It's like thinking beyond today is impossible and that using logic to predict behavior is akin to being a conspiracy theorist.
Some now popular conclusions around AI were flagged comments ~3 months ago.
No wonder software is so terrible today.
Look at how the "Wuhan lab leak theory" was flagged and buried, how it took Jon Stewart to voice it and take a lot of heat before anyone paid any attention, and now it's mainstream and the Federal government is supposedly officially investigating it. https://nypost.com/2023/02/28/jon-stewart-recalls-outrage-af...
Sometimes governments use it to gaslight everyone: https://www.outlookindia.com/website/story/how-jimmy-carter-...
On HN, the orthodoxy was that you shall not talk about everyday causes of depression or ADD. But then in 2022 we had this meta analysis that found no evidence supporting the serotonin theory and SSRIs: https://www.sciencedaily.com/releases/2022/07/220720080145.h...
Now it's OK to talk about it, whereas before it was not. So now we have a lot of taboos to talk about overdiagnosis of ADHD and gender dysphoria, talking about the DSM V criteria etc: https://www.theguardian.com/society/2011/feb/04/adhd-diet-fo...
So yes.. right now the thing is to hate on Web3 and to do strawmen like "how will it fix AI". It won't fix AI, it's just that Web3 may be roughly a zero-sum game for society, or at least not nearly as negative-sum as AI.
> Off-Topic: Most stories about politics, or crime, or sports, or celebrities, unless they're evidence of some interesting new phenomenon. Videos of pratfalls or disasters, or cute animal pictures. If they'd cover it on TV news, it's probably off-topic.
As much as you're entirely right that HN's readership decides what succeeds or not, pretty much every example you've furnished is defined clearly as "off topic" submissions. The Wuhan lab leak theory, regardless of how we view it in hindsight, was paraded as a sensational event up and down party lines. It never had a chance on this site, like Donald Trump drama or "Ask HN: What did you think of Kardashians last night?". Maybe it is relevant - but framing it sensationally is like toxic waste to this site's readers.
> the orthodoxy was that you shall not talk about everyday causes of depression or ADD
> So now we have a lot of taboos to talk about overdiagnosis of ADHD and gender dysphoria, talking about the DSM V criteria
They're not taboo. They're just not appropriate conversation topics for a hacking/finance website. The few times that it has become a big discussion topic has been through the lens of engineering lifestyles and personal experience. Even that is arguably too sappy for a non-Reddit website.
> Now it's OK to talk about it
> So yes.. right now the thing is to hate on Web3
Very subtle foreshadowing - you almost make it sound like Web3 has a second chance or something.
Truth is, like I said at the beginning - HN chooses what they want to hear. Most websites that give their users any degree of control over their experience let this happen (see Twitter, where users block the verified users). If you demand engagement on a topic nobody cares about, it stands to reason that you'll feel repeatedly disappointed by this website.
I think it absolutely is relevant, given that the conversation itself was about changing narratives.
> you almost make it sound like Web3 has a second chance or something.
This thread itself, is one of the rare ones where bringing up web3 actually is relevant, yes. It is reasonable for him to make that argument, because the topic is literally about how to authenticate things when you have an opposing central authority.
> If you demand engagement on a topic nobody cares about
His point wasn't about engagement. Instead his point was about how unpopular opinions are being proved correct, and that people were way to quick to dismiss ideas.
His point was that people should be able to look ahead, into the future, instead of immediately dismissing things out of hand, just because the thing isn't directly effecting the world right now.
IE, if you were downvoting/disagreeing with opinions about how AI is going to make it hard to authenticate information online, then you were wrong and were way to quick to dismiss an opinion out of hand.
I mean, it's a stretch. I could use that stretching metaphor as an excuse to talk about taffy for 4 paragraphs, but it doesn't contribute anything meaningful to the conversation. Scrounging up discussion about Web3 on an Indian government exposé is off-topic, and readers will flag it. It's equally substantial as those "this is good for Bitcoin" posts that get flagged under every article about a government treasury.
> the topic is literally about how to authenticate things when you have an opposing central authority.
If so, his flagged post didn't reflect it. It was a vapid comment that made no attempt to reconcile either topic.
> his point was about how unpopular opinions are being proved correct
If you want to argue that the Blockchain could have prevented this from happening, then by all means - write that comment. But that's not what his flagged post was. It detracted from the conversation and quite literally read as flamebait.
> if you were downvoting/disagreeing with opinions about how AI is going to make it hard to authenticate information online, then you were wrong
Spoilers: it has been hard to authenticate information online for as long as the internet has been around. In fact, the most significant events are corroborated with real life occurrences that challenge these lies. These women can attest that they did not do this in the photo. If government officials insist that the version with them smiling is the true master copy, then you're the fool for believing them.
I'm baffled that this is even being used as a platform to discuss Web3, a collection of technologies mostly reputed for... being manipulated and exploited by bad actors that cannot be incentivized to stop.
It really isn't. Because it is one of the normally very rare use cases for web 3 stuff, and the topic is directly about a case where such a thing would be needed.
> it has been hard to authenticate information online
That's nice and all. But the claim was that people were previously dismissing opinions about AI that are now mainstream.
So yes, you'd be wrong if you were outright dismissing these opinions about AI before, which are now popular.
And it is useful and interesting to point out how previous opinions that were dismissed out of hand, are now mainstream, using this situation as an example.
> a case where such a thing would be needed.
So... suggest a solution. What is your magical Blockchain panacea for this situation? To plug all digital cameras into an online ledger and hash every image? I'd love to hear a non-pipe-dream based idea for this.
> But the claim was that people were previously dismissing opinions about AI that are now mainstream.
This could be said about just about everything. I still dismiss the flying car, even though I was "proved wrong" in the 60s and 70s when some coked up engineers built one. I still dismiss cryptocurrency, even though many coins will outlive me. There is no world in my lifetime where cryptocurrency obsoletes the IMF. Betting against the powers that be is like playing chicken against a shotgun.
I am not sure if you read the rest of the comments and suggestions in this whole post, but multiple people did, over multiple discussion threads.
Feel free to go back and read them.
Also, it is still interesting and useful to suggest part of the solution, without writing dozens of pages on the exact specific details of it. Even just recommending crypto, is interesting enough, due to the fact that real usecases for crypto are so rare.
> This could be said about just about everything.
Ok, well in this case, related to AI, lots of people were wrong. And those people who were wrong, should figure out what mistake they made, to be so wrong, and why they were so quick to incorrectly dismiss these arguments.
Bombs have been around for a long time. What if everyone could have many, cheaply and easily? Why, since nothing happens with N of them, then 20,000 x N is just a little more of the same. And so what if the bots are now able to hold conversations and pass the Turing test 24x7 at scale? We have been able to hire humans to do that.
Except… no. That’s not how it works and you know it. Allowing anyone to deploy swarms of bots with superhuman memory, strategy, coordination, and now the ability to do many tasks with acceptable or good results at scale, is a game-changer and the question is whether it will make society better off.
The same questions asked of Web3’s downsides should be asked of AI. Not upsides, downsides. The fact that the guy in charge of safety at OpenAI says he thinks there’s a 20% chance of human extinction and you choose to shrug it off and focus on some other part — is the issue.
This isn’t just an academic discussion. It is like the people who profit off of coal and other things, regardless of their emissions or effect on people or climate. It matters and should be regulated.
Why was GP flagged? They made a good point.
The point was that 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.