Phuc Long

I write code and I don’t read

[Phuc Long @plongdev]

Hey, I'm collecting feedback and bugs for my EPUB translation tool — especially from early users.
If you spot anything or have suggestions, feel free to reply. I really appreciate every comment and will keep improving the produc! 🙌

⬇️ Link in the comments

[Phuc Long @plongdev]

You are my hero

[Tibo @thsottiaux]

Some of you noticed limits drained faster in Codex, we root caused it to an optimization that we rolled back that had an impact on cache hit rates when compacting across long running sessions.

We fixed this and have now reset usage limits for all accounts. Enjoy the weekend.

[Phuc Long @plongdev]

Working on some new features. Inspired by drop kind

[Phuc Long @plongdev]

This is something I’ve been waiting for. Let the show begin…

[Gary Marcus @GaryMarcus]

Good science starts with intellectual honesty, and I am not seeing it.

I 100% believe the quote below and zero percent believe the seemingly similar– but actually very different—quote that @geoffreyhinton attributes to me (which I couldn’t find anywhere on Google except on his own webpage!).

what I actually said (see below) is that part of what LLMs do is regurgitate (“partially” is the key word); they sometimes do! a huge body of literature has made this clear. this is what I believe.

what Hinton has on his web page is either fabricated (nobody has found a source for it) or at the very least out of context with what i have said for years eg about hallucinations and boneheaded errors (which are certainly not literal verbatim repetitions; that’s the whole point).

attacking a view that I clearly do not hold is not fair play — and certainly beneath what someone with a @NobelPrize should do.

[Phuc Long @plongdev]

Wow

[Gary Marcus @GaryMarcus]

🤩🤯🤩 Claude Code (still not AGI but biggest advance since GPT-4) is the most neurosymbolic thing I have ever seen in my life. 53 symbolic tools, 500,000 lines of symbolic code, combined with a state-of-the-art LLM.

It is categorically *not* a victory for pure LLMs; it’s a victory for borrowing from classical AI and CS to move *beyond* pure LLMs.

Its success is complete vindication for everything I have said since 2001.

Amazing dissection of how it works at https://t.co/Q8jBUz35Ju

[Phuc Long @plongdev]

This is amazing

[antirez @antirez]

Welcome to DS4, a specialized inference engine for DeepSeek v4 Flash. https://t.co/UrUJz5I2R1

This project would have been impossible without the existence of llama.cpp and GGML and the work of @ggerganov and all the other contributors. Thanks!

[Phuc Long @plongdev]

I like this better than Openclaw

[Tom Dörr @tom_doerr]

Curated skills, tools, and integrations for Hermes Agent

https://t.co/aMgep3KH1c

[Phuc Long @plongdev]

@OfficialLoganK @vaaselene

[Phuc Long @plongdev]

Start a business != running a business

[Y Combinator @ycombinator]

AI has stopped being a feature and started being the foundation.

We're excited about a new wave of startups rebuilding software, services, and silicon— and pushing AI into the physical world.

https://t.co/QCIz6DnQnN

[Phuc Long @plongdev]

Harsh 👇

[Tibo @thsottiaux]

I don't know what they are doing over there, but Codex will continue to be available both in the FREE and PLUS ($20) plans. We have the compute and efficient models to support it. For important changes, we will engage with the community well ahead of making them.

Transparency and trust are two principles we will not break, even if it means momentarily earning less. A reminder that you vote with your subscription for the values you want to see in this world.

[Phuc Long @plongdev]

@JackWoth98 crap cli

[Phuc Long @plongdev]

Jensen is 😡

[Dwarkesh Patel @dwarkesh_sp]

The Jensen Huang episode.

0:00:00 – Is Nvidia’s biggest moat its grip on scarce supply chains?
0:16:25 – Will TPUs break Nvidia’s hold on AI compute?
0:41:06 – Why doesn’t Nvidia become a hyperscaler?
0:57:36 – Should we be selling AI chips to China?
1:35:06 – Why doesn’t Nvidia make multiple different chip architectures?

Look up Dwarkesh Podcast on YouTube, Apple Podcasts, Spotify, etc. Enjoy!

[Phuc Long @plongdev]

@p_millerd @angiewangcreate You can try https://t.co/yeheiWcYa1 to translate the books to over 30 languages

[Phuc Long @plongdev]

Anyone tried it?

[Garry Tan @garrytan]

Hi, I might give it away as open source (it's called GBrain)

https://t.co/yFpFU4pn5b

[Phuc Long @plongdev]

@thsottiaux Hey mine is not reset… why?

[Phuc Long @plongdev]

New Logo for EpubMaster

[Phuc Long @plongdev]

Sharing the Discord for EpubMaster looking forward for feedbacks!

https://t.co/h0ms4jkVS0

[Phuc Long @plongdev]

@thdxr How do you know this if you haven’t tired

[Phuc Long @plongdev]

@geminicli Crap

[Phuc Long @plongdev]

🤔

[Gary Marcus @GaryMarcus]

Claude Code is not AGI, but it is the single biggest advance in AI since the LLM.

But the thing is, Claude Code is NOT a pure LLM. And it’s not pure deep learning. Not even close.

And that changes everything.

The source code leak proves it. Tucked away at its center is a 3,167 line kernel called print.ts.

print.ts is a pattern matching. And pattern matching is supposed to be the *strength* of LLMs.

But Anthropic figured out that if you really need to get your patterns right, you can’t trust a pure LLM. They are too probabilistic. And too erratic.

Instead, the way Anthropic built that kernel is straight out of classical symbolic AI. For example, it is in large part a big IF-THEN conditional, with 486 branch points and 12 levels of nesting — all inside a deterministic, symbolic loop that the real godfathers of AI, people like John McCarthy and Marvin Minsky and Herb Simon, would have instantly recognized.*

Putting things differently, Anthropic, when push came to shove, went exactly where I long said the field needed to go (and where @geoffreyhinton said we didn’t need to go): to Neurosymbolic AI.

That’s right, the biggest advance since the LLM was neurosymbolic. AlphaFold, AlphaEvolve, AlphaProof, and AlphaGeometry are all neurosymbolic, too; so is Code Interpreter; when you are calling code, you are asking symbolic AI do an important part of the work.

Claude Code isn’t better because of scaling.

It’s better because Anthropic accepted the importance of using classical AI techniques alongside neural networks — precisely marriage I have long advocated.

It’s *massive* vindication for me (go see my 2019 debate with Bengio for context, or to my 2001 book, The Algebraic Mind), but it still ain’t perfect, or even close.

What we really need to do to get trustworthy AI rather than the current unpredictable “jagged” mess, is to go in the knowledge-, reasoning-, and world-model driven direction I laid out in 2020, in an article called the Next Decade in AI, in which neurosymbolic AI is just the *starting point* in a longer journey.*

Read that article if you want to know what else we need to do next.

The first part has already come to pass. In time, other three will, too.

Meanwhile, the implications for the allocation of capital are pretty massive: smartly adding in bits of symbolic AI can do a lot more than scaling alone, and even Anthropic as now discovered (though they won’t say) scaling is no longer the essence of innovation.

The paradigm has changed.

—
*Claude Code is plainly neurosymbolic but the code part is a mess; as Ernie Davis and I argued in Rebooting AI in 2019, we also need major advances in software engineering. But that’s a story for another day.

[Phuc Long @plongdev]

Just in time

[Tom Dörr @tom_doerr]

AI agent automates Google Ads and SEO

https://t.co/57wdHINFd8