qarl

7 ms·

91.
by qarl 1mo ago
How else are they going to learn?!?
92.
by qarl 1mo ago
Hm. I suspect you're being sarcastic. Let me respond to that - LLMs will definitely produce racists content if that's what they've been trained on. They just repeat what they've been fed. For example, there was a tim
93.
by qarl 1mo ago
How do you kill a botnet?
94.
by qarl 1mo ago
I love this new era of AI powered creativity. I expect we're just getting started.
95.
by qarl 1mo ago
> I'm not convinced information theory... I'm not convinced my molecules follow the laws of thermodynamics.
96.
by qarl 1mo ago
Yes - exactly. Because they don't hallucinate in the same way - they are perfect for detecting each other's hallucinations. But when you ask the SAME LLM (with the same context) they remember the hallucination so it doesn't
97.
by qarl 1mo ago
Ah. I missed that. Yeah. "Backprop alternative" is the wrong framing.
98.
by qarl 1mo ago
Personal experience using agents and seeing this happen frequently. Try it yourself. Get one to hallucinate, then paste that text into a new window and ask it to verify the facts. EDIT: Also - Cohen, Hamri, Geva & Globerson, "LM v
99.
by qarl 1mo ago
> But when an “algorithm” does it, the process is empiricism-washed. I don't understand his example. He says "math can't be racist" but the story he cited is that the person who made that statement was corrected and
100.
by qarl 1mo ago
They state replacing backprop is not their goal. Their goal is to understand how distributed systems which cannot do backprop (the brain) can still do learning.
101.
by qarl 1mo ago
Well - yes - there is no agreed upon definition. Which makes the issue even harder to clearly determine.
102.
by qarl 1mo ago
Oh. Yes. Sure. If there is an error in the training data then they will obviously cite that error. I'm mostly talking about random coding errors.
103.
by qarl 1mo ago
Yes - those failures (like strawberry) are. And those failures are very rare, which is why you had to reach for the Pit Imp MTG card, which I had to Google to understand your point.
104.
by qarl 1mo ago
I'm sorry you're so upset, but it's true. Try it for yourself - get your LLM to hallucinate something, and then paste that text into another chat window and ask it to verify the result for you.
105.
by qarl 1mo ago
Well, yes, I understand that testing does not create a provably correct solution. But I'd love to hear the source of your "extremely inadequate" or your "10%" claims. I mean - there is a reason why it's used
106.
by qarl 1mo ago
Well - it's not hard to find. But ok: Birch, The Edge of Sentience (2024), ch. 16 - "simply no way to assess sentience in an LLM" Schwitzgebel, AI and Consciousness, (2025) — "we won't know before we've alread
107.
by qarl 1mo ago
Yes. If we allow a file with unbounded storage (like your exception for physical computers) then an agent can easily perform the necessary state transitions and tape reads/writes.
108.
by qarl 1mo ago
> It even disagrees with itself. Whenever Claude writes a message that takes a position on something, its final one or two paragraphs will try to dismantle its own argument. I have agents adversarially check each other. When I switched
109.
by qarl 1mo ago
It's the anti-sycophancy stuff, isn't it? But yes - I agree the behavior is new and VERY annoying.
110.
by qarl 1mo ago
While this is absolutely true - I'd hesitate to discount using similar agents for checking each other. Two agents will almost never hallucinate in the same way, regardless of their weights - and by having a second one (with a differen
111.
by qarl 1mo ago
I think that AI is dangerous and could be used as a weapon. So yes, I would like to be protected from all parties. I don't think that's nuts.
112.
by qarl 1mo ago
You understand that robust testing can, should, and often does test directly for reliability and robustness, right?
113.
by qarl 1mo ago
Those are two different issues. One is about typical speech patterns and one is about liability. I agree with you on the liability issue, but I don't think there much question about this issue outside the anti-AI conspiracy campaigns
114.
by qarl 1mo ago
I think those companies are referring to other companies - say Chinese AI companies - who we also need protection from.
115.
by qarl 1mo ago
I hate to spoil your mood - but it is currently unclear whether agents can be self-aware. And it's very likely something that can never be known.
116.
by qarl 1mo ago
That distinction doesn't matter to my point.
117.
by qarl 1mo ago
This distinction is silly. We say "Google's web crawlers scape web pages." We don't insist you say "Google uses web crawlers to scrape web pages." We describe software as having agency all the time. It's
118.
by qarl 1mo ago
In my experience - a robust testing harness will get you the safety you need. And most software you describe has such testing. I think the Bun port is a great example where testing enabled a very successful implementation. (Both the origi
119.
by qarl 1mo ago
Fair enough - it will take some time before the science comes back. But I feel the need to point out - the goalposts for "does AI work" shift daily.
120.
by qarl 1mo ago
Really? OK - then let me stand up and vouch that I am at least 20X more productive with AI.

More ›