Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Kimi K3: Open Frontier Intelligence

kimi.com

Post reply on HN
Print view View original

991–1000 of 1001 posts

Re: Kimi K3: Open Frontier Intelligence

#991

Post by preg_match »

bottlepalm wrote: post #472
lukebechtel wrote: post #386

> Chip Design > As an early proof of concept, Kimi K3 designed a chip to serve a nano model built on its own architecture. In a single 48-hour autonomous run, K3 built, optimized, and verified the chip using open-source EDA tools on the Nangate 45nm library. Within 4 mm², the chip closes timing at 100 MHz and sustains over 8,700 tokens/s decode throughput in simulation, packing 1.46M standard cells, 0.277 MB of SRAM,…

Really feels like end game type stuff - AI designing its own next versions, designing its own chips, etc.. The advancement is slow, but fast - like a plant growing. We really are the boiling frogs now aren’t we? And the people with eyes wide open are us, and anyone that frequents this site really. Is this Milliways?

The bottleneck is physical stuff: chip fab is still a whoooooole beast. Doesn't matter too much if AI can design chips, the lead times are still in the years. Assuming you use existing fabs. If you need new fabs, now it's decades.

And there's not too too much that can be done here. Robotics, sure, but robotics are very behind AI because physical space is just hard. We can't, right now, just spin up like 1000 robots and build a fab.

Re: Kimi K3: Open Frontier Intelligence

#992

Post by preg_match »

criley2 wrote: post #284
cavemandaveman wrote: post #245

Earlier quoted context omitted.

Can you host the model for a lower cost per token than you'd pay Anthropic or OpenAI for a similar level of intelligence? I doubt you're beating their efficiencies of scale.

No, and the reason is simple: Usage is bursty and if you don't maximize usage of the hardware you're going to lose on price. Ok you can host this model once. What if I want a dozen subagents? Ok you can host it 12 times at once. What if we go a whole week only using max 4 at a time? Etc etc. The limits imposed by self-hosting might be bearable for a variety of reasons, but it's going to be more expensive and less con…

The flip side is if you already have the hardware and can utilize it. Also, if this is for security or IP concerns, you largely don't have a choice. The US players are out.

The marginal cost goes down significantly if you have datacenters. Baring in mind that US API pricing is kind of absurd. Even if you, say, only utilize your DC 1/10 of the time... you might still be ahead of API pricing by a wide margin.

Re: Kimi K3: Open Frontier Intelligence

#993

Post by katzenversteher »

rhdunn wrote: post #873
katzenversteher wrote: post #718

Earlier quoted context omitted.

Hey Simon, I noticed one thing all LLMs are currently pretty bad at and maybe we could create a benchmark from it. Let an LLM play the role of a dungeon master and tell it to strictly stay in the script/story and only allow realistic player actions. You will notice that they are easily brought off track. E.g. - Tell the LLM that you as a player noticed a strange glow in an NPCs eyes -> the NPC becomes an enemy. - In…

Have you tried the https://huggingface.co/LatitudeGames models? They are used by the https://play.aidungeon.com website, but can also be downloaded and used with llama-server in conjunction with something like SillyTavern. But in general, I've experienced things similar to you. I've also found that LLMs are bad at subtext, e.g. hinting at an NPC being a werewolf or vampire.

I have not used them directly but I also experimented with aidungeon and basically found all the issues I mentioned in my original message.

Re: Kimi K3: Open Frontier Intelligence

#994

Post by epistasis »

qeternity wrote: post #357
epistasis wrote: post #349

Earlier quoted context omitted.

I have severe complaints about Anthropic's product managers on this front. Their preference for hiding, obscuring, and trying to wrest control from the user are a bit harrowing. It would be wonderful to go back to Claude Code from before March. It seems like every release destroys value for me!

It's a defensive tactic to reduce the effectiveness of distillation. Say of that what you will, but it's not because they want to wrest control from users. It's because they don't want Chinese companies to do exactly what Moonshot (Kimi creators) and others have done.

It's funny because Anthropic is harming my use of their product, to not even stop the supposed theft of their thought chains. Something that any other supposed upstarts could presumably grab from Moonshot or whatever now.

And my complaint also extends to all their tool use explanations, or rather the lack thereof. I get prompted continually for tool use that I can't examine, that's a poorly formatted 1kb bash script, etc. The PM desire to hide valuable information while requiring extensive interaction has really driven the product into a very unusable place compared to where it was a few months ago. (Or perhaps that's just Claude responding to its memory of my use, and I have somehow driven it to be excessively verbose and difficult to use, which would be unfortunate... Perhaps there's a way to reset the memory.)

Re: Kimi K3: Open Frontier Intelligence

#995

Post by oblio »

anigbrowl wrote: post #702

This might be the most impressive website generator demo I've seen: https://macos27.kimi.page Context from the person who prompted it: https://x.com/mweinbach/status/2077827886149439547

Interesting, it reminds me a bit of the Linux in the browser thing.

Re: Kimi K3: Open Frontier Intelligence

#996

Post by oblio »

rmonvfer wrote: post #964
anigbrowl wrote: post #702

This might be the most impressive website generator demo I've seen: https://macos27.kimi.page Context from the person who prompted it: https://x.com/mweinbach/status/2077827886149439547

Wait until you see the bare metal version (by the same guy) https://x.com/mweinbach/status/2078231897843302517 https://github.com/mweinbach/swift-os

Interesting. I wonder if this could work in the opposite direction, forcefully opening Apple hardware to other OSes. Think not just Asahi Linux, but regular Linux distros on Apple hardware.

Re: Kimi K3: Open Frontier Intelligence

#997

Post by oblio »

m3h wrote: post #86

> Kimi K3 is Kimi’s most capable model to date, with 2.8 trillion parameters. This puts them on the top of the largest open models list: Kimi K3 2.8T DeepSeek-V4-Pro 1.6T (49B active) Kimi K2.6 ~1T (32B active) GLM-5.2 754B (40B active) DeepSeek-V3.2 685B Mistral Large 3 675B That's one mighty large model! Moonshot is going to need the USD 500 million reportedly raised earlier this year to run this model.

I wonder if we're a 5-10 years from running this on beefy consumer hardware, or more like 20+ years.

Re: Kimi K3: Open Frontier Intelligence

#999

Post by satvikpendem »

linzhangrun wrote: post #670
satvikpendem wrote: post #66

Now, will they actually release the weights? Seems like Chinese model providers are slowly closing up, like Alibaba's Qwen 3.6 which did release weights (but not the biggest parameter count ones) and none for 3.7.

Qwen's team has been reorganized, and not open-sourcing also fits Alibaba's usual corporate style... Other companies have not shown similar problems so far. China has many government agencies and state-owned enterprises that need models which can be deployed locally. This is also part of "Xinchuang" (self-owned systems, self-owned hardware, etc. New government computers all run special Linux versions on domestic CPUs…

Looks like I spoke too soon. As a result of Kimi competition Qwen is releasing their 3.8 weights as well.

Re: Kimi K3: Open Frontier Intelligence

#1000

Post by nkmnz »

sudosysgen wrote: post #44
nkmnz wrote: post #29

Earlier quoted context omitted.

> > ...ranks second only to Claude Fable 5 and GPT-5.6 Sol. So... it ranks THIRD?

The literal interpretation of that sentence is "when it is second or third, it is only behind Fable 5 or 5.6 Sol". And indeed they give benchmarks where it is ahead of one but not both models.

That would be “ranks second to x OR y”
Post reply on HN