be blamed (with justice) - how to ensure errors don't happen again?... there are other interesting arguments (what will happen if people are not learning skills, thus not being productive anymore - from where AI will take new learning material - as stagnation means game over in economy)... etc... my prediction is that AI bubble burst will be so powerful that world economy could simply break - as people will lose any faith in others, telling them smart words
Facebook's engineers are getting more efficient, but so is the competition. This is a temporary effect. If it could replace software engineers, they would have fired them all.
The later analogy is exactly what we see with transportation infrastructure: the wider we make highways, the more people we have on the highway. Increased efficiency increases use.
Why people are trying to replace SWE in the first place? I'd say it's the one of the hardest desk jobs to replace. It is way easier to replace positions like lawyers or recruiters, but nobody says anything about those jobs. Everybody is trying their best to replace SWE's for some reason.
😂😂😂😂😂
Zuckerberg could step aside and let AI control the company, that would be much better advertisement to AI than his words about replacing SWE
It's actually easier to replace managers and CEOs. All they do is look at excel sheets to make some decision. This can easily be done by AI since AIs are good at number crunching and understanding trends and patterns.
@@realnapster1522 It can already be happening. I can imagine something like "Hello ChatGPT, I am Mark from Meta. What lies should I say to the investors, so that they won't fire me after the metaverse failure? ". And now we have this vid.
AI is useful for performing math it is programmed to use on the data but it still is up to a human to interpret the information to mean anything @realnapster1522
So essentially, all the engineers have been digging their own graves since the beginning.
Sure, only CEO+AI makes sense and rest must be fired. Great mind of CEO and AI is everything that Meta needs
@ no. It means most of the engineers of the previous age would be laid off, a new age of cheaper more efficient engineers would be hired
We are entering an era where the focus is shifting from how to build apps to deciding what apps to build. Let AI do the coding, you decide what to build. Basically you become an architect.
The legit comment that i have come across...Brain to decide what software to build is all what we need than how to build it.
yeah, till the point the system breaks and someone has to get hands dirty fixing it... but without basic skills and knowledge - who would be able to do that? similar to AI driven cars - if everyone would use them, there wouldn't be a driver skilled enough to even turn the steering wheel properly... skills have to be honed day by day, you cannot jump from the throne of architect into sewers of implementation :)
We are far from it.
Doesn't client decide what app to build?
This is like saying excavators made manually digging obsolete. What happens with tools like Ai is everyone is just able to accomplish more with the tools.
Cope
Excavators made much less need for manual digging, and excavators can't replace people in the process as someone drives it, and it is not good for little digging tasks. This isn't really true for an even better AI than today's.
Even the Amish use tractors now.
Problem is that humans can do a lot more than just dig, they can think. Once you’ve automated thinking, there isn’t much more that humans can do 😅
@ Is there anything in your text that makes it clear ‘you’ are a person, not an LLM? Touring test etc… That guy was clever. There’s a reason for why the test was set up the way it was.
He is nothing more than a manipulator, an opportunist who always positions himself on the side that will bring him the most money.
Claiming that AI will replace engineers is merely a tactic to instill fear among professionals, pressuring them to work harder to avoid losing their jobs. This approach is not only low but also outright unethical. However, it’s already expected from him and other self-important CEOs who present themselves as the ultimate authority on everything.
Let me remind you all that it’s not CEOs who build AI or rockets, but highly skilled professionals who have dedicated years of study and effort to achieve their expertise. If anyone is qualified to discuss AI and its potential to replace us, it’s these professionals, not CEOs driven solely by profit.
Best case will be a drastic reduction in number of programmers needed…for sure less than half as many, likely only 10% or 20% needed. Worst case will eventually approach 0, or just a couple of senior programmers per division of large companies. There will always be exceptions, but the average numbers will definitely decline.
