91–100 of 125 posts
Something is wrong at moonshot.
They may have all this amazing architecture but Kimi has been super dumb recently. I reckon they’re under compute pressure and quantising to stay afloat. I was a heavy k2.5/2.6 user earlier in the year and built serious features with it, but even k3 now does stupid shit like fail tool calls and get stuck in endless thought trains. K2.6 was spinning its wheels on a problem for over 10 minutes today and then deepseek v…
They may have all this amazing architecture but Kimi has been super dumb recently. I reckon they’re under compute pressure and quantising to stay afloat. I was a heavy k2.5/2.6 user earlier in the year and built serious features with it, but even k3 now does stupid shit like fail tool calls and get stuck in endless thought trains. K2.6 was spinning its wheels on a problem for over 10 minutes today and then deepseek v…
Deepseek is also struggling with simple tasks for the last two days. Claude is perfect, with the additional credits.
Earlier quoted context omitted.
People don't remember now, but in the old HN comments like "this is so true honestly" would have been downvoted. If a comment just agrees with the parent, people would have said that's what the upvote button is for. Adding a "I agree" comment is just noise.
This is so true honestly.
Earlier quoted context omitted.
It's entirely reproducible from the available documentation (which is why you see vLLM, SGLang, MLX etc all racing to produce optimized implementations). (As an aside, this is why the "open weights are not open source" thing is a complete misunderstanding. The weights themselves along with the documentation give you enough to fine tune the LLM. You can't rebuild it from scratch, but you can't do this even with the da…
> You can't rebuild it from scratch > It's entirely reproducible from the available documentation You have a very interesting understanding of "reproducibility", I'll give you that :) But even with that, there are plenty of technical details (especially in regards to the training process) missing from the tech report that leads to these weights not being reproducible in any sense of that word.
He was talking about two different things, hence the parentheses. The architecture is reproducible, not the model weights.
They may have all this amazing architecture but Kimi has been super dumb recently. I reckon they’re under compute pressure and quantising to stay afloat. I was a heavy k2.5/2.6 user earlier in the year and built serious features with it, but even k3 now does stupid shit like fail tool calls and get stuck in endless thought trains. K2.6 was spinning its wheels on a problem for over 10 minutes today and then deepseek v…
The expression (vt −St−1 kt )ktT is literally the gradient step of a linear regression model updating its weights (S) on a single sample using Mean Squared Error loss.
Interesting what other labs comes up with from this.
[flagged]
Earlier quoted context omitted.
> You can't rebuild it from scratch > It's entirely reproducible from the available documentation You have a very interesting understanding of "reproducibility", I'll give you that :) But even with that, there are plenty of technical details (especially in regards to the training process) missing from the tech report that leads to these weights not being reproducible in any sense of that word.
> You have a very interesting understanding of "reproducibility", I'll give you that :) He was talking about two different things, hence the parentheses. The architecture is reproducible, not the model weights.
But what's the point of even saying that? Of course it is, otherwise how is it supposed to run in the runtimes? You cannot release model weights that others can run, without also releasing the model architecture, it's in the code at the very least...