Thread #108677296
File: AILOSE.png (77 KB)
77 KB PNG
Why did AI lose?
66 RepliesView Thread
>>
>>
>>
>>
>>108677306
>meanwhile, in reality...
https://www.theverge.com/news/909640/microsoft-removing-copilot-window s-11-buttons
>>
>>
>>
>>
>>
>>
>>108677306
>cuz of ai
cuz ai does the work correctly and they don't need people?
or cuz they went all in on ai and now they can't afford people?
in my company, we just add an all hands about how "maybe we should be a bit more selective" about spending tokens
this is after 2 months of aggressive "IF YOU'RE NOT SPENDING $500 A WEEK ON TOKENS THEN YOU ARE NOT WORKING AT YOUR FULL POTENTIAL!"
>>
>>108677296
Over leveraged.
It also wont understand things for you. People who understand things and have good abstraction and systems level thinking are having an explosion in productivity, even if they're not tech workers or are familiar with it to begin with. You also need a high level of responsibility and attention to detail to catch its hallucinations, its bullshit, and manage the mass of content it can produce.
Business wise, the pilots would go well if they didn't overinvest. Just consider a company gently pulling in a non-edge, slightly outdated GPT API, just telling their employees "hey, its here, infinite tokens for general use case. If you need anything in particular or more specialized for your work, reach out to our tiny AI department embedded in software or IT"
Then if you want to be certified to use the API, just do a small 20 minute corporate training module explaining pitfalls, a small quiz, and a waiver that you will verify all outputs, can't say "the AI did it1!!!" etc etc. That's it. So simple and cheap.
>>
>>
>>108677348
>>108677418
>>108677440
>>108677525
you guys are going to have a really rough year
git gud at tardwrangling the bots or be part of the 2/3rd of staff who get laid off
>>
File: 1771612743990.jpg (14.7 KB)
14.7 KB JPG
>>108677510
Moore's law mindfucked two generations into thinking that everything limited by computing power would grow exponentially. Plus all of the dumb crypto money that has stagnated in the last 5 years is desperately looking for the next big thing and went all in on AI without realizing all of the fast achievements have already been made and it's already plateauing.
>>
>>
>>108677306
>cuz of ai
It's not because AI is doing the programming, though. It's because all the capital being funneled into FAGMAN during covid and the Biden years, which triggered an absurd hiring buble, is now diverted into winning the AI race. So all the project zoomettes who got hired to make tiktok and instagram videos about eating snacks all day and doing no real work at SV Tech Companies got laid off. A lot of the engineers working on totally irrelevant side projects with no effect on the bottom line also got the axe. The whole thing was kicked off by Elon firing the 80% of pure bloat at Twatter.
>>
>>
>>108677589
>git gud at tardwrangling the bots or be part of the 2/3rd of staff who get laid off
I can git gut while simultaneously believing leadership is aimless and they're making a huge mistake
and if I get fired because the company fails or because ai replaces even the tard wranglers, it's still the same to me
>>
>>
>>
File: 1726846184954740.png (330.4 KB)
330.4 KB PNG
>>108677296
>new study on AI
>look inside
>it's Llama 3 or GPT 3.5
>>
>>108677296
In medieval times, they thought the human mind was a book.
In the industrial age they thought it was a machine
Now they think its a computer.
Science, maths and physics are predictive models for the real world. We have no fucking idea how the mind works, so we are using very advanced statistics to PREDICT what an inteligence would produce without understanding the underlying. Imagine calculating a satelite trajectory without understanding the physics behind it, or even concepts of velocity and weight. That is what AIs are and why they are fundamentally limited.
>>
>>
>>
>>
File: 1769444899576690.jpg (29.9 KB)
29.9 KB JPG
>>108677745
Please share your bespoke, meaningless benchmarks with us. They help your case a lot.
>>
File: crying long face red eyes soyboy soyjak tagme-5.png (522.1 KB)
522.1 KB PNG
>y did firs gen GAYi flop, duh chuds wer telin me itd take ova duh werld
>>
>>
>>108677306
Good this means the 95% of failing startups will have a lot of good employees on the market
OH WAIT THEY CAN'T BECAUSE THOSE EMPLOYEES ALL SIGNED NON-COMPETES AND HAVE TO SIT FOR LIKE A YEAR OR TWO
AMERICA LAND OF THE FREE
>>
>>
>>
>>
>>108677296
>The data also reveals a misalignment in resource allocation. More than half of generative AI budgets are devoted to sales and marketing tools, yet MIT found the biggest ROI in back-office automation—eliminating business process outsourcing, cutting external agency costs, and streamlining operations.