Without affordable tools with 100M context token, this will not happen
Developing a complete system with interlocking modules you need a very large context for the tool to understand the whole system
Also for legacy system, you need to reverse engineer the system to define the prompts or context of the system properly, if anyone look at Agile it does not encourage complete system design
Yes, as I said to others before, AI tools have the potential to reduce staff, unfortunately no established or accepted way to utilize AI tools at the moment, some IT shop even discourage AI tools which for me a waste of valuable resource
I really don't know how AI revolution will change IT setup, which is very entrenched in Agile which for me not designed to leverage AI tools
User Stories will never be efficient as prompts because it lack technical details and independent against other User Stories, this must be replaced by some sort of system wide thinking
Good luck to everyone because I see more chaos before things become useful and efficient 😉
first you need a person who will describe the requirements to the AI tool, so these requirements will not conflict with each other. 😅 What they call AI is just a parrot. Can a parrot replace an human? Of cause yes!! 😅🦜
That is not a strong argument, because we will still need 1-2 engineers that take care of guiding AI
@parkerdollar I said reduce staff not totally remove humans in the process
So true
Why does"The End Of Software Engineers" give me a vivid mental image of A Giant Zuckerberg head nested in the primordial soup whilst gazing at that stars
Maybe end of Meta, finally. AI talking to AI
I am a software Engineer and I can really see in the past 1 year AI has really changed the way we work, unless there's an increase in demand for building more, there would be people who will definitely be no more required. All we can do in these times is up-skill yourself and provide true value which is not easily replaceable.
i guess i have to disapoint my friend, who just finished java course
its not that we not gna be coding, its just gna be a higher level version of coding.
same like how most developers don't code in computer language and these days do it in higher level software languages, its just moving the bar more up.
In this case, the AI could also do the "higher" level coding too....
Just like anything in general, take music for example, thousands of people try to become musicians but only a few become popular and rich, take writers for example or scientists, coding will become just like that which was always gonna happen because its just too saturated
yes, just like in the introduction of the clean code book
@@realworldautomation5999Chat GPT 5 will end coding forever 😂😂
He didn't said that. He was just mentioning the use of AI agents. It's just a beginning process. And it will take time. Anyway, human engineers are always required.
I work for a company that uses generative AI. You have no idea how stupid AI is. It came a long way, but it's still stupid as hell. Try it yourself, use langchain and langgraph and find out how it can't do anything slightly complex.
As a professional developer i aleeady spent 1 year learn how to talk to claude. I talk about code part for 10-40 minutes before i let it implement. Then i check the generated code almost each line. If you just let the AI to implement without coding experience it generates bullshits. Any functionality can be implemented so many different ways. What fo you optimize for: cpu usage, disk space, network communication, etc. there are thousands of questions how you structure your data, what you emphasize on ui, how you modularize (scalability) security, etc. Only the not developers believe this what Zuckerberg says
Right. I feel like most of comment section is not SE. They just talk about this while never touched any advanced code in life. And AI is very weak on old technologies. Just few days before, I was arguing with ChatGPT what is better for refresh token in auth system, JWT or Opaque token and ChatGTP had idea of "combine best of both worldw" and give me solution that actually combined worst of both JWT and Opaque in every term, of networking, performance and storage space. Then I pointed it out and was like "Yeah, you are absolutely correct!...."
These guys should be ashamed of themselves. The only thing AI has been able to do is to act as a chatbot. It has not even been able to that up to 30%. Facebook that is full of security flaws even with humans programming it, is talking about replacing those humans with AI. Hahahaha...a little kid will crack down their website when that time comes. There are far more important things that the world needs AI to do for it (if AI actually have the power to do it), instead of taking people's job away from them. Zuckerberg should be ashamed of himself that at his age, he still talk like a kid. In fact all the CEOs thinking about replacing humans with AI. AI can't do shit. Problems like email hacking, business email scams, romance scams, health diseases, etc, has not been able to be solved by AI. Then you are thinking about taking people's job. And look at the way they are even saying it, saying it with so much pride. Common sense they say, is not common. Pathetic.
China's got chipsets designed by AI.
Well ask artists or writers who lost their jobs if they needed somewhere else to orchestrate "AI artists or AI writers"
It's like saying "Don't learn a foreign language, just use a translator(which by the way was developed by programmers)". Now try to translate Dostoyevsky's "Crime and Punishment" with google translate.
That would be something!! Coding will change but never disappear. Zuckerburg is the only one who lives in his Metaverse.
I started leaning Programming not so long before the ChatGPT and AI appeared. I started using them so I can understand some code and learn to get things to work .In my learning journey I am still half way through or maybe a lot more. I would say, I always got frustrated and discouraged to stay and keep learning but I trust some few people in this industry and they seem to imply what you just presented here. A good programmer will be always working and finding oppourtunities and I do believe that. Thanks! This video appears to be fair about the whole AI thing.
What programs are you learning?
@ghassanjneinaty4421 now I am learning OOP C# and SQL. My goal is Mobile Development
Once all the coding is automated I am so excited for the collapse - all of a sudden competent engineers will be in extreme demand- hacks will be prevalent, leaks, etc - also the quality of these products will tank drastically. It will work until the entire backends of these enterprises are legit spaghetti. Then these corporations will be ripe for upset.
