The LLM Euphorics—the types who might report about being poor for a few months because they “splurged” on a multi-tens of thousands of dollars LLM server—are now concerned about LLMs for the people. Yeah okay.
Those of us who are negative about AI for political reasons have been saying from the start that the biggest problem with AI is power. People can’t now all of a sudden be thinking that huh nation states have power (along with Big Tech and the rest of the power brokers).
But this is in fact quite a tortured fear, all wrapped up in the usual hyping—though this part is expected of LLM Euphorics. The usual story of simply making human labor less valuable and concentrating hardware for compute is just, you know, this rotten economic system working as it is intended. No need for weapons, subterfuge, three-letter agencies, much more straightforward, and just a natural evolution of X-CLASS CAPABILITIES.
What kind of world are we building where advanced technology is just used to stop living as full human beings? Where’s the desire to self-actualize? Did it start dying when people got glued to the idiot box?
It’s good that clankers are not afraid of throwing away code. The biggest problem with code generation (that is version controlled) is maintenance. It’s better to throw away questionable code rather than say eh, we don’t quite understand this part (and our agents can’t make a compelling story about it) but we spent a lot of effort on it and it apparently works so we better keep it.
.. only if you know what the code is doing, though. Often the requirements get scattered and lost to the winds and the code is the only record of its own idiosyncratic behavior. And yes, someone's depending on the bugs in it.
This could be true but, no matter what, a streetwise person would never trust him after he has taken the money. If he wanted his opinion on AI to be trusted then he should have made his money some other way.
The "good" argument is that people trust other people's opinion more who have not been paid to advertise. I trust the doctor who personally recommends a drug more than the doctor who was paid to recommend the drug - even if they recommended the drug before they were paid. That's a fact of life, it's not "character assassination". Tao didn't do anything wrong by making an ad but he can't expect people to take his opinion seriously after someone gave him a lot of money to state an opinion that favors them.
It's proper to suspect arguments that are motivated by self-interest. The stronger that self-interest, the more one should suspect the argument. This is what you're saying?
In that case, the anti-AI Luddite arguments are maximally impeached, since they are motivated by fear of personal disaster. Tao doesn't need AI to succeed; the Luddites desperately need it to fail. So they are willing to say anything, jumping right to ad hominem arguments when they lack any real substantive rebuttal.
I don’t know that Terry much cares about the opinions of people who judge claims based on innuendo and cynicism rather than the actual merits of the claim.
> I think Terry Tao is a great litmus test for AI zealotry (both pro- and anti-). Just in this thread, we have people twisting themselves into knots about how he "sold out" or "not doing math the right way" or whatever. To him, AI is a tool, like any other.
We've been flooded with "AGI is 6 months away!" for a few years now, mostly by Anthropic/OAI/XAI, which is clearly nonsensical hype. Also, almost everyone has been walking back their previous claims that "AI will replace ~80% of white-collar jobs."
> Also, almost everyone has been walking back their previous claims that "AI will replace ~80% of white-collar jobs."
They started walking those claims back right around the time someone tried to set Sam Altman's house on fire.
Not making those claims anymore doesn't necessarily mean they don't still believe those claims, it is very possible they just realized saying the quiet part out loud was a bad look for them even if it was/is what they believed to be true.
In the last decade or so I have never seen so much layoffs across the industry. This may be suggesting that evidence supporting the latter hypothesis is not maybe too far fetched.
In the last decade, the software engineering industry has turned into a grift that has pushed out hundreds of thousand of low-quality "engineers" through coding bootcamps or online courses. Many of these people have no passion for the craft or interest in building products.
Then, when money was cheap during COVID, companies over-hired unscrupulously. Now, given that markets are cooling off and there's some political, geopolitical, and economic uncertainty, companies are hedging their bets, and laying off is usually the right move, especially as interest rates are going back up.
There are perfectly viable explanations for the situation the industry finds itself in without invoking the AI boogeyman, especially considering that just about every study out there shows that AI use correlates with a fairly modest increase in productivity, and that it won't turn anyone into a "10x engineer" overnight.
Over-hiring could be one way of explaining the effect we are seeing, however, where are those "coding bootcamp" or "online courses" engineers? I honestly ask because I have never worked with one in almost 2 decades of being in the industry, and I worked across many different domains. What I see is on the contrary - the people who are getting laid off are people with legit engineering degrees from legit engineering Universities.
