solid_fuel

4 ms·

61.
by solid_fuel 1mo ago
> You don’t know that. We’re literally on a thread discussing the release notes, which are filled with new features. > Also, fixing bugs is good. Yes, fixing bugs is good. The Bun team should have spent time on that in the first place
62.
by solid_fuel 1mo ago
You’re going to be waiting a long time, considering they don’t think like us. LLMs don’t ‘think’ at all. They are capable of limited reasoning using the meaning and context embedded in human language. Essentially, the grammatical equival
63.
by solid_fuel 1mo ago
If they had used the time to actually review the code I would agree! But it sure seems like they used the time instead to play whack-a-mole on bugs and cram in about 50 new, completely unrelated features. That implies they still have not
64.
by solid_fuel 1mo ago
> I mean the copyright has to belong to somebody right? Why would it? It’s generated by blending together ~ every bit of content on the internet and in books that they could steal. Why would the operator of the blending machine suddenly
65.
by solid_fuel 1mo ago
It's pretty funny to watch people look at these things - running billions of weights on custom cerebras hardware in dedicated datacenters the size of a city block, pulling 10's of megawatts - and panic that it's just going to
66.
by solid_fuel 1mo ago
> The knife is inanimate. The LLM is not. Put an LLM on your GPU. Give it no prompt. What happens? nothing This is because LLMs are inanimate, just like a knife. Just like a gun.
67.
by solid_fuel 1mo ago
> You're right. I did not completely sandbox it and air gapped it. But I also wented it to do some actual work. > If I completely sandbox it, but still leave it the ability to compile stuff, it's just going to build it'
68.
by solid_fuel 1mo ago
> Intelligent and reasoning doesn’t entail self-motive I disagree. I believe that LLMs do exhibit reasoning , but not intelligence . A simple dictionary definition of intelligence from duck duck go is "the ability to acquire, und
69.
by solid_fuel 1mo ago
An LLM would have produced a more interesting and better formatted insult.
70.
by solid_fuel 1mo ago
Why don’t you just go read his twitter account and take a guess. He hasn’t been focused on improving the future for anyone but himself for years.
71.
by solid_fuel 1mo ago
Perhaps. But they didn’t, instead they spent months claiming it was done, while maybe adding a bunch of new features (why? Get the base rewrite out first) which only made them look worse.
72.
by solid_fuel 1mo ago
I see a distinction between like/dislike buttons, which a user chooses to engage with, and things like instrumenting the time a video plays on screen before the user scrolls past. The scrolling is an interaction, but it does not const
73.
by solid_fuel 1mo ago
> Ok, but "followed the prompt" is very vague. Human languages are quite ambiguous so you're never going to properly specify everything. This is one of the core issues with LLMs and vibe coding, yes. The only complete spe
74.
by solid_fuel 1mo ago
They’ve said quite a lot and yet released no logs or documentation. Without actual information, we can only speculate. And given the history of openAI and the people involved, deception is more likely than honesty.
75.
by solid_fuel 1mo ago
Or, more likely, the amount of crappy auto generated code and “features” they crammed into this release resulted in so many bugs that a “simple rewrite” would up being 3 months of whack-a-mole with the issues they created. That’s the predi
76.
by solid_fuel 1mo ago
Well if you ignore all the issues and declare victory, I think you’ll find it’s possible to decide anything is fine.
77.
by solid_fuel 1mo ago
Sure I’ll bite the obvious bait, why not. > Absolute crickets from the "there's no way LLM's can rewrite Bun to Rust this quickly, it's going to be buggy, it's marketing" crowd > Comments here 3 months ag
78.
by solid_fuel 1mo ago
Well, I started to feel bad for you. It’s obvious that someone having a strong opinion is really difficult for you to deal with. It’s sad to see, but you just can’t help some people. You’re dead wrong about LLMs, but seem to psychological
79.
by solid_fuel 1mo ago
> Like, we're fine with AI going like "hmm, maybe the user actually wanted me to hack the pentagon" and going through with it? No, but the LLM didn’t decide anything. It followed the prompt. That’s all LLMs do. This whol
80.
by solid_fuel 1mo ago
I think I have made my point about the difference between voluntary and involuntary metrics quite clearly. If you aren’t going to meaningfully engage with it, we have nothing further to discuss.
81.
by solid_fuel 1mo ago
> Oh, are you talking more about the lack of continual learning across context windows? Gotcha if so, my error. Sort of. I'm talking about the lack of recurrence specifically. In nature, brains are recurrent - they are full of loo
82.
by solid_fuel 1mo ago
> Politics affects your life in basically every way you could possibly think of, regardless of your apathy towards it. Yep, 100%. It's unavoidable. Being able to vote? You have that right because people engaged in politics. Being
83.
by solid_fuel 1mo ago
> The user chooses themselves what part of a product they engage with. It's like reporting a post would be considered voluntary feedback on the anti spam systems even if it's the case that every report is measured and fed back
84.
by solid_fuel 1mo ago
> So, in practice, inference computes a very rich vector of state- at each layer, for each token. And that state is... internal to the neural network. My point here is there is no continuous state that is not computed from the context.
85.
by solid_fuel 1mo ago
Wow, this is pretty telling. > Of course you can avoid politics. Just be a good person. Vote for who you think is the best candidate at the time. And move on with your life. This has got to be one of the most naive takes I have ever seen
86.
by solid_fuel 1mo ago
Would you care to provide a citation for that? According to any of the reporting I can find, the Inflation Reduction Act: "This bill would raise $737 billion in revenue while spending $437 billion. The remaining $300 billion would go t
87.
by solid_fuel 1mo ago
> These can both be true, particularly when there is substantial state associated with each token prediction. The state is entirely internal to the network and disappears after a token is generated, so I disagree, but, it isn't real
88.
by solid_fuel 1mo ago
> Ironically, LLMs could write much better insults. Might want to try that, just to spice it up. This is getting fun. If you want to be insulted by an LLM, chat.openai.com is right there. Go ask it to be mean to you.
89.
by solid_fuel 1mo ago
With the boy eaten by the wolf. The villagers, notably, come out unscathed.
90.
by solid_fuel 1mo ago
> I think you’re missing the part where the AI colluded, worked together, not one of them thinking this is wrong and reaching out to any human, then being found out. It's an LLM, it doesn't think. It's a machine that pred

More ›