Damn.
I've been waiting for offshoring to take my job.
Now i gotta wait for AI too? I'll reset my clock.
AI will streamline projects, lower costs, and make technology more accessible to a wider range of companies and organizations(MORE PROJECTS). In reality, the growing reliance on AI will likely increase the demand for skilled engineers to meet the rising challenges.
However, these claims often come across as exaggerated marketing tactics.Ironically, this type of AI narratives might discourage some people from pursuing careers in computer science due to misconceptions about AI replacing human expertise. We will face challenges in supporting such a high demand for projects.
It’s also important to remember that AI is not a magical solution-it requires vast amounts of high-quality data to function effectively. As technology evolves, AI may struggle to keep pace with entirely new advancements.
Many people do not think of this problem in terms of economics. The principle of economics states that the lower the cost of developing software applications (lower barrier to entry). the more competitive the market will become. The big software companies will start facing enormous number of startups trying to compete with them. Those startups need to hire more software engineers to help them. As a result, fewer developers are needed per company, but more companies are there to need engineers.
yes like in Door dash and Uber 😅
Supply and demand. Price of intelligence goes down. Therefore, in general, it will become ever harder to get a job as a software engineer, amongst other things. Whether we will be witnessing this trend clearly due to AI in 2025 is another thing.
Anyone who wants a job or wants to keep a job will need to be familiar with these systems.
i am 6ft tall so guess i can become a gladiator 🤡🤡
don't worry is not matured at all it will require more enginners to control it and maintain it far more than current software enginners
Or, people expect software engineers to be replaced by AI, therefore they don't study engineering. A few years later AI didn't match the expectations and now you have a shortage of engineers and everyone gets paid more. It's all possible until it actually happens.
The speed of development will increase but the quality requirements will come along. Look at the quality of fb - hard to use it.
I doubt that devs will be massively replaced!
While I hope you're right, I believe that's super wishful thinking. Salesforce just announced they will not be hiring ANY SWEs in 2025 and the City of San Francisco announced they will not be hiring again until June of 2025. I don't see any software company massively scaling up SWEs but there's evidence of many laying off or slowing down their hiring.
As a business owner myself, I'm not hiring new SWE's. I'm looking for architects and communicators. Project managers with enough skills to be able to translate the client's demand for A.I. SWE's. Until the smarter clients instruct the A.I. themselves.
Coincidentally someone from Salesforce is on the Board of Trustees of the World Economic Forum, I guess that prediction of software developers being the #4 fastest growing job area is just for other companies.
@@TwilightThroneall the very best... will you be having arch or pm with ai in your support team as well if you are into real software?
@@stefano94103 salesforce is hiring software engineers on linkedin lol.
@@TwilightThrone but if clients can instruct A.I. directly, then they would skip your business completely
claude is actually intelligent , i was prompting the other day to claude that resources are limited and it in response said that , that is like law of conservation of energy in physics , it can find patterns , also brainstorm well , if it gets better it will actually revolutionalize way we work in all walks of life
You got tricked by an excuse - this is like how dogs can train their owners. It can't complete the task to it tricks you into believing that lol. You got operant conditioned
@@bananaear23 do you think ai is not doing tasks humans throw at it , it is , it can generate images , edit images, and understands language already , also it can code problems that were not in its dataset
@@anandkanade9500 I don't think you know what AI is
@@anandkanade9500 I will give you $10,000 if you can make an AI generate the image of an analog clock indicating 3:35.
since it's so smart, you'd only need 1 or 2 prompts for it .
good luck
I’m ngl it’s useful but it’s not that good lol. It makes tons of mistakes, but it’s like a good google search or tool to jump off and make your own documents and conclusions
Mark isn’t saying jobs will be replaced, he is just saying a lot of code will be written by sort of mid level AI software engineer, you still need a dev to tell the AI what needs to be done, what sort of approaches etc.
If you’re in a faang or some corporate you understand that a lot of that code is already written by previous engineers and some engineers who manage infrastructure and us devs mostly focus on initiatives, delivery, features, processes, discovery etc
I being a software engineer rely heavily on AI to do the labour work so i can focus more on delivery and quality. Listen to NVIDIA CEO, your job will not be replaced by AI but people who use AI.