Also, over-hiring by the very definition implies a sudden surplus of engineers on the market. I can't quite understand where did these engineers all of the sudden come from? The number of engineers outputted by the Universities YoY is pretty much close to O(1) so I am not convinced to this theory at all and I see it only as a good excuse that companies are making in order to make them look better.
I spoke with my friends few days ago, and one of them runs the company so he asked me on the opinion of the AI frenzy. I gave him my view and by the end of it he told me that he feels uneasy but that he has to let go part of his employees because he simply does not need them anymore - they are literally replaced by the AI model and one or two or N-M engineers operating the model. Yesterday he needed 10 people for the job, today it is 2 or 3 people.
So, I think that the AI has already changed the landscape dramatically, and what we are seeing are not the post-COVID effects.
Where I'm from and peripheral countries, the industry is riddled with bootcampers and button pushers. My company even has a big bootcamp for reconversions
What are you trying to suggest? That people without the University degree who have been trained for monkey coding do exist? Sure but that's not what I was saying nor does it skew the picture in any significant way.
What I'm trying to convey and you fail to understand is the picture you have in your mind is very much affected by your reality. The fact that you don't see these bootcampers doesn't mean they don't exist.
If "the latter hypothesis" of parent commenter was that "AI will replace ~80% of white-collar jobs", then that hypothesis clearly not supported by the current layoffs. AI isn't replacing workers, AI just happens to be an easy excuse for it. Could as well have been "COVID" or "tariffs" or "the economy" or "the end of Zero interest-rate policy"
Why not? I have literally got several first hand examples where people are fired because of how good the AI models became. Why do you find that questionable?
Objective reality is that many people have lost and are still loosing their jobs. If you don't have anything useful to add to your response please refrain from polluting the discussion.
No, it is not the same at all. I intentionally frame my words by saying that after all there may be an indication that such an event or correlation exists but I am explicitly not stating anything, therefore it is rather an invitation for discussion and not one-sided talk.
Are you a trained lawyer? Okay but presumably not practicing in the last twenty years.
You know that all contributions to the Git project has to be signed off as either being made by yourself or being handed over by someone who has signed off on that certficate of origin. For everyone on every change. Even the lead developers so to speak. And you spend some thousands of dollars and run an AI analyis tool to wash your hands?
Who are you to do that? Oh wait I forgot, you are Mr. Chacon. A hand in everything Git and friendly with everyone in Git who matters for twenty years. Remind us next time as well so I don’t forget.
Judging by all the trivial errors made in pre-AI code, many of us are far from rigorous a lot of the time.
It’s like that story about the programmer who wants to send the car down the slope one more time to see if it does the same thing again (or whatever it was). The ephemerality makes iteration possible and appropriate, but also makes rigour less important.
Just changing words and saying (perhaps) that one is “just about code” and the other is about “architecture” rings hollow. LLMs can think (like submarines can swim) so the only question is how well.
I keep being negative about Digital Tech in general[1]. But this is worse than my habitual negativity towards D. Tech sans AI (AI is a whole chapter onto itself).
And what can be done? The comments usually say a big fat nothing.
- Any fool already knew this comments: “shouldn’t be a surprise”
- I guess I should call my representative comments
- Just boycot tech comments
Usually nothing much actionable. Building the Ad/Surveillance/Privacy Invasion society? Very actionable, good pay, many mouths fed and FIRE accomplished by HN posters. There’s even at least one acronym for this life achievement.
Shoutout to digital activists that are doing something. I’m but an armchair complainer on this front.
Those of us who are negative about AI for political reasons have been saying from the start that the biggest problem with AI is power. People can’t now all of a sudden be thinking that huh nation states have power (along with Big Tech and the rest of the power brokers).
But this is in fact quite a tortured fear, all wrapped up in the usual hyping—though this part is expected of LLM Euphorics. The usual story of simply making human labor less valuable and concentrating hardware for compute is just, you know, this rotten economic system working as it is intended. No need for weapons, subterfuge, three-letter agencies, much more straightforward, and just a natural evolution of X-CLASS CAPABILITIES.
reply