311–320 of 393 posts
Earlier quoted context omitted.
If your moat is “please don’t copy my outputs”, you don’t have a moat. There is no such thing as a distillation “attack”.
How very machiavellianist-libertarian of you. Don't even try to combine it with any notion of "leadership" then, however, since distillation is literally "copying the actual leader"
Anyone want to bet that much like speculative execution, speculative decoding is going to introduce a whole slew of vulnerabilities in the ways LLMs work?
DeepSeek continues to not only push the boundaries but also publish these incredible papers explaining how they achieved their gains - something the American labs no longer do unfortunately. Chinese labs are doing the most interesting work in AI right now.
>publish these incredible papers explaining how they achieved their gains - something the American labs no longer do unfortunately. Google is still releasing a lot of llm architecture research. They introduced speculative decoding of LLMs in 2022[1], then released the code to perform sceculative decoding for their Gemma 4 model this year[2] [1] https://arxiv.org/abs/2211.17192 [2] https://github.com/google-gemma/cook…
Earlier quoted context omitted.
They weren't the first to do MTP like this, and arguably did it wrong: the MTP heads are kept in a separate file and have to be welded in by the inference engine. Qwen 3.6 shipped with working MTP first, and had working MTP in llama.cpp first.
Given the MTP drafter is basically a separate model, keeping it separate makes more sense IMO. It's out of my wheelhouse but it seems like you could adjust the MTP drafter model separately from the main model, too. Ultimately though the real explanation, I think, is Google doesn't care since for their own purposes (in LiteRT-LM), they do bundle them. As far as I know, anyway.
Earlier quoted context omitted.
[flagged]
I'm quite certain I was criticising a set of ideas, not you personally. That I misunderstood your point in context is a different issue, in which case, yup, my mistake.
> This form of cold, American libertarianism you espouse is pure poison in the body politic, both in this US and globally. It degrades all of human interaction to transactions. Its no wonder that the US is where sociopaths like Zuck were birthed.
Was an outright attack. But whatever, all water under the bridge, stuff like this happens. I might attack someone too if I misinterpreted something. Not a big deal.
> That I misunderstood your point in context is a different issue, in which case, yup, my mistake.
Thank you for taking the extra time to understand.
DeepSeek continues to not only push the boundaries but also publish these incredible papers explaining how they achieved their gains - something the American labs no longer do unfortunately. Chinese labs are doing the most interesting work in AI right now.
Earlier quoted context omitted.
No parent is right. The core root driver of the world is capitalism, open source exists downstream of that. Software engineers need money to survive. If they exclusively work on open source stuff where are they getting money from to survive? Follow the money trail… even a donation… eventually it leads to an incentive based source or action.
> If they exclusively work on open source stuff where are they getting money from to survive? From open source. You can earn money from open source. Open source is not opposed to capitalism, idk where you got that idea.
You’re right. One of its big uses is for big companies to destroy the ability of individual developers or small companies to make a living selling software. It’s also used to harm larger competitors whose revenue depends on some non open source software. Essentially, open source is the new “dumping”.
That’s not to say open source is bad. But in a capitalist context, it’s certainly a very double-edged sword.
I’ve been using DeepSeek v4 pro for a month now in Kilo Code and its great. Fast, reliable, large context window and cheap as… Did 1,5B tokens this month and cost me 40usd (majority cached, but still).
Is there a way to see how many tokes one does with claude code (pro)?
> Local-first session search, analytics, insights, and token use statistics for coding agents, supporting Claude Code, Codex, and more than 20 other agents.
solid piece of software
Earlier quoted context omitted.
Chinese labs are also still behind, so they’re incentivized to collaborate and have no reason to do it in private. I suspect their tune will change if they ever take the lead..
Not everyone is motivated by greed
Earlier quoted context omitted.
The question is also what game they're playing. Deepseek came out of a hedge fund. I think it's no coincidence that their publications tend to have a large impact on AI stock prices. Destroying the growth story of overvalued stocks is an interesting investment strategy. It's not even new. Shortsellers understandably get terrible rep from execs, but their actions are more often in the public interest than you'd think.…
> The question is also what game they're playing. Deepseek came out of a hedge fund. I think it's no coincidence that their publications tend to have a large impact on AI stock prices. Its revealing that they always seem to publish after some big announcement by American AI companies. But regardless, this is one of the benefits of a duopoly.
There's always an announcement by one of the frontier labs.