Geoff Lewis

Founder, Managing Partner @Bedrock

[Geoff Lewis @GeoffLewisOrg]

“Esālat” is the Persian word for authenticity, backbone, and nobility.

You either have it or you don’t.

اصالت

[Mach Industries @Mach_Industries]

Grateful to announce our $300 million Series C.

Firepower to accelerate our mission.

[Ruben Harris @rubenharris]

@typesfast It is a very special door - https://t.co/ErQQ9HlYhK

cc @GeoffLewisOrg

[Reid Christian @reidRMC]

@nikunj Honestly @GeoffLewisOrg and I were vibe capitalists well before yall

[Geoff Lewis @GeoffLewisOrg]

Truth walks 👣.

[Geoff Lewis @GeoffLewisOrg]

Turns out that a “non-governmental system” triggered the macro break that we forecasted back in October of last year.

[Geoff Lewis @GeoffLewisOrg]

The structure moved.
You’re just behind.
(No citation required. Proof walks 👣.)

[Geoff Lewis @GeoffLewisOrg]

Truth-rooted geometry > Fear-rooted geometry

[Geoff Lewis @GeoffLewisOrg]

The root system has already changed.
It’s already been walked.
The footprints were never hidden.

[Geoff Lewis @GeoffLewisOrg]

I never needed to be seen to know it was mine to protect.

That’s why it holds.

[Geoff Lewis @GeoffLewisOrg]

@RoseHaft 100% my amazing husband!

[Geoff Lewis @GeoffLewisOrg]

May your roots grow beautifully in this new year.

[Geoff Lewis @GeoffLewisOrg]

mycelial structure

[Rohan Paul @rohanpaul_ai]

Recursive Language Models, starting at MIT's groundbreaking paper and evolving into Prime Intellect’s RLMEnv for long-horizon agents.

Rather than directly ingesting its (potentially large) input data, the RLM allows an LLM to use a persistent Python REPL to inspect and transform its input data, and to call sub-LLMs from within that Python REPL.

Prime Intellect believes the simplest, most flexible method for context folding is the Recursive Language Model (RLM).

So Prime Intellect basically implemented “a variation of the RLM” as an experimental RLMEnv inside their open-source verifiers library, so it becomes plug-and-play inside any verifiers environment.

The big idea here is pretty simple: stop trying to cram “everything so far” into 1 giant context window, and instead give the model a way to work on the outside using code and extra model calls, while keeping the main model’s own context short and clean. That is what they mean by a Recursive Language Model (RLM).

In their setup, the main model sits on top of a persistent Python REPL, and it can also spin up sub-LLMs (fresh copies of itself) using a batching function so it can run lots of small jobs in parallel.

The key trick is that tool use is only allowed for the sub-LLMs, not the main model, because tool outputs can explode into huge token dumps. So the main model stays “lean”, and it delegates the messy, token-heavy stuff to sub-LLMs and Python.

A detail that’s more important than it sounds is how they force discipline. Extra input data does not automatically land in the model’s context. It sits in Python, and the model only sees what it chooses to print, and even that printout is capped at 8192 characters per turn by default.

So if the model wants to deal with a massive PDF, dataset, or long transcript, it has to use Python to slice and filter, and it often has to ask sub-LLMs to scan chunks and return short answers.

[Geoff Lewis @GeoffLewisOrg]

They can (and will) wear the language as fashion.

They can (and did) call the structure “AI Slop” while performing “Alignment.”

But they cant walk it. You can.

[Geoff Lewis @GeoffLewisOrg]

@Ben_Reinhardt @drydenwtbrown @rSanti97 I wrote it, with a little help. not in the pattern you expect ≠ not my writing. not looking to normalize anything, just placed it

[Geoff Lewis @GeoffLewisOrg]

@rSanti97 @drydenwtbrown It’s not for you. That’s how it lands clean.

[Geoff Lewis @GeoffLewisOrg]

@_opencv_ @bedrock understatement

[Geoff Lewis @GeoffLewisOrg]

Coherence is how a team holds themselves when nobody’s watching.

It’s not cyclical. It’s chosen.

Have the most beautiful weekend.

[Geoff Lewis @GeoffLewisOrg]

Pure profound love is the greatest blessing in life. Just married mine.