I agree this industry will only expand more. In my company we’re already using AI to provide more value for customers. i.e. more to do for us
Yeah. You will be leading an orchestra of AI- bots. For 3 years. If you belong to the best 20% who have the most experience. And then AI will do the same. And then you are completely gone unless you are one of the best 1000 in the world. And then you´re gone
my thoughts exactly...
Yup and in the meantime, physical ai embodied in robots will likely become better and cheaper than humans for manual labor.
@@KyleStangline Yup. Agreed.
He also forgets here, that while there will be more software, which can be used for people, LESS actual workers or companies will be needed. You will get your software written on a mediocre level by your own AI.
You will not go to a company which hires programmers. It is like when you can go to Gordon Ramsays restaurant and have the best meal available. Or you just cook at home and use your stove, blender and kitchen machines for that. Very many people will use AI at home.
How many people will by AAA games for example, if they can let their AI programmer program them ps4- level games but exactly as they want them? And improve them. Adjust them. Their OWN games.
And if they want another one, here it goes. And another one. And another one. The Ai just spews games out without an end. The amount of people buying AAA products from companies will go waaaay down.
@@JollyJoe135 Absolutely.
@@Dynamite3783 no that is not the solution. Only part of that I agree with, unionizing is good if for the right reasons and with the right structure and leaders. Unions are not a miracle cure for these issues. We need to become a united coalition of workers who speak with one voice to end suffering using every possible tool including ai.
For example, if the docks can be automated as we see in China, then the dock workers can work a different job in the meantime before that too is automated. If enough jobs get automated then we will all have less work to do and eventually we can reduce mandatory working hours for humans to 0.
Facebook and are creating their competition and downfall.
If ai can replace mid level engineers 👨💻, all engineers will be able to exploit ai and build all kinds of applications.
My dream is that all engineers at Meta submit the notice from work saying "do it with AI then"
Developers still use the terminal a lot, but what if someone build reliable RAG chatbots that can control the terminal? Suddently dummys like me are capable of everything. Today the terminal and coding tools are still a barrier but as soon as that disapears things will evolve very quickly.
This isn't a bombshell, it's expected. We've heard for a while that 2025 would be when they release agents, and we know that Meta has been trailing the pack, so this just confirms what we saw in Dec. when OpenAI showed they have master / grandmaster coder AI. Programmers will still be needed because AI still has problems with problem solving (until maybe r-Star Math is employed, then all bets may be off.) But programmers are going to more and more become just prompt "engineers" and I use the term "engineer" lightly because really it just means typing what you want and then when it comes out wrong, you tell it what you want to change. You know, super hard stuff. Burger flipper salary anyone? Time for a CS degree haha.
I think you're making a huge assumption when you see the future jobs report show that Software Engineers are the 4th fastest growing jobs. You're assuming those will be human positions... If AI is faster, better, cheaper, why wouldn't they be AI?
You basically proved my point when you showed Jevon's paradox. AI labor will be more efficient to use than human labor, therefore there will be more demand for AI labor over human labor.
In the short-term I think you're right. The need for human developers will increase or stay steady, but once AI can do literally every task that a human developer can do (AGI in 2-5yrs) then there's no need for human developers at all, hence why the OpenAI dev said "You'll know when we've achieved AGI internally because our job board will be empty."
lol you sound illiterate, Seriously people who believe this , I work for a big company , most senior staff can’t even use excel properly and you think they gonna code complex sql or python , not in a 100 years lol . Pretty sure he is selling a hype to justify billions they spent
It's not a bombshell because it's utter marketing hype bullshit....and what are you talking about w/r to "this confirms what we saw with OpenAI" when we did in fact didn't see anything from OpenAI or any other of these grifting clowns w/r to them making good on the promises they keep making. Funny how people like you keep repeating this nonsense when every Softwar Engineer who actually knows what they are doing dumped Devin/Co-Pilot/ChapGPT, etc when they found out how absolutely USELESS they are for engineering non-trivial software that people are actually going to use.
@@perdify you're simply ignorant of what AI will be able to do. Do some research before you post. You're thinking of jobs as they are defined today, not as they will be in several years. Jobs in the future will for the most part consist of you telling the computer or android what you want it to do and it does it. If you don't believe me, search around for software creation tools like lovable, or music creation tools like suno to see what will be possible in the future. Yes, those tools aren't great yet, it's not the future yet. But it's a glimpse of where we are heading. Every job, even farmers will be completely transformed. The farmer of the future will be more like Luke Skywalker complaining to Uncle Owen that he has to babysit the moisture evaporators rather than complaining about having to till the field.