So, successful AI usage is mostly using it to mop up behind the scenes and get the office in order. It's low hanging fruit. If you had some competent programmers, office managers or a consulting firm, you could have done the same thing years ago.
>How companies adopt AI is crucial. Purchasing AI tools from specialized vendors and building partnerships succeed about 67% of the time, while internal builds succeed only one-third as often.
Custom made and narrowly focused AI usage is yielding success, which is what you would expect from experts using software to solve a problem. The generic chatbots seem like they can do anything and most people don't understand their limitations yet, which is probably why there is so much failure. However, if AI becomes and expert in using AI, then things will really start to get weird.
I think in time, AI is going to be oriented towards off loading AI workloads on to regular software and custom neural networks. The price of sending questions to the huge data center based AI's will increase, but AI and it's users will be experienced enough to pick the tools for the job instead of burning gigawatts of energy. If you want a solution from the big brain, you'll be working with a smaller local AI that knows what to hand the big brain. You and the local bot will flesh out the details and the pricey round trips to the big brain will be at a minimum. It's like using a contractor or other outsourced solution provider.
>>
>>
>>
>>
>>
>>
File: f607fd5649c6301388782d83378381d4[1].jpg (25.4 KB)
25.4 KB JPG
>>108677296
Did you know high-speed photography was invented at MIT? By Harold “Doc” Edgerton.
It allowed the capture of images like a bubble bursting.
>>
File: 78A7BF30-FCB7-446B-8840-5E5539C7B1ED.jpg (1.3 MB)
1.3 MB JPG
>>108677296
Luddite cope thread
>>
>>
File: disco_record_sales.jpg (83.4 KB)
83.4 KB JPG
>>108677598
>>
>>
>>108677306
They aren't firing people because AI is replacing them; they're firing them because AI is failing and they can't afford to keep them on. In a sense they're firing them "because" of AI, but not in the way you're trying to get at.
>>
Wow companies are like states now. They make stupid business decisions and then just do a woopsie, lay off everyone and can continue making retarded business decisions until the next round. Basically too big to fail are slowly failing while startups are just crashing to the ground
CEOs that drank the AI koolaid (mine still does) are just going full on delusional and think some verbose BS is going to save them. LMAO.
But at the end of the day it's us all with our jobs suffering retarded and greedy leadership
>>
>>108677306
>it's april 2026 and microsoft and meta are about fire 20k people cuz of ai, lilbru
Followed immediately by 20k H1B applications. AI is Actually Indians, you let them get into your corporations' management and now they're replacing you.
>>
>>
>>
>>
>>108677663
Living in conditions so bad due to overcrowding that the population doesn't grow is horrific and makes armies of horrible shitty people with nothing to live for.
Populations should never ever be allowed to reach that point.
Kill yourself, retard.
>>
>>
>>
>>
>>
>>108677296
If an employee had a massive failure rate and kept making seemingly random mistakes while also forgetting everything you told them then you'd just fire them.
AI basically infects your employees with this disease, only the truly competent know how to properly use the tool in the first place. Midwits basically become a massive net loss in productivity at best and unhinged liabilities at worst because they'll just obey the slop when it tells them to do garbage.
>>
File: forever-seething.jpg (285.2 KB)
285.2 KB JPG
>>108677296
>>108677348
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>108677296
Companies are already short reviewers, and in practice the only things that get proper review are mission critical changes. This has been fine until now because the least competent people could not physically produce code all that fast that compiled or even pretended to work. So there was a kind of equilibrium. In many places mid level engineers can slack by fixing showstopper bugs produced by gangs of junior jeets and changs. The seniors are too busy spending 75% of their time politically infighting, 15% steering the ship away from the rocks, and 10% (if that) reviewing things.
AI has fucked this up beyond repair. The most retarded are now turbo-enabled to produce an unfathomable amount of code churn and also to undercut the lazy mid level guys by encroaching on their turf, trying to oneshot bugs they themselves caused. This is causing massive anxiety at that mid level.
For their part, the seniors are either becoming delusional in 100 different ways or cynical and "fuck this gay earth". There is a near universal feeling that things must crash and burn before a new order arises from the ashes.
Did I miss anything? Oh yeah, all layers of management have opinions on everything, but they're all becoming irrelevant so fast that they're having trouble mentally processing this fact. It's complete mayhem.