I have been on HN since something like 2008, and I've seen many trends come and go here (including JS framework fads :)
I remember when blockchain and Web3 were massively popular (2014) with only a few dissenting voices and then became total anathema to the point where VCs like Andreessen Horowitz are spit-roasted for riding Web3 hype and selling to the public. (OpenSea was incubated in YCombinator btw.)
So I am genuinely curious because the same seems to be happening with AI, but the result seems absurd to me. With Web3, we are in the "this sucks and has no good applications" phase, but with AI we are in the "but it could have so many... amazing applications" phase.
Yes, I agree that AI might have some good potential applications, and I'll even (for argument's sake) grant that Web3 and smart contracts have no good applications at all. But the question here is about the DOWNSIDE to society. After all, that's what we should be talking about.
Web3 was mildly zero-sum for society, while AI is shaping up to be massively negative-sum, as more people get access to it. It's like giving bombs to everybody and hoping no one will use them in bad ways. At least with nuclear bombs we knew who was enriching the plutonium. Here, everyone has access to weapons, to be used as they please.
The key to when things go bad is swarms. When HN is overrun with swarms of bots a few years from now, you'll remember reading this. Web3 by contrast is just an immutable ledger, its worst applications are convincing people to voluntarily put some money into a buggy smart contract.
With AI, it's not voluntary participation, you can't opt out of being assaulted by swarms of bots peddling fake information, impersonating people, and actually being socially and economically forced to interact with it all, the same way Facebook and LinkedIn were highly needed for social and professional life online. And this is to say nothing of disinformation at scale, like we just saw with the officials in India.
Perhaps this video will bring it home, but for online usage: https://www.youtube.com/watch?v=HipTO_7mUOw
What I'm curious about is, isn't HN supposed to exercise reason and logic, and not just ride the waves of hype and disillusionment like the less-tech-savvy public? Reason says AI can have much worse consequences than Web3 ever did. And the experts do too... this guy who used to work at OpenAI Safety Researcher says 20% chance it leads to most humans dead:
https://news.yahoo.com/ex-openai-safety-researcher-says-165725772.html
As you say, AI is overhyped. And most of the potentially "bad" applications are also things you could do with a shell script. If nothing else, it has some kind of symmetry where if it turns out it has really evil uses, it will also have really good uses, which will be more attractive. Same as computers in general.
Web3 was almost 100% scams and gambling with very little productive utility. The good uses were almost nonexistent. It was one of the more vapid hype waves in tech history.
No. This is a finance site, reason and logic has absolutely no bearing on these people once you tease infinite returns like this. Sam Altman just waved AI demos in front of Microsoft's face and got a business deal closed in less than a week. It wouldn't surprise me if YC's diligent investors were the same way with Web3.
Personally speaking (having previously been part of the Web3/crypto movement circa 2015), I think AI is also zero-sum. The race to the bottom looks different from cryptocurrency though, and less harmful than it is annoying. Cryptocurrency's failure was obvious - a currency without guardrails or guarantees is a king's ransom of sand. It falls through your fingers over months of rug pulls, blockchain drama and nervous sideways-trading. The only way to get good at cryptocurrency was to stop using it, in my experience. In the age of KYC it's barely even useful as a fast alternative to wire transfers.
On the other hand, AI is indeed harmful but not in a terribly destructive way. It promotes false knowledge, can distributes harmful information and even make humans feel hopeless and depressed. The internet also promoted these sentiments though, and the "business" side of cryptocurrency was almost entirely centered on swindling and misinformation. The "lowest common denominator" scenario for each side is dire, but the messaging around cryptocurrency legitimizes it when at almost every level it is fundamentally unsafe. That makes it the more dangerous technology, in my eyes.