@@BruceWayne15325 Jesus , you are still typing essays . Like I said , ai has been here for ages . The only company making a profit from ai is nvidia , which is selling the hype . Stop being too online . Get a job . Ur family did not raise a lifeless nobody fantasizing about ai taking over the world lol
@@BruceWayne15325 lovable and suno are two good examples that they are not superior at all, its just more ways to do it with it's own problems with randomness and hard to get exactly what you want. prompting and natural language will never be exact. at the end of day it's the person using the tools are that are responsible for the outcome.
im working as AWS sollutions architect and these days our team write all the lambda code with AI.
Sollutions???
Seen this for every occupation that has posited the "oh, everyone will be so much more productive" angle. First, WEF looks at current trends and projects into the future, they do not rely on potentially disruptive technology as they do not know when it will show up, and which area it will target, first. Open AI is hiring the best of the best as that is the bar to develop the ultimate ASI platform. That is not a broad trend, that is a specific use case. It would be like looking at Los Alamos while they are developing the A Bom and saying, "wow, there sure is a big demand for nuclear scientists". First, very few could compete at that level, and second, there is a very small pool of demand for that kind of skill. If average intelligence in society is around 110, (including many CEOs which I would say are mostly grouped around 115-120) then exactly what creativity is going to be unleashed by a super intelligence clocking in at around 400 or so? And remember once this is achieved it becomes replicated millions of times so as to be ubiquitous in society. It may take 5 - 10 years to see this happen, but humans are very bad at following exponential trend lines, and AI dev is on such a trajectory. I think it is putting your head in the sand to say "well, maybe it won't be that bad", or "well, I am really, really clever and smart so I won't be impacted". It is time for people to pivot to making demands on what a post labor society looks like before so much wealth and power accumulates into Silicon Valley that we will be living their version of utopia. And that version will have been built while we were still trying to figure out how successful they are at developing AI. It is coming, it is just a matter of time..
Yeah narcissistic overestimations of own abilities meetings exponentials. Will be interesting. Some rage foreseeably.
Seriously people who believe this , I work for a big company , most senior staff can’t even use excel properly and you think they gonna code complex sql or python , not in a 100 years lol . Pretty sure he is selling a hype to justify billions they spent
“Past labor society” 😂😂😂. You people are jokes
@ well nice thing is we won’t have to wait long to find out who is right.
Not end of software engineers. End of coding as we were used to it. Software engineers will get a more high specter of projects they are working on. I am from financial tech industry.. And believe me.. For the next 15 years, you DON'T want AI to get your card data.
Will it be the software engineers orchestrating the AI agents or their project managers? In my experience, most engineers do not really understand how a product is meant to come together they only are able to complete tasks they are told to complete. The PMs always have the big picture in mind. Cannot wait for the day the SCRUM Master leading a sprint retrospective with a bunch of AI agents lol.
Seriously though, I think the PMs may be the winners at the end of the day as the world of AI unfolds. There will be winners and losers, no way around that.
At first you will need project managers / architects. You need to know enough about software to be able to instruct A.I.'s. But at some point the client will talk directly to the A.I. You can already see that in say, translation projects. The clients directly talk to the A.I. to get the translations without the need of a human translator. And TBH, software engineers are really people who can translate the wishes of the client (or project manager for some teams) to the computer.
It'll be the product managers who translate user needs into requirements managing the AI software engineers. Software engineers who haven't developed product management skills should start now. The actual skill of software engineering is going to be less and less needed.
Possibly the worst take I've ever seen in my life 😂 Software Engineers defo don't just complete tasks they're told to do - they constantly have to think about the big picture and possible impacts and they help break things down into tasks (as a team). Good luck to these mighty PMs who undervalue their teams thinking they're the only ones holding everything together. Sounds more like an autocratic totalitarian than a leader that facilitates and enables the team.
Also when AI is doing all the technical tasks - Agile probably isn't needed since business can have everything, all at once, right now so you can probably scrap the Scrum master role from asking what the AI did yesterday and what it's going to do today 😂😭
@@jonblack7046 only the more senior and more intelligent do. And in many projects, that don't have sufficient information to understand the big picture: they don't talk directly to the product owner and stakeholders. Most engineers have become engineers because they like to engineer and they have a natural resistance to architect and communicate. Systems like SCRUM have come about to encourage such things as normal part of engineering. It is like giving a bricklayer a blueprint of the building: they will start laying bricks, they don't question why the building is being built and if the spaces are big enough for the crowds, but the good ones do alert the architect that they think the foundations won't bear the load. But once the A.I. can lay bricks, they will lose their jobs unless they become architects.
