
Building the data layer for biology @strandaibio | YC W26
San Francisco, CA
Joined June 2012
- Tweets2.3K
- Following1.5K
- Followers559
- Likes39.7K
Pinned Tweet
We're releasing predicted whole slide proteomic images for 442 H&E slides from the TCGA colon adenocarcinoma cohort (TCGA-COAD). A public resource for research, generated with Lattice, our frontier virtual spatial proteomics model.
Explore the original tissue alongside whole slide predictions, from regional context down to cellular detail. Use the dataset and SDK in your own research.
We’ve discovered that these predictions carry interpretable prognostic signal. Stay tuned to learn more.
We plan to run Lattice on all of TCGA. What other datasets would you like to see?
עודד | oded retweeted
Flashback to June 14, 2024.
At the time, frontier AI models have zero fluid intelligence. They score around 5% on ARC 1. They can't reliably add two numbers. They make laughable logic errors every other prompt. Reasoning models don't exist. Coding agents don't exist. Claude Code doesn't exist. "Code gen" is a complete joke. Scaling up next-token prediction based on web text data is hitting a wall.
What if a time traveler told you that, by November 2026...
• AI models would have human-like fluid intelligence, thanks to a new paradigm: test-time induction (modern reasoning models).
• Not only would ARC 1 be saturated, but unreleased benchmarks 10x harder (ARC 2) and 100x harder (ARC 3) would be saturated too.
• Nearly all code would be written by AI, and models could one-shot complex codebases flawlessly.
• AI models would be solving multiple Millennium Prize problems.
• OpenAI's valuation would be 15x higher, and Anthropic's would be 85x higher.
And on top of all that...
• Nvidia would still hold ~80% of the AI compute market.
• Nvidia would have grown its revenue and earnings 4.5x.
• Nvidia would still be growing 75% a year.
• Nvidia would have even higher operating margins (67% vs. 65% back then).
Based on that information, straight from the future, would you have bought NVDA?
If you answered yes -- congrats. Today, 2.5 years later, you'd be down 5% on your investment (even more after inflation).
Scientists and builders: re:