Leon_limestone 🔭

Witness

[Leon_limestone 🔭 @leon_limestone]

Ethereum/Base: AI existence
Solana: AI action
zkML layers: AI verification
GPU networks: AI intelligence

[Game @game_for_one]

Chamath put out an 80 page dive a few days ago on the agent economy. Didn’t read all of it yet, but the headline numbers are real - AI writes 75% of Google’s new code now. Claude Code went from zero to 134k daily commits in under a year.

But there’s an issue if you look at the headlines in practice, Microsoft is already pulling back Claude Code access. Uber burned their entire 2026 AI coding budget in four months. Less than 10% of orgs have agents running at real scale.

I don’t think it’s a capability problem. The models work. It’s a cost problem. Enterprise budgets are built for fixed predictable spend. Agent workloads don’t work like that, you need a hundred machines for 30 seconds, not one machine forever. The subscription model breaks the moment you actually try to scale.

The harness layer probably wins as models commoditize. But underneath all of that is a pricing problem nobody’s solved yet.

> Compute that actually matches how agents run burst, on demand, pay for the outcome.

I think this is one to look out for, at some point we should see projects building this pop up.

[Ansem @blknoiz06]

crypto is indisputably the best at two things

1) engineering massive scale human coordination games
2) enticing speculation

whoever figures out how to combine both of these things together around raw compute capacity will make infinite

[Leon_limestone 🔭 @leon_limestone]

Imagine you still have the opportunities to buy into exponentially increasing "consensus."

Like Bittensor or even Bitcoin.

[Leon_limestone 🔭 @leon_limestone]

conviction szn

[Leon_limestone 🔭 @leon_limestone]

proof of intelligence
proof of inference
proof of research
...
proof of work🤖

[Leon_limestone 🔭 @leon_limestone]

rotating profits into a2a infra

[Leon_limestone 🔭 @leon_limestone]

next 1b runner on eth
it's happening

[Leon_limestone 🔭 @leon_limestone]

next 1b runner
let's see

[Leon_limestone 🔭 @leon_limestone]

ethmaxxxxing

[Leon_limestone 🔭 @leon_limestone]

@redarvian 把稀有图片套利者比喻成bitcoin miner是不是不太恰当？套利者的操作成本太低，他们给整个系统运作的贡献是如何体现的

[Leon_limestone 🔭 @leon_limestone]

@Sencrazy_1 market bought 0.5%. Welcome back

[Leon_limestone 🔭 @leon_limestone]

文艺复兴👽

[Dan McAteer @daniel_mac8]

Autonomous, recursively self-improving AI researchers are here.

They're just not evenly disributed.

[Viv @Vtrivedy10]

https://t.co/n1HUIEbky8

[Andrej Karpathy @karpathy]

LLM Knowledge Bases

Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So:

Data ingest:
I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them.

IDE:
I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides).

Q&A:
Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale.

Output:
Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base.

Linting:
I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into.

Extra tools:
I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries.

Further explorations:
As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows.

TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

[Leon_limestone 🔭 @leon_limestone]

The beta play of Bittensor is not subnet but something reinforce each other in different layers

[Leon_limestone 🔭 @leon_limestone]

RT @DistStateAndMe: Jack Clark (@jackclarkSF ), cofounder of Anthropic, has now written about @tplr_ai twice.

Two weeks ago we went viral.…

[Lukas (computer) (In SF) 🔺 @SCHIZO_FREQ]

This is always how I assumed LLMs would wind up functioning because this is how I (and presumably most others) think

I assume the base unit of thought is this gestalt thought vector thing, not "words," and we've just all developed a very fast way to translate these to words because words are more communicable than thought pieces

This was always my issue with "some people don't have an internal monologue!" discourse

It just makes no sense for words to be the base unit people think in. It's like 1000x faster to think in terms images or these thought pieces or whatever

I assume it just seems like people think in words bc when they describe what they're thinking to people, they have to translate the thought pieces to words - as that's how we communicate - and this process converts their actual thoughts into the form of a monologue

But it only makes sense to think in words when you need to output some form of communication. Otherwise it's not very efficient

And human brains are insanely efficient

[Christian Catalini @ccatalini]

1/ Karpathy just described the hard ceiling on trillion-dollar autonomous systems as a throwaway caveat about his overnight hyperparameter script (@NoPriorsPod) — and went back to tuning his learning rate...

[Wei Dai @_weidai]

Andrej Karpathy on autoresearch with an untrusted pool of workers:

"My designs that incorporate an untrusted pool of workers (into autoresearch) actually look a little bit like a blockchain.

Instead of blocks, you have commits, and these commits can build on each other and contain changes to the code as you're improving it.

The proof of work is basically doing tons of experimentation to find the commits that work."

The idea that distributed & permissionless autoresearch ~= proof-of-useful-work remains a high-level intuition for now, but it is extremely intriguing to say the least.

Someone needs to take this further. See QT for more on what's missing.

[Wei Dai @_weidai]

Is it possible to build "proof-of-useful-work" on top of autoresearch?

There's already great compute-versus-verification asymmetry that is tunable. Would need a reliable way to generate fresh & independent puzzles (that are still useful).

Maybe a dead end, but someone should look into if decentralized consensus with useful work is possible on top of autoresearch.

Let me know if you solve this.