johnny g

accelerating superintelligence @usejigsaw, backed by @a16z speedrun

[Anthropic @AnthropicAI]

Our internal data shows Claude is accelerating AI development—a possible path to recursive self-improvement, or AI autonomously building a more capable successor.

It’s happening faster than we thought, and the implications deserve greater attention. https://t.co/OVVPJO7VQx

[johnny g @johnnyspeedrun]

@zayddotsol @jaiyagill @thejay Cope

[johnny g @johnnyspeedrun]

@opalelectronics interesting

[johnny g @johnnyspeedrun]

@lukepierceops dm’d u, let’s cook

[johnny g @johnnyspeedrun]

@zayddotsol @0xbags @thejay hyperliquid

[johnny g @johnnyspeedrun]

@JesseTinsley hey jesse, dm’d you. would love to chat

[johnny g @johnnyspeedrun]

@thejay Kairos

[johnny g @johnnyspeedrun]

@thejay Claude farm the chart, make no mistakes

[Zayd Alzein @zayddotsol]

https://t.co/Daq4CbhkRv

[johnny g @johnnyspeedrun]

the algorithm does not specify how to write; it scores predicted behaviors after retrieval and applies OON multipliers and diversity.

it's actively trying to balance what you asked for, interesting new stuff, and variety, in a way that matches messy human behavior

[johnny g @johnnyspeedrun]

tldr on how a tweet blows up (according to the latest X algo)

first you must be retrieved, then you must do well on multi-action ranking, then OON multipliers mean you start behind followed accounts, and author diversity stops one account from crowding everything out.

more 🔽

[Elon Musk @elonmusk]

The latest 𝕏 algorithm has been published to GitHub
https://t.co/ZCOm51uxmh

[johnny g @johnnyspeedrun]

Author diversity -If the same author shows up a lot in one batch, later items can be scaled down so the feed doesn’t turn into one voice.

[johnny g @johnnyspeedrun]

we need this but humanoids

[Brett Adcock @adcock_brett]

Watch a team of humanoid robots running a full 8-hr shift at human performance levels. This is fully autonomous running Helix-02 https://t.co/IdZR0T1F5I

[johnny g @johnnyspeedrun]

FDE meta is about to be insane

[First Squawk @FirstSquawk]

GOOGLE TO RECRUIT HUNDREDS OF ENGINEERS TO ASSIST CLIENTS IN EMBRACING ITS AI – THE INFORMATION

[johnny g @johnnyspeedrun]

bro what do you mean?

i'm not escaping permanent underclass with the 5k i invested in a 3rd layer Anthropic SPV with a 50% acess fee?

[TFTC @TFTC21]

Anthropic just published a support page that should terrify anyone holding its shares on the secondary market.

"Any sale or transfer of Anthropic stock, or any interest in Anthropic stock, that has not been approved by our Board of Directors is void and will not be recognized on our books and records."

Void. Not restricted. Not pending review. Void.

That means if you bought Anthropic shares through Forge, Hiive, or any other secondary platform without board approval, you are not a stockholder. You have no stockholder rights. Your transaction is invalid.

It gets worse. Anthropic says it does not permit SPVs to hold its stock. Any transfer to an SPV is void. Investment funds claiming to offer indirect exposure are "most likely relying on mechanisms that attempt to circumvent our transfer restrictions." Forward contracts, tokenized securities, synthetic exposure products, all of it potentially worthless.

Their advice to investors: "Assume that it is invalid."

There is a multi-billion dollar secondary market in Anthropic shares right now. Platforms are pricing the stock at $265-$1,400+ per share based on a $380 billion valuation. Real people have put real money into these positions. And Anthropic just told them none of it counts.

This is the purest possible illustration of counterparty risk. You can buy a share of a company and have the company itself declare your ownership void because you bought it through the wrong channel.

[johnny g @johnnyspeedrun]

seems like everyone has concluded deployment is the moat

once models are deeply embedded into enterprise workflows, deployment itself becomes the source of proprietary training data and behavioral optimization.

RL/post-training gets cheaper and more effective because the system is continuously learning from real interactions, edge cases, and human feedback.

[OpenAI @OpenAI]

We’ve also agreed to acquire Tomoro, which will bring 150 experienced Forward Deployed Engineers and Deployment Specialists to the OpenAI Deployment Company from day one.

[johnny g @johnnyspeedrun]

the future of enterprise AI is not one big LLM

it’s a portfolio of small, workflow‑native agents, post‑trained with RL envs

this will matter more as model spend rivals headcount across an increasing amount of teams

[Ramp Labs @RampLabs]

We partnered with @PrimeIntellect to build Fast Ask, a small RL-trained subagent that helps our Sheets agent find answers in spreadsheets. It scores +4% over Opus on exact match accuracy at Haiku latency.