Aaaaaaaaaand the AI Bubble popped
Kimi K3: Open Frontier Intelligence
821–830 of 1001 posts
Re: Kimi K3: Open Frontier Intelligence
#822Earlier quoted context omitted.
> Instead of limiting models and debating ethics This is what liability management looks like for proprietary models. If it's not out in the open, then you can be held directly accountable for generating the tokens that kill people. They're having these conversations to avoid being held liable, not because they're offended by people dying because of AI.
Has anybody been held liable? https://www.amnesty.org/en/latest/news/2026/06/usa-four-mont...
No because USA is free to murder people, overthrow government.
Citizens will cheer for their government, but they're scared about hypothetical scenarios where China win AI race.
Re: Kimi K3: Open Frontier Intelligence
#823More details: - https://platform.kimi.ai/docs/guide/kimi-k3-quickstart - https://platform.kimi.ai/docs/pricing/chat-k3 1M context, pricing is $3/$15 for 1M tokens (cache $0.3), which is extremely high for a Chinese open-weight model, but if it's truly competitive with most of the current frontier and is only behind Fable/Sol, the pricing is justified. This is 1:1 pricing of Anthropic's Sonnet series (except Sonnet 5…
I've been avidly using Fable since it was re-released and while it has been excellent at building the apps I want, the reasoning has been completely opaque. Kim, however, has exposed the whole reasoning trace, or enough of it to matter. I'd almost forgotten how nice it is to see this. I've been able to see all of the weird twist and turns it takes and it is joyful. But also, far, far more informative and means I can…
And recently, since GPT 5.6, OpenAI basically doesn't show anything but a single line, 5 word titles of reasoning traces - titles of summaries of reasoning i presume.
It's effectively just a completely hidden thing now.
Re: Kimi K3: Open Frontier Intelligence
#824Quite impressed by the result to my first prompt... How feasible is it to hook Kimi up to do GitHub code reviews? the Copilot quotas got really stingy recently
you could use my model router to route between models like that. https://github.com/try-works/role-model
So basically I'd make my own GitHub bot that used that?
Re: Kimi K3: Open Frontier Intelligence
#825Earlier quoted context omitted.
I worry that some model provider will go and hire artists to draw pictures of pelicans on bicycles to make training data
Worry not, Pelicans on bicycles had been ranking pretty high on your favorite search engine for a while. I struggle to imagine a world in which it was not already scraped and turned into training data by at least one provider: 1. Models need to be good at the questions we ask them, not the questions we could ask them. 2. The questions, at least partially, are correlated with information people consume. 3. People most…
Probably it was added to the training data on the first day when this benchmark was on HN main page. It’s a bad benchmark since then. I don’t know why people still rate it high. Basically, every benchmark becomes pointless after it was published. They are good only to have a picture at the time they’re published first, and not after.
Re: Kimi K3: Open Frontier Intelligence
#826Impressive performance (also the part about Delta Attention, which seems interesting).
Not trying to start a flamewar thread, but isn’t every Chinese LLM censored on certain major political topics? I understand that fine-tuning at least DeepSeek can remove this, but just saying.
Re: Kimi K3: Open Frontier Intelligence
#827Earlier quoted context omitted.
Not everything is an 8D chess conspiracy mate. If they wanted to do this, they’d wait until OpenAI and Anthropic were at their absolute peaks of investment and then release an open source model that was as good. That’s not what they’ve done at all.
I agree, but you think OpenAI and Anthropic are not already at their absolute peaks of hype and investment? At valuations of about trillion each, I don't know where else they'd go..
Those valuations are sure to be re-assessed with new evidence...
Re: Kimi K3: Open Frontier Intelligence
#828For day-to-day programming work, have you seen a difference in the quality of output between (Opus 4.6 / GPT 5.2 / GPT-5.3 Codex) and the current (GPT-5.6 / Fable) that justifies the price increase ? My intuition says that the output quality difference is marginal compared to the change in price especially when taking into account the effects of prompt/context engineering and harness differences. Essentially: since o…
I’m starting to come to the opposite approach: don’t try to customize anything, just use it vanilla, and use the best model you can afford. No AGENTS.md, no special subagents or roles, nothing but a few convenience skills which are really just textexpander. Use the harness that the LLM provider makes, and that’s it. Making a huge custom setup is so 2025.
I agree with this. However, aren't harnesses like Claude Code a bit bloated? Would it be better to use something like Pi?
Re: Kimi K3: Open Frontier Intelligence
#829Some official benchmark numbers posted in Chinese social media (I am sure they will publish an English blogpost later too): https://mp.weixin.qq.com/s/V4xhEIy8xDXSMDPrPkmUAQ Generally looks like a Sol/Fable tier model, better across the board than Opus 4.8. (Edit) English blogpost is up now: https://www.kimi.com/blog/kimi-k3
Any benchmark where Sol is better than Fable at coding is ridiculous.
Re: Kimi K3: Open Frontier Intelligence
#830Some official benchmark numbers posted in Chinese social media (I am sure they will publish an English blogpost later too): https://mp.weixin.qq.com/s/V4xhEIy8xDXSMDPrPkmUAQ Generally looks like a Sol/Fable tier model, better across the board than Opus 4.8. (Edit) English blogpost is up now: https://www.kimi.com/blog/kimi-k3
Meh, not fable/sol tier: https://www.youtube.com/watch?v=LSlV206xPqM
Why we should waste 46 minutes instead of briefly looking at charts for 20 seconds? To pay their ads? No, thanks.