Pinned Tweet
excited to announce that BLENDERGPT has raised $0.001M led by my mom to build the best text to 3D on the planet
we are so back chat
aarya retweeted
On August 31, the first two participants were injected with @PopVaxIndia’s first fridge-stable RNA vaccine candidate in a Phase I clinical trial in Australia. We built this vaccine at our RNA Foundry in Hyderabad – from computational design to GMP manufacturing. Here's how:
aarya retweeted
My friend, co-conspirator, and colleague @jajoosam has written a wonderful essay about the collaborative baking of half-baked ideas. I highly recommend it, and not just because he says nice things about me:
jajoo.sh/medium-rare
aarya retweeted
Wrote something with @theachyuttiwari on where India stands on AI, what we can do, and why we need Homi-style independent thinking and action.
swarajyamag.com/commentary/w…
Happy Independence Day folks 🇮🇳
i’ll be visiting cape town next week! hit me up if you’re around and interested in shenanigans and tomfoolery
aakarsh is incredibly talented and super fun to work with!
hi! i’m a recent Berkeley grad looking for roles across evals, ops, or early-stage biz dev roles.
most recently i was an AI engineer at the Center for AI Safety. before that i built BASIS at Berkeley, now one of the largest student AI safety communities (secured a $257k grant from CG).
especially interested in AI for security, science, on frontier research problems.
would love intros, pointers, or hear about any companies i should talk to.
thanks!
aarya retweeted
with Mira, AI can now live on your face.
capture every conversation. create the most personalized form of AI ever.
order now.
aarya retweeted
If you write firmware, you know: test plans depend on whoever wrote them, edge cases slip through, and bugs show up after you've shipped when they're 10x more expensive to fix.
I built a tool that understands your architecture and generates a comprehensive test plan your team can actually execute. Procedures, expected values, failure diagnostics, all traced back to the spec. Export-ready for the compliance reports, traceability matrices, and sign-off docs your managers and clients are waiting on.
Here’s a glimpse of it in action:
orgo is positively the best infrastructure i’ve used for building computer use agents (<100ms latency for launching cloud desktops!) and spencer’s been working extremely hard to deliver this. incredible product and founder—give it a shot!
Introducing @Orgo
Orgo provides virtual computers for AI agents.
Build computer-using agents and deploy them across thousands of instances in seconds.
Now available for developers.
aarya retweeted
What if bigger models are actually more interpretable?
Mixture of Experts (MoE) have become central to scaling LLMs and are used in nearly every frontier model.
Yet we still lack a mechanistic understanding of how MoEs represent features differently than dense models.
How do experts affect superposition (how networks represent more features than there are neurons) and how do experts specialize?
In our recent NeuIPS workshop paper, we found that MoEs are more interpretable than dense models.
work done w/ @foundinrome @JeremiNuer at Arcadia Research Team
aarya retweeted
introducing Halo X, always listening AI glasses.
vibe think. never use your brain again.
pre-order now.
aarya retweeted
🚨 We have $1M and 90 days to build something the world can’t ignore.
*We’re hiring for the summer*
This summer, we’re launching The Arena: a 10-week sprint in SF where 10 people, 1 house, and $1M come together to launch a new version of our product every 7 days.
Over a summer, Caine and I lived in a Harvard makerspace, built flamethrowers, iron man helmets, and slept on workbenches. Our projects have hit 73M+ views and got covered by The New York Times, Forbes, and TechCrunch.
Now we’re back, and the stakes are higher. Our only KPI: get Halo to 100,000 users by August.
If you're a hacker, designer, or builder who:
- ships fast
- is creative
- loves futuristic tech that improves the world
this is your chance to build something legendary with us!
super well written essay -- so much valuable science gets trapped in private lab notebooks and never gets published. there's a really interesting opportunity to build the "github for research" here and massively accelerate science.
Documenting and sharing research in real-time is underrated in discussions about open science. @jainhiya_ and I think software can help change problem selection, collaboration, and funding.
We write about how and why we should create real-time, open lab notebooks.
aarya retweeted
today, i'm excited to release a reinforcement learning guide that carefully explains the intuition and implementation details behind every single fundamental algorithm in the field. enjoy :)
naklecha.com/reinforcement-l…
