[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


Let's be honest: AI will leave 99% of tech workers jobless, will it not?
>inb4 it can't replace my job now / muh requirements
Do you think it will not be significantly better in the next 10 years? And even if your job is safe, maybe you are one of the lucky ones, and you won't lose your job. But be honest: does this apply to the overwhelming majority of tech workers?
>But muh embedded / physical world stuff
Again: this doesn't apply to 99.999999999999999999% of tech workers.
>>
>>109596954
yes.
in 5-10 years, realistically
>>
>>109596954
Far Cry soundtrack said it best
https://www.youtube.com/watch?v=KW5kcUUu2Ac
>>
>>109596954
99% of tech workers are code monkeys building corporate shovelware relying on highly abstracted libraries, so yes. Most don't even code anymore, they're babysitting AI agents while paying to give away training data.
>>
>>109596954
I do find it a bit ironic how so many tech people were laughing at artists in 2023-2024, yet tech workers are the people getting dabbed on today.
>>
>>109596954
I thought covid proved we could get rid of 99.9% of jobs if we wanted but we'd rather have corporate daycare so we all went back
>>
File: clin-d-oeil-2.jpg (60 KB, 800x534)
60 KB JPG
>>109597118
corporate daycare exists to prevent people from becoming competitors
>>
>>109597118
No, COVID just proved that most office jobs didn't physically require people to be on-location, since the only time their physical presence was actually "needed" was for the odd meeting here and there. And even then, that wasn't really the case.
>>
The big bet is where in the S curve on the LLMs we are. The models have barely gotten better in the past 6 months, in some ways its been a regression and a lot of people are going back to older versions to cut out pointless wordiness and bullshit, it's mostly all the agentic harness stuff around it that has improved how to guide the LLMs to get shit done.
>>
LLMs can't think, innovate or create. They are fancy search engines. But yes if anything it proves that 99% of "tech" people shouldn't be employed because they have no business working in it.
>>
File: 1743166055821255.jpg (31 KB, 960x850)
31 KB JPG
>>109597546
This is my favorite cope.
>>
>>109598034
Absolute majority of people in any profession is consisting of not innovators but just people doing the chores and mundane work. If every white-collar professional was truly innovative, we would be the type 3 Kardashiev civilization already
>>
File: 1708454553432.jpg (61 KB, 526x526)
61 KB JPG
>>109596954
>>
>>109598097
Then I don't really see why they should be kept around.
>>
>>109598151
Because you don't want to live in a South African scenario where 30% of people are unemployed and ready to shank you for looking like you may have $50 in your pockets.
>>
File: 1773935833575635.jpg (398 KB, 1024x588)
398 KB JPG
>>109598141
Bluecollar chuds think this is an own on the whitecollar wagies, but its really just meant to demoralised them into finally taking a bluecollar job. Helping drive down labour job wages and give unions more money from membership dues.
Its mainly the /biz/ chads who win again
>>
>>109596954
I don't know, but what I do know is that most people are fucking retarded, using AI or not, especially so middle management.
>>
>>109596954
not my problem, i'll be retired early in another 5 years. if that turns into 10 because i have to work at a grocery store, i'm fine with that honestly.
>>
>>109596954
It most likely won't, but any opinion that deviates from the hype is dismissed as emotional cope by you retards, so it's not a conversation worth having
>>
>>109598500
Explain why it most likely won't happen. Coding agents *today* can do a better job and are faster than 99% of tech workers (and not just programmers: this goes from help desk, to sysadmin, to infrastructure work, to cybersecurity etc.). Let's be objective about this.

If you make a coherent argument, I won't dismiss it as cope. I just can't see, given what I said above, how total tech worker unemployment is evitable.
>>
>>109596954
>Do you think it will not be significantly better in the next 10 years?
Yes, I do believe that it won't be significantly better.
>>
>>109599100
How so?
>>
>>109597546
I want this to be true anon, but this is just cope. The models can replace most tech workers already. Even if growth slows down, tech workers are fucked anyway.
>>
>>109596954
>AI will leave 99% of tech workers jobless
all you niggers deserve to be jobless fighting over scraps for my enjoyment
>>
>>109596954
web dev -> ai can build a production ready site in minutes
data science -> ai can analyze all the data, and generate complex models using that data
devops -> ai can take care of all your cycle without batting an eye
and so on

do you want a job? learn electrical engineering
>>
>>109599125
right now our access to this technology is propped up by unsustainable levels of venture capital, bailouts, and a bizarre game of musical chairs.
if you remember the minor AI winter that happened between 2020 - 2022, it can happen again. In 2019-2020 the public had a taste of frontier models via AI Dungeon ...etc. This went away when the money dried up and it didn't resolve until ChatGPT came out to kick off the bubble in full force.
the same way movies haven't gotten cheaper after moviepass went under, there's no reason to assume we'll have a real paradigm shift.
>>
>>109599228
Yes, I agree with you. But don't you think in 10 to 15 years they'll find a way to make it more economically viable? Maybe some hardware progress, maybe they figure out how to make these models run on less power etc.
After all, the most rich and powerful corporations on Earth are spending literally billions upon billions of dollars on this thing and they desperately want to automate tech job. How could you even compete with them?

