Hacker News: qsera

New comment by qsera in "OpenAI backs Illinois bill that would limit when AI labs can be held liable"

qsera — Fri, 10 Apr 2026 16:21:29 +0000

It won't be, because it is marketing, and we are eating it up.

New comment by qsera in "OpenAI backs Illinois bill that would limit when AI labs can be held liable"

qsera — Fri, 10 Apr 2026 16:20:51 +0000

Another marketing gimmik...

New comment by qsera in "OpenAI backs Illinois bill that would limit when AI labs can be held liable"

qsera — Fri, 10 Apr 2026 14:24:07 +0000

https://en.wikipedia.org/wiki/National_Childhood_Vaccine_Inj...

New comment by qsera in "White House staff told not to place bets on prediction markets"

qsera — Fri, 10 Apr 2026 14:17:40 +0000

Mmm..yea. But stock market is not something new, and this possibility always existed.

So I was only thinking about Poloymarket and stuff like that.

New comment by qsera in "OpenAI backs Illinois bill that would limit when AI labs can be held liable"

qsera — Fri, 10 Apr 2026 14:09:57 +0000

>But that would be an insane thing to claim.

The trick is to make people behave like that without actually claiming it. AI companies seems to have aced it.

New comment by qsera in "OpenAI Backs Bill That Would Limit Liability for AI-Enabled Mass Deaths"

qsera — Fri, 10 Apr 2026 14:08:01 +0000

> it simply shouldn’t be possible with a proper approach to testing.

It just has to be delayed. Like many years after application. Or trigger on very specific and rare circumstances. Not likely in a trial, but near certain at a population scale.

Or both...

On top of that, If I remember correctly, this liability wavering also exist for Vaccines.

New comment by qsera in "White House staff told not to place bets on prediction markets"

qsera — Fri, 10 Apr 2026 13:56:57 +0000

It also incentivize people to follow the events and remember stuff so that they are better equipped to make predictions.

I think it incentivize people to use their head and may be even select better leaders.

New comment by qsera in "White House staff told not to place bets on prediction markets"

qsera — Fri, 10 Apr 2026 13:54:50 +0000

I think the proper thing to do would be to forbid these markets to make bets on stuff that can be won by using only insider information.

New comment by qsera in "ML promises to be profoundly weird"

qsera — Fri, 10 Apr 2026 01:12:12 +0000

> LLMs are circuit builders

I think they are circuit "approximators". In other words, a result of a glorified linear regression..

New comment by qsera in "Claude mixes up who said what"

qsera — Fri, 10 Apr 2026 01:10:15 +0000

Obviously, a real intelligent entity would consider risk/benefit analysis and act accordingly.

New comment by qsera in "ML promises to be profoundly weird"

qsera — Fri, 10 Apr 2026 01:08:38 +0000

I think It matters for the question that I was responding to.

New comment by qsera in "ML promises to be profoundly weird"

qsera — Thu, 09 Apr 2026 14:59:06 +0000

https://arxiv.org/abs/2603.09678

New comment by qsera in "Claude mixes up who said what"

qsera — Thu, 09 Apr 2026 14:10:03 +0000

>If you were there, what would you do?

Show it to my boss and let them decide.

New comment by qsera in "ML promises to be profoundly weird"

qsera — Wed, 08 Apr 2026 18:17:34 +0000

But that is bolted on and is not a core behavior.

New comment by qsera in "ML promises to be profoundly weird"

qsera — Wed, 08 Apr 2026 17:34:17 +0000

Source?

New comment by qsera in "ML promises to be profoundly weird"

qsera — Wed, 08 Apr 2026 17:28:28 +0000

For starters, natural brains have the innate ability to differentiate between things that it knows and things that it have no possibility of knowing...

New comment by qsera in "ML promises to be profoundly weird"

qsera — Wed, 08 Apr 2026 17:15:26 +0000

I am not the OP, an I have only used ChatGPT free version. Last day I asked it something. It answered. Then I asked it to provide sources. Then it provided sources, and also changed its original answer. When I checked the new answers it was wrong, and when I checked sources, it didn't actually contain the information that I asked for, and thus it hallucinated the answers as well as the sources...

New comment by qsera in "ML promises to be profoundly weird"

qsera — Wed, 08 Apr 2026 17:04:24 +0000

But why do you need an example? Isn't it pretty well understood that LLMS will have trouble responding to stuff that is under represented in the training data?

You will just won't have any clue what that could be.

New comment by qsera in "ML promises to be profoundly weird"

qsera — Wed, 08 Apr 2026 17:00:39 +0000

>95% is not my experience and frankly dishonest.

Quite frankly, this is exactly like how two people can use the same compression program on two different files and get vastly different compression ratios (because one has a lot of redundancy and the other one has not).

New comment by qsera in "AI Assistance Reduces Persistence and Hurts Independent Performance"

qsera — Wed, 08 Apr 2026 15:00:29 +0000

The AI marketing just need the publicity people generate during the honeymoon phase. Falling out is generally gradual and typically in silence.