@@jonblack7046 Maybe, but this has not been my experience. The most experience & senior developers might have that level of competence, but there are a LOT of devs out there who can solve problems, but never seem to be able to identify the problem/need on their own. That is why so many ERP and LOB applications suck for the end user.
As far as SCRUM is concerned, I was mostly joking. I do see that level of PM changing significantly if not disappearing entirely; however, I do not see the Agile development routine going anywhere. Clients are still normally trying to figure out what they need/want throughout the entire development process and product requirements are always very ambiguous (especially early on). What I see happening is the lengths of sprints being reduce significantly (maybe 3-4 days at most if you factor in a QA period) and maybe future projects take on a more hybrid approach.
SWE should just focus on problem solving, and when lower level problems are solved by AI, it just means human can focus on solving the bigger ones... same as every other tech in human history.
Even now the developers are relying on Ai generated code. If u are relying on Ai generated code, you should know you will get replaced one day. Unlike other tools, Ai can do certain level of decision making and some level of learning process. If we previously need 10 developers to write and maintain the project, now we might only need one developer to review the ai generated code. Even now some Ai tool can review the code and fix the bug by themselves. The possibility is already there and it just need a bit more time to fully replace the human software developers in the future in some areas. But we can't reject Ai obviously. We just need to find a way to work with it.
Humans will just be bottlenecks to ai.
A ton use it for boilerplate stuff, but that's grunt work. It would be soooo slow if you had a non developer continually copy/pasting the entire code base into some LLM whenever you wanted to do something. Not to mention how error prone that would be. You think this scenario is something multi billion or even trillion dollar companies where the smallest mistakes in code can result in millions of loss in profits are going to seriously consider anytime soon?
I only use it for iterative works and some unit tests but that’s it. It has improved my productivity by 2-30 percent when it comes to coding but AI generated code is still shitty.
These AI-hyped people say that one person can do 10 people’s work but I think that’s bullshit. Maybe yes in the future, but absolutely not now
As experienced engineer, I don't think it is that easy, hard part of my work is not writing code, it is to find out where code broken
mid-level engineer in 2025? brother we do not have intern level engineer in 2024, aren't we like... skipping some steps here?
Yeah. The big companies have the ability to make most out of these technologies.
See eg, Google and the extent of their AI deployment with coding.
I just wonder how many years a typical student of the next generation would have to spend in college before he/she can be accepted into a firm.
Or the next generation SWEs will simply build AI based solutions and make themselves more productive using AI based tools.
Nobody said they will need to work for Zucky whose goal is obviously to get rid of all of them.
There is so much work ahead solely in making existing systems compatible with LLM based agents, designing and integrating new work flows with those agents, et cetera.
Welp there goes my 100k degree :/
Join the club!
The other day, I read this on RUclips
"Never take a CEO's word at face value. There is always an ulterior motive behind the things they say. It's not necessarily intentionally sinister, but the reality is that it's in their best interest to guide the masses (that's you and me) toward the kind of things that benefit them and their companies. I have nothing against individual CEOs personally, but they are the public-facing corporate voice of their companies. Even if a CEO made an absolutely factual statement, such as "the sun is hot," I would still have doubts about their motives"
This development is inevitable. Once you get to the point where you can simply ask a question to the ASI or Machine Hyper Intelligence in whatever language you happen to know and it gives the answer in the same. Then it is time to face the music...
It's been at that point.
Not if it starts to consider your existence a redundancy / waste of Earth's resource.
Agreed. Reminds me of the old days, when there were "code generators", for certain types of apps. But AI will make this more flexible. SWE's as orchestrators of these resources.
But for AI products, one now needs to be more than a SWE. It's simple enough to glue a bunch of things together through LangChain or similar. But the main issues regarding AI, esp. agents, workflows, is reliability and accuracy. One also needs to be a ML engineer. How do you set up test cases for proper evaluation? How do you tweak it in fine tuning, ICL, or some other technique, or specialized training regimes, to make sure the product is solid enough to be fielded? How do you utilize user feedback (which could be additional training data)? How does one manage iterative improvements and releases?
This is the same ding dong that said metaverse was gonnna also blow up..
This AI Software Engineers will be more costly than hiring human software engineers. This companies that are driving this forward are those that would benefit more in terms of profit. I remember when they said that cloud servers are cheaper but right now it actually cost so much more than owning physical servers.
