Arnav

Currently building a consumer ai lab 🍀

Prev: CS @UCSanDiego, 2x AI startups in SF, and deep ML research:

[Arnav @ardotnav]

@10x_er someones gotta write em lyrics right

[Arnav @ardotnav]

this is the bar guys, this is the bar. this is how it should be done. take notes.

[Farza 🇵🇰🇺🇸 @FarzaTV]

Watch me control my computer with just my voice. This is the future of operating systems.

No hands.

GPT-Realtime 2.0 is very, very underrated.

Demo:

[Josef Chen @josefchen]

Launching our new paper on arXiv: we trained the largest multilingual food model ever built.

4.1M recipes. 7 languages. 1,790 ingredients. 300 dimensions.

All of human cooking compressed into 2 megabytes.

[Arnav @ardotnav]

@TrungTPhan

[Arnav @ardotnav]

@shujanshaikh you can download older versions of xcode compatible with older macos versions from apple’s archive

[Arnav @ardotnav]

@himanshustwts yo this is very a valid line of thought

[Arnav @ardotnav]

@paularambles wait this is so iconic

[Arnav @ardotnav]

@anything let’s gooo!

[Chhavi Yadav @chhaviyadav_]

We built Safety-Nudges, a lightweight intervention layer, designed to promote awareness & reflection while using LLM chatbots.

We chat with AI bots every day now. But there’s a growing problem not talked about enough: People are starting to overtrust AI chatbots.

Especially users outside tech.

When a chatbot sounds confident, empathetic, and fluent, it’s easy to forget: it can still be wrong, biased, manipulative, or misleading.

So here comes Safety-Nudges!
https://t.co/psEcus4WF6

[Arnav @ardotnav]

@0interestrates writing is supposed to have an undertone of emotions that come from a lived experience. ai text feels off because you know it is just mimicking emotions.

code is merely logic, as long as it does what it is supposed to, why would you care, rather you can trust an ai more lol

[Arnav @ardotnav]

@hopes_revenge purchased 31 times?!

[Benjamin @bschne]

it's always cool and a little disorienting to watch someone native to the new way of doing something vs. someone whose workflows and tricks were basically calcified by the time they adopted the new tech

[Lenny Rachitsky @lennysan]

Great advice

[andrew pignanelli @ndrewpignanelli]

The new Cofounder site is literally a step-by-step guide on how to start a company.

When i started my first company there was all sorts of stuff i had to learn and all of the guides were SEO maxxed slop. Now there's a real one written by experts and u can download it.

[Goodfire @GoodfireAI]

Neural networks might speak English, but they think in shapes.

Understanding their rich *neural geometry* is key to understanding how they work – and to debugging and controlling them with precision.

Starting today, we’re releasing a series of posts on this research agenda. 🧵

[Alexander Whedon @alex_whedon]

Introducing SubQ - a major breakthrough in LLM intelligence.

It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA),

And the first frontier model with a 12 million token context window which is:

- 52x faster than FlashAttention at 1MM tokens
- Less than 5% the cost of Opus

Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention).

Only a small fraction actually matter.

@subquadratic finds and focuses only on the ones that do.

That's nearly 1,000x less compute and a new way for LLMs to scale.

[Soham Parekh - in/acc @realsohamparekh]

Excited to introduce @PhyseraAI.

Physera is a research and product lab rethinking applied intelligence.

We’re building across three fronts:

1. Making AI systems efficient under real-world cost and latency constraints
2. Simulating human decision-making with high-fidelity, multimodal environments
3. Turning model understanding into measurable, reliable outcomes

AI progress isn’t just about capability, it’s about making that capability useful, predictable, and deployable.

We’re rethinking the stack from first principles.

We’re assembling a small group of thoughtful builders around a mission to create something meaningfully different.
If that sounds interesting, I’d love to chat.

[Arnav @ardotnav]

@nikitabier we had a class in school called anti social computing and for the final there was a project where i built this exact feature for all social media! so cool twitter is getting this almost freaked me out there for a sec

[AnhPhu Nguyen @AnhPhuNguyen1]

with Mira, AI can now live on your face.

capture every conversation. create the most personalized form of AI ever.

order now.

[Kevin Naughton Jr. @KevinNaughtonJr]

people don't realize how insane high frequency trading (hft) firms are

someone i know recently told me they moved to a more sensitive team which increased their garden leave time to ~2 years

for anyone who doesn't know, garden leave is what you're put on when a company (mostly hft firms) can't stop you from going to a competitor but don't want you to take your knowledge of their systems to them immediately either

because of this you're placing on "garden leave" where you're still paid and employed by the company but restricted from coming to the office, accessing their systems, or doing any work. it's like a 2 year sabbatical with the idea being that by the time you join the competitor your knowledge of your previous employer's ip is outdated

if that's not crazy enough here's how the convo went with this person i know:

"they moved me to a more important and sensitive team which increased my garden leave to ~2 years"

"wow that's crazy so when you leave you just get paid normally for 2 years?"

"yeah so i got a nice bump from that"

"wait, got a nice bump? they only pay you when you leave?"

"yeah except they had to pay me to sign the document that increased my garden leave"

"...wait...they had to pay you to get you to sign a document that says they'll pay you even longer if and when you decide to leave the company"

"yeah"

"...how much did they pay you to sign that piece of paper?"

"multiple 6 figues"

imagine getting a quarter million dollars to sign a piece of paper that says you accept a company paying you even more money

these companies aren't real😭💀