[dead]
Claude Sonnet 4.5
811–819 of 819 posts
Re: Claude Sonnet 4.5
#812Earlier quoted context omitted.
Here is a definite scientific nail down and solve for non-determinism in LLM outputs (Mira Murati's new outfit but really credit the author) https://bff531bb.connectionism.pages.dev/blog/defeating-nond...
Requires a login?
Looks like it's probably https://thinkingmachines.ai/blog/defeating-nondeterminism-in...
Re: Claude Sonnet 4.5
#813Earlier quoted context omitted.
Why do you assume that LLMs are only coding Tailwind slop...? I've used LLMs extensively in C++ development of Unreal Engine games.
You went months without writing a single line of code by prompting in Unreal, or you didn't read the thread carefully and jumped to non-sequitur?
Your original comment implies that "vibe coding", AKA, using LLMs to code, produces suboptimal results. I'm telling you that I've used LLMs extensively to write Unreal Engine games that work well, not just "Tailwind slop".
You're not having a discussion in good faith.
Re: Claude Sonnet 4.5
#814Earlier quoted context omitted.
You went months without writing a single line of code by prompting in Unreal, or you didn't read the thread carefully and jumped to non-sequitur?
Your original comment implies that "vibe coding", AKA, using LLMs to code, produces suboptimal results. I'm telling you that I've used LLMs extensively to write Unreal Engine games that work well, not just "Tailwind slop". You're not having a discussion in good faith.
So you don't know what vibe coding is.
The comments I replied to cover it quite well:
> Very interesting observation. I haven’t written a function by hand in 18 months.
> Same. I haven't written any code by hand in some time. Oh well. I guess I'm just doing it wrong.
Using LLMs to help doesn't make it vibe coding: these are people claiming they write no code at all and only prompt agents.
You can't build a complex product with good UX/UI this way... but you can generate Tailwind slop.
Re: Claude Sonnet 4.5
#815Earlier quoted context omitted.
> A lack of determinism comes from many places, but primarily: 1) The models change 2) The models are not deterministic... models themselves are deterministic, this is a huge pet peeve of mine, so excuse the tangent, but the appearance of nondeterminism comes from a few sources, but imho can be largely attributed to the probabilistic methods used to get appropriate context and enable timely responses. here's an examp…
The User: The stove keeps burning me because I can't tell how hot it is, it feels random and the indicator light it broken. You: The most rigorous definition of temperature is that it is equal to the inverse of the rate of change of entropy with respect to internal energy, within a given volume V and particles N held constant. All accessible microstates are equiprobable over a long period of time, this is the very de…
Well really, the reason why I gripe about it, to use your example, is that then they believe the indicator light malfunctioning is an intrinsic feature of stoves, so they throw their stove out and start cooking over campfires instead, tried and true, predictable, whatever that means.
I think my deck of cards example still holds.
You could argue I'm being uselessly pedantic, that could totally be the case, but personally I think that's cope to avoid having to think very hard.
Re: Claude Sonnet 4.5
#816Earlier quoted context omitted.
> A lack of determinism comes from many places, but primarily: 1) The models change 2) The models are not deterministic... models themselves are deterministic, this is a huge pet peeve of mine, so excuse the tangent, but the appearance of nondeterminism comes from a few sources, but imho can be largely attributed to the probabilistic methods used to get appropriate context and enable timely responses. here's an examp…
The previous poster is correct for a very slightly different definition of the word "model". In context, I would even say their definition is the more correct one. They are including the random sampler at the end of the LLM that chooses the next token. You are talking about up to, but not including, that point. But that just gives you a list of possible output tokens with values ("probabilities"), not a single choice…
but, to be fair, simply calling the sampler random is what gives people the impression like what OP is complaining about. which isn't entirely accurate, it's actually fairly bounded.
this plays back into my original comment, which you have to understand to know that the sampler, for all its "randomness" should only be seeing and picking from a variety of correct answers, i.e. the sample pool should only have all the acceptable answers to "randomly" pick from. so when there are bad or nonsensical answers that are different every time, it's not because the models are too random, it's because they're dumb and need more training. tweaking your architecture isn't going to fully prevent that.
Re: Claude Sonnet 4.5
#817I haven't shouted into the void for a while. Today is as good a day as any other to do so. I feel extremely disempowered that these coding sessions are effectively black box, and non-reproducible. It feels like I am coding with nothing but hopes and dreams, and the connection between my will and the patterns of energy is so tenuous I almost don't feel like touching a computer again. A lack of determinism comes from m…
Maybe just take a weekend and build something by writing the code yourself. It's the feeling of pure creative power, it sounds like you've just forgotten what it was like.
Re: Claude Sonnet 4.5
#818Earlier quoted context omitted.
Yesterday someone posted an example of the same prompt but changing it to a human and it was basically trash, the example you've posted actually looks good all things considered. So yeah I do think its something they train on, same way they train on things in the benchmarks.
The easy way to tell is to try it yourself - run "Generate an SVG of a pelican riding a bicycle" and then try "Generate an SVG of an otter riding a skateboard" and see if the quality of the images seems similar.
How about a narwhal spacewalking from the ISS, with Earth visible below (specifically the Niger delta)?
https://claude.ai/public/artifacts/f3860a8a-2c7d-404f-978b-e...
Requesting an ‘extravagantly detailed’ version is quite impressive in the effort, if not quite the execution:
https://claude.ai/public/artifacts/f969805a-2635-4e30-8278-4...
Re: Claude Sonnet 4.5
#819Anecdotal evidence. I have a fairly large web application with ~200k LoC. Gave the same prompt to Sonnet 4.5 (Claude Code) and GPT-5-Codex (Codex CLI). "implement a fuzzy search for conversations and reports either when selecting "Go to Conversation" or "Go to Report" and typing the title or when the user types in the title in the main input field, and none of the standard elements match, a search starts with a 2s de…
Codex is like an outsourced contractor. You give them the spec and eventually you get a full feature to review.
Claude is more like a pair programmer, you iterate in short cycles and adjust.