Joseph Perla

Building something new | PhD in AI (dropout) | 4 time exited founder | Ex-Facebook, Lyft, Hangout, Princeton

[Joseph Perla @jperla]

@Altimor Use https://t.co/5szlor241X and it’ll be way easier to switch

[Joseph Perla @jperla]

Attestation is All You Need

https://t.co/3i9yV8Psik

Secure End to End LLM routing, all open source

[Bernie Sanders @BernieSanders]

I will soon be introducing a bill to give the public a 50% ownership stake in the largest AI companies in America.

This would guarantee that the trillions created by AI are used to improve the lives of all of us — and block oligarch decisions that harm the American people.

[Joseph Perla @jperla]

I want to make this but for skydiving

[Lucas Beyer (bl16) @giffmana]

Yooo wait a minute this is a super cool idea for AR glasses!

But it should go much further: instead of just a minimap, actually render the ghost in 3D in front of you like in Mario Kart.

And then... actually put some item boxes in the air that you need to jump to collect. And you can shoot your ghost with a green shell. And you can put banana peels for your ghost.

This may be what it takes to get me into running. Maybe a new side project, if nobody is doing this yet?

[Joseph Perla @jperla]

@TencentHunyuan Hey can we offer a 2-week free period of Hy3 on https://t.co/5szlor241X ?

[Joseph Perla @jperla]

@hypersoren I said this in May 2025 https://t.co/75zL0NRZAj

[Joseph Perla @jperla]

Chloe be

[Joseph Perla @jperla]

This is how u got perfect on my SATs

[Nishant Balepur @NishantBalepur]

🚨 New Paper! 🚨

One of my first Ph.D. papers found that LLMs can answer multiple-choice questions without seeing the question 🤔

At #ACL2026, I'm presenting a follow-up showing that current reasoning LLMs can still do this! And quite similarly to a clever test-taker 🧑‍🎓🧵

[Joseph Perla @jperla]

People are starting to realize that it’s a full blown virtua machine / runtime

https://t.co/MJtiALYcg7

[Jean-Michel Lemieux @jmwind]

Spent time today reading through one of Anthropic’s massive Claude prompts after sniffing the network traffic. A 1,500 line prompt!

What surprised me wasn’t the personality tuning. It was the amount of embedded operational logic. Large parts were more akin to a programming language. API doc outlines, code samples, tool routing rules, formatting protocols, javascript snippets with dos/donts.

I used to think prompting is about wording, but these models are now invisible runtime systems behind a chat UI.

[Joseph Perla @jperla]

@jmwind Yes they’re are a runtime / virtual machine https://t.co/m3CVFgnoSo

[Joseph Perla @jperla]

@christophersaum https://t.co/kU2nCyaaGN

[Joseph Perla @jperla]

We’re organizing a call with Alex Bores soon, want to be involved?

https://t.co/jpRAAcfVoC

[Joseph Perla @jperla]

@signulll https://t.co/WBhp0SJF4V working on it

[Joseph Perla @jperla]

Hey @clattner_llvm languages are dead. Come join me

[Joseph Perla @jperla]

@heyblake https://t.co/WBhp0SJF4V

[Joseph Perla @jperla]

Pydantic AI's OpenRouter provider docs get right to the trust issue with routing controls. TrustedRouter's short privacy claim is: "We never log your prompt or the output." https://t.co/jyY1HIdf8a

[Joseph Perla @jperla]

Games https://t.co/75zL0NSxpR

[Noah @NoahKingJr]

What’s coming after AI?

[Joseph Perla @jperla]

@NoahKingJr Games https://t.co/75zL0NSxpR

[Joseph Perla @jperla]

Designing how to get an OpenClaw computer into everyone’s hands

Quill Feather

~$100/each