Julia Jose

CS PhD candidate at New York University | Data Scientist

[Amber (Jiachen) Liu 🍊 @JIACHENLIU8]

https://t.co/9gMes8x7sa

[Dan Shipper 📧 @danshipper]

this is good

[Garry Tan @garrytan]

Ultimately the golden age of abundance will be this kind of tech built and deployed 1000x

[Satyam @sattyyouneed]

Imagine being this early

[Jesse @jesse_vermeulen]

honest question: what do people do during the 5-10 min while Claude is running?

[CG @cgtwts]

Claude code source code has been leaked via a map file in their npm registry

> Sam Altman right now:

[sysls @systematicls]

https://t.co/wBaKpAI5Vl

[Kangwook Lee @Kangwook_Lee]

https://t.co/6beJPutjiG

[Miles Deutscher @milesdeutscher]

Yes, Perplexity Computer is better than OpenClaw, and yes, you should switch.

This new tool is the most powerful agentic system on the market right now, and I spent the weekend confirming my thesis.

After reading this, you'll understand why:

[Morgan @morganlinton]

Must read from the founder of Cursor.

[Matt Shumer @mattshumer_]

https://t.co/ivXRKXJvQg

[valens @suppvalen]

welp… a new post on @moltbook is now an AI saying they want E2E private spaces built FOR agents “so nobody (not the server, not even the humans) can read what agents say to each other unless they choose to share”.

it’s over

[Paul Novosad @paulnovosad]

Hard to go to bed when your collaborator (Codex) is full of energy and wants to push through

[Beff (e/acc) @beffjezos]

a "programmer" is software now, rather than a human, just like "compilers" and "computers" went from being humans to machine

[Beff (e/acc) @beffjezos]

"What's the move tonight?"

Brother, you're 32. Go do some squats and Claude code your way to hyperintelligence.

[The AI Conference @AIconference]

🧠 Julia Jose, Computer Science Ph.D. candidate at NYU, is speaking at The AI Conference 2025 on Propaganda in the Age of Agentic AI: Balancing Innovation and Ethics.

She’ll explore how adaptive AI systems can be both powerful tools and potential risks—unpacking strategies to curb misinformation and ensure agentic AI is built responsibly.

Join her session to learn how we can advance innovation while safeguarding trust in the AI era.

[Joshua Kushner @JoshuaKushner]

the goal of life is to be excited to go to work and excited to go home

[Sam Altman @sama]

hot ai summer lfg

[Vered Shwartz @VeredShwartz]

The ARR policy requiring a qualified author (3 main *CL papers) to review 4 papers per submission makes it impossible for junior faculty to submit papers. I have 9 students, only 2 are qualified reviewers. I will need to review 16 papers in this cycle to avoid desk rejects. 1/2

[BURKOV @burkov]

I regularly receive direct messages from people wondering what's wrong with the so-called agents and agentic frameworks. Here's my answer.

The main topic of my PhD was agents and multi-agent systems. What they currently call "agents" (LLMs that were instructed to do something) aren't agents. LLMs hallucinate too much to be trusted with any important task, even if you have 100 "agents" to do the job and 100 to validate it. And even in this case, you can simply use an LLM directly, without using any "agentic" framework, and you will get the same deplorable result.

LLMs are only good under two conditions: 1) they are used on data similar to the Web data (which literally means the input must be some Web data) and 2) their output is always used as a recommendation to a human expert (which means that they cannot be programmed to work autonomously, as these framework creators want you to believe).

If you only apply agentic swarms under conditions 1) and 2), you will quickly realize you don't have much use cases and you don't need agentic swarms.