Even if the prices skyrocket and you have to pay (say) 1000$ a month for this tech, it would still destroy the tech job market, because it's much less than the average tech salary and it's much faster than human workers (also, AI doesn't take breaks, can work 24/7 etc.)
>>
>>109596954
anyone who isn't a retarded brainwashed leftitst can see the trajectory
the problem is what will happen after
people call me a schizo but I'm certain that they are going to kill everyone
>>
>>109599416
I don't think you are schizo at all anon. After all, why are they spending astronomical amounts of money on this tech? There HAVE TO be ulterior motives in their agenda.
>>
>>109596954
2 more time units
>>
>>109596954
Tech is the first sector, but everyone is going to feel the pain here. If you're not already rich, and you don't have the top .1% of looks or charisma... it's over.
>>
>>109596954
All tech?
fuck no
Code monkeys? Yes
There's no fucking intelligent CTO that will give AI complete control to infrastructure
>>
>>109599086
The entire premise of your argument is wrong. Vibe coding doesn't work that well. It's good for shitting out demos and for simple python automation scripts, but it quickly breaks down for any serious project.

To make anything decent you need humans which use LLMs as a tool for obtaining information and generating small amounts of code at a time. This workflow, when done sensibly, is not much different from the old one with the use of google, stack overflow and docs. Many companies are pushing their developers to rely on LLMs too much out of FOMO, but that will eventually pass for the most part once the hype does.

You could argue that it's going to get good enough, but I don't see that happening with transformer based LLMs. Their intelligence is "jagged" in the sense that for some things they are superhuman, while for others they are completely retarded.

Their improvement is also not happening for everything at the same rate, and for some things they are simply stuck. One example of something they suck at is software architecture. You can look at the scientific literature on this, you don't have to believe me.

You could argue that a new powerful architecture will come along in a few years, and while that is possible, it's a baseless guess and no one can really predict that.
>>
>>109599217
The only reason electrical engineering (i.e. digital design, EDA, PCB design, etc...) looks safer is that there are much less available data due to the proprietary and fragmented ecosystem. But it's only about time.
>>
>>109599546
Fair enough, thank you for the answer. I didn't know there was scientific literature about their inability to do code architecture or similar. I will look into it.
>>
>>109599546
Just for curiosity (I ask you since you seem to be informed on this topic): is there anything else where there is scientific research showing they are stuck at something (other than code architecture of course)?
>>
>>109596954
Even without AI, 90% of "software engineers" are redundant.
>>
>>109599217
>learn electrical engineering

Honestly, this is bad advice. The demand for EE in the labor market is quite low. Meanwhile, while you're studying EE for four years, new AI agents and new models will emerge that will likely be just as good at EE as they are at coding now. I don't think a university education makes any sense right now. By the time you're studying, your profession will disappear due to AI. It's better to study to be an electrician in vocational school, not electrical engineering.
>>
File: 1786747599184919.jpg (54 KB, 1024x770)
54 KB JPG
>>109597097
>so many
No they weren't. Just because you're an emotionally disregulated monkey that gets ragemated by the small but loud minority does not mean everyone hates you.
>>
>>109599371
>some hardware progress
Hardware has barely progressed in recent years. They've made minor changes, but believe me, chips from 2026 aren't much different from chips from 2016. Whereas between 2006 and 2016, there was much, much more progress, as we saw in PC and server hardware. Not to mention 1996-2006.
>>
Yeah, the only skill that matters now is taste. What's taste? Uh, you know, variable naming, data structure and algorithm knowledge, design patterns, API design, ummmmm database efficiency uhhh
>>
>>109599086
All white-collar jobs are already dead because all that work (not just coding, unfortunately) can be done by an AI agent.
>>
File: 1763208921784110.jpg (217 KB, 1536x1920)
217 KB JPG
>>109604386
>emerge

Stopped reading here. Reddit is down the hallway sir
>>
>>109604415
And what's the solution? UBI?
>>
>>109599450
the ulterior motive of making money, yes
>>
>>109599546
Scientific literature is a meme when it comes to assessing AI capabilities, it always lags by a year and reviews older models.

METR benchmark results keeps going up and there is no reason to assume that it will stop anytime soon.
>>
>>109597478
What if all the remote workers were actually disappeared years ago by the gov and a governmental conspiracy has AI avatars that look like the disappeared remote workers on their zoom calls?

Its be the ultimate silent takeover.
>>
>>109604600
You guys have to stop the obsession with money. It is not about the money for them, stop being so stupid. THEY ARE THE ONES THAT MADE MONEY IN THE FIRST PLACE WAKE UP IT'S JUST A MEANS TO AN END
Marxists are doomed because they are OBSESSED with the material world
>>
>>109604833
I assure you that Anthropic or OpenAI did not invent money.
>>
>>109596954
Look up GEN-1.5, nontech workers are fucked too. We're a few years away at most from robots that can operate in the real world.

NEETs won, wages lost
>>
>>109604600
They are burning money at an astronomical rate. None of the companies doing something adjacent to AI are making any money except Nvidia and other hardware manufacturers.
>>
>>109604720
This. Also, the methodology is mostly retarded and the sample size is always tiny.
>>
>>109604994
Every business burns money until it doesn't. The revenues keep going up, that's the thing that matters
>>
>>109604994
data plane companies sure are!
>>
>>109605132
>Every business burns trillions of dollars, until it doesn't
>The revenue keeps going up, but the costs remain constant bro!
>>
>>109596954
Why should I have a positive view on AI?
LLMs won't lead to AGI and even if they did the AGI would be misaligned.

LLMs as they are now increase the workload on senior programmers and reduce the quality of the code they produce but worse still they stop new programmers from properly learning and if we don't progress from LLMs the current stage continues for a significant time we will have stopped intergenerational transmission and that's civilization ending.
>>
>>109605532
According to some sources Anthropic is already profitable. In any case, the expense:revenue ratio is not huge

https://www.forbes.com/sites/jonmarkman/2026/08/17/anthropics-groundbreaking-second-quarter-delivers-115b-in-revenue/



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.