qarl2

5 ms·

31.
by qarl2 1mo ago
> You still don't get it. HEH. I just read your comment three times and I definitely don't get it. I think its entire point is to express outrage about not getting it, rather than actually explain it.
32.
by qarl2 1mo ago
Really though? You have to try really hard to make your code that unstable - and well written code isn't that. Seems more like voodoo.
33.
by qarl2 1mo ago
You yourself would be non-deterministic if we could rewind time. So what's the issue here? Not to mention, LLMs can be made deterministic if that were actually valuable. Seems like determinism is just the latest excuse to hate.
34.
by qarl2 1mo ago
The author's product (Realness) heavily promotes the fact that it works without the use of AI... ... but clearly AI was used in the development? That seems deeply cynical.
35.
by qarl2 1mo ago
I wonder why.
36.
by qarl2 1mo ago
Sure. And that's what I mean by a practical solution. But it seems suspect to accept a practical solution in one area (humans) and deny it in an almost identical (machines with neuron based brains) area. No?
37.
by qarl2 2mo ago
YES - I agree entirely. But how are we going to escape solipsism? These questions need a practical solution, not a perfect one. That's the problem. We need to stop chasing the perfect one.
38.
by qarl2 2mo ago
This is the question of qualia - and it's surprisingly intractable. Realize that in this context, I can't be sure that you are conscious, either. Arguments here will crash into solipsism in the end: you guys can never prove you&
39.
by qarl2 2mo ago
> There’s no third party test for a fundamentally first person phenomenon. Agreed. But that also means I have no empirical evidence that YOU are conscious either.
40.
by qarl2 2mo ago
Well - yes - I have. And the test came back positive. But it's always possible I'm just being naive in some way. An external challenge makes it more robust. I don't think I'd want to do an entire linux distro - it&#x27
41.
by qarl2 2mo ago
Come come, let's spec something real. You seem extremely doubtful, why don't you let me show you?
42.
by qarl2 2mo ago
Oh, I believe agents are fully capable of it. Would you like to issue a challenge like that? I'd love to demonstrate their capabilities to you.
43.
by qarl2 2mo ago
"That said, the recent shift feels more profound: almost every day, it feels that the lineup is dominated by stories focused on AI, written by AI, or commented on by AI." I interpreted that as surprise. I apologize if you disagre
44.
by qarl2 2mo ago
The author of the article seems to be surprised. As well as many people in these comments.
45.
by qarl2 2mo ago
> "They lie about it being dangerous!" and simulataneously "They hacked a system with zero day exploits! Everyone involved should goto jail!" FWIW: The simultaneous "it's stupidly weak" and "it&#x2
46.
by qarl2 2mo ago
Surely you're aware of the various theories that AI is going to up-end the entire world. How can you be surprised it's up-ending a geek forum first?
47.
by qarl2 2mo ago
OH - I think I understand now. You're like LeCun - dismissive of LLMs in all their forms. So of course agents can't lead to AGI, because LLMs just can't. Which I've always found so strange. Language was always conside
48.
by qarl2 2mo ago
> In my experience the use of agents and harnesses and other scaffolding around them doesn’t improve the performance of LLMs much I have to say - that seems straight-up crazy to me. Would you mind digging into that discussion? As just o
49.
by qarl2 2mo ago
> I love how everything is recorded now. I mean, except when it's censored. That part is a shame. HEY! Is there a tool that tracks the censored comments? I bet there is. And if there's not... that will be a delicious new p
50.
by qarl2 2mo ago
Given how fast your comment is being swarmed, I suspect so. Also seems like exactly the kind of thing Elmo would fund. EDIT: And now it's been censored. What a strange coincidence!
51.
by qarl2 2mo ago
I remember when I caught PG on reddit arguing with some guy who'd said something mean about him. He didn't reveal who he was. But you could tell from his history - his first post was from before reddit opened to the public. Good
52.
by qarl2 2mo ago
You were making some pretty strong statements about the usefulness of harnesses that seemed to me to be entirely detached from reality. Maybe I misunderstood you? Because I can provide evidence to correct that misconception. Another of my
53.
by qarl2 2mo ago
Sigh. "I'm right, you're wrong" doesn't make a very good proof. EDIT: I'm going to bed. You may have the last word.
54.
by qarl2 2mo ago
Satire is merely an analogy. I think maybe you're having trouble with metaphor and ambiguity.
55.
by qarl2 2mo ago
I'm sorry I was not able to explain to you how things can be two things at the same time.
56.
by qarl2 2mo ago
HEH. Never used agents, huh? The failures are mostly uncorrelated. Why are you ignoring basic principles of engineering (knowing what you're talking about before opening your mouth)?
57.
by qarl2 2mo ago
Would you like to see some of my own examples of wild success? I have a system that entirely reverse engineers old arcade games. Creates semantic symbol mappings that were considered impossible just a couple years ago. Granted, it took me
58.
by qarl2 2mo ago
You know - that's why we use 2 agents for any important decision. 1 in 10 becomes 1 in 100. And you can get any degree of accuracy you desire. If only humans were so reliable. Why are you ignoring basic principles of engineering?
59.
by qarl2 2mo ago
> presuppose competence in LLMs which simply doesn’t exist. Then how do you explain the wild success at using them for development?
60.
by qarl2 2mo ago
Target.

More ›