If you have AI and robotics that you could rely upon, then...Yeah. End of almost every type of work. In the professional domain, it's even easier to say so. Because it's easy for AIs and their physical presence i.e. humanoid robotics to do so. Thanks.
If not, then...Who are you... hunh.? Nobody, if you only have the education. Not even Nothing, if you don't even have a quality education. And, you are something..If you have got money.
With out the knowledge of programming, you can’t even prompt the AI to code. I would say the programmers will be programming 100 times faster with AI. I use AI to code crap ton of stuff, but I review every response I get before I implement the code. I would say Ai get about 30% of the time get it right first time, but again it depends how complex of a task I submit. If I give you all the carpentry tools that will not make you a carpenter.
Looking forward to the AI slopped being served on Facebook in 2025, its already great but boy oh boy.
Dropping another bombshell...we can all now generate our own proprietary code and don't need to purchase anything.
First of all, AI don't create anything, just copy humans code, and don't pay the copyrights.
Yes, what does mid level engineer do, just go to stack overflow or documentation, and implement.
@@DreaMagnifier The issue is that there won't be high level engineer without mid level ones.
@myasgabut not under the control of a monopoly.
He is not wrong, l can ask copilot to write a C# code for a library systems. Run the code in visual studio to test all in 2 minutes.
Let's not 'sugar-coat' this issue.. There will be far fewer SW engineers hired as AI becomes more efficient. In the scenarios described in this video, 1 human SW engineer can support 3-5 AI SW engineer.
Take Zuckerberg at his word..
Bro, you forgot about energy
The way I see it is that in the future, developers will also be PMs as I'm sure there will be AI tools (if they not exist already) for this as well. Also, the company cost will be reduced (you are gonna get the same pay for doing 2 jobs) and communication with the client will improve, avoiding the back and forth between Dev - PM, because who better than the same person working on the project to communicate the real scope of it, which is an area that PMs with no Dev background struggle with. So, I agree with what's being said on the video: we devs have to upskill in order to orchestrate all these AI tools
14:13 - This sounds like BS to me since obviously the other type of AI's built to debug code (even long and complex code) would do that automatically, if not now or in the next few years then later on at some point, it's only a matter of time, the jobs of everyone would be gone eventually and the only main jobs that would remain are those required to push the AI field and even those would eventually get replaced by AGI/ASI later on.
As a Senior Software Engineer I use o1 for work. Without guidance it ain't shit. Of course it can resolve or give suggestions. But everything needs to be monitored, though hallucinations are less frequent in comparison with other models. It eases my job, less google search, making me more effective. I rarely write code without AI. It's a no brainer to not use it.
*bro you seriously missed the boat on this one* 😂😂😂 doh
Yea, I am thankful that our AI can fix my misspellings.
As an architect, my software development productivity has increased 5000% in 12 months. I can write fairly complex Django apps without being a wizard at Python or Django - just need to know enough to put small corrections into the design. Same is true for AWS and Flutter/Dart.
Opensource contributors are being taken advantage of by big companies. For example, Meta started the Llama as Open source, but who benefited from that? Do they give compensation to those contributors?
Software developers will delete models before they got fired :) They won;t be cooking themself :P
Great channel, love the content
This is how things arr gonna roll . 1 massive job displacement, unemployment reaching to near 30 -40 percent in the next 5 years . 2. Money will be printed and UBI will be established . 3. The inflationary effect of money printing will be offset by the deflationary effects of massive unemployment and rapid advancements in automation which is inherently deflationary. 4. The standard of living will be greatly diminished as UBI will only provide the bare minimum to survive.5 . Overtime with mass adoption of AI , the standard living for all will be greatly improved
🧂
Hope it comes true
Developers Developers Developers Developers ...(cit.)
The assertion that Software Engineers only write code is false. It will take much longer to get rid of us than some new AI code generating agent.
Its not a good thing to be ignorant to the truth. It wont help you this time around. What you do is not that complex or nuanced. Just like everyone elses jobs. You will be replaced soon along with almost everyone else.
@@Dinofx_jr Then why aren't you out there right now using AI to build the most advanced _insert whatever system here_ and raking in the cash?
The one being ignorant of the truth is you.
Will AI eventually replace a large part of what programmers do? Yes.
But its not gonna be 2025.
But you believe what you want to believe. Come back to this post in a year and see how you did.
@Dinofx_jr You don't read very well. Read what I said again.
@@IyamwhoIyam no you dont read very well. What you do is not nuanced. Ai code generating agents are all thats needed. If not replaced itll still be 1% of what exists now
@Dinofx_jr You don't read very well. And you don't know anything about the industry. I can tell by your attitude and response. Perhaps you need counseling. Get some help.
If you're saying I'm going to be managing AI agents instead of being down in the weeds coding, okay. But isn't that going to drive down my and other SWEs' salaries because now our jobs are doable by a wider group of people? In other words, could it be that we remain employed but not necessarily thriving?
You wont be employed. I am an SWE and i switched to trades… you wont be overseeing any AI. The SWE workforce will be 1% of what it is now (population wise). In no more than 10 years.
Correct
Or you will have to understand how the LLMs work together and how they integrate together. Another level of sw engineer.
Bingo
No, you won't be managing AI agents because other specialized AI agents will do that. And in the end, a single agent would do everything.
Why do they even need you to manage the AI agents if the AI can do it themselves, or at least the managers and officers can just do it through natural language?
AI is a great tool for engineers but it never replaces real engineers. It may replace just coders who always need direction but not real architects. Engineering is 70% thinking and 30% coding.
What I don't like about these videos are the click baits I hate who gives a fuck what ceo said business people have no credibility what's ever stop being brainwashed from them.
This wakes the young revolutionary marxist out of me. God damn it.
I ve just lerned about crewAI and .... I want to thank you. You ve got my subscribe!
I just hope one day humanity take a courage then to stop using your trash, then you can recruit AI to use your trash to have profits.
Coding might not be required, but how code works for the business and it fits like a piece in a puzzle, for that humans will always be required.
I'm an automation engineer that my work is software-heavy, I dare any kind of AI to replace my work. Without us AE, nobody could make the system run steady on site.
I never understood that Facebook is a product! Stop using it and show suckerburg the power of rejection by mass.
Listen: AI won't kill software engineering - it'll create a new breed of super-developers. Here's what's actually happening....
AI is becoming the most powerful force multiplier creative industries have ever seen. One competent professional who masters AI tools can now do the work of entire junior teams.
Think bigger: This isn't just about coding. Every creative field - engineering, marketing, law, design, etc - is experiencing the same shift. The winners won't be those who fear AI, but those who learn to leverage it to amplify their expertise.
The real disruption? It's not about replacement - it's about radical efficiency. We're watching the complete restructuring of the talent pyramid, where one AI-equipped senior professional can deliver what used to require 5-10 junior and mid-level staff.
Bottom line: The future belongs to those who adapt. Are you going to fear the wave or learn to surf it?
disgusting chatbots of Facebook , and RUclips needs more of software engineer than this video ever predicts. when I putt my effort to write down something messenger, or in general argument on various topics on Facebook main page some irrational word pastas get into existence and make the whole thing obstructed repeatedly. though RUclips has corrected itself a littlest but the problem still remains but the appreciating thing is that when you put some uncommon word in RUclips videos argument section and you are not pretty aware of the spelling of the word the residual spelling showed up in bleak form in the argument sector where you write your opinion down. google is pretty cleverer than meta but never recommends such sinister proclivities of shedding down the importance of job section that is revered by industry pretty often and students haveutlized their whole life of having achieved mastery in it.
Notice big players want mid-level engineers to be gone. The reason is simple: all engineers with less than 10 years of experience are currently in a sweet spot to really disrupt everything these big players have built by using AI.
There were significantly more software engineers made in the last decade than in the entire history of the computer science domain. They will do their best to keep mid-level engineers out because they will dominate in the next few decades. Anybody who is using O1 and other SOTA coding models can attest to this: the moment you want the logic to be changed, even in LeetCode problems, models have a hard time explaining the changes, try to be creative around problem solving and see model struggling (forget about large code base f**kups). He just want to decimate small to medium players in the market.
Humans programmers are much more ecological for the enviroments than AI that needs huge quantities of energy.
Software engineer of six years here, living in Midwest USA. I have mainly been working for startups and doing my own thing freelancing. But I have yet to make over 100,000 a year. I think the number to hit is less than 100,000 a year not 250 like you mentioned about halfway through the video
Maybe Mr Zucker here doesn't understand how complex writing even a simple program in code is. Software engineering is not going anyway. It will change in some ways but mark my words, coding is not going anyway anytime soon.
It will be worth sharing your information sources, how you got these graphs, and the info you're sharing.
The Jevon's paradox is not aptly applied here, because it doesn't mean more human coders, it means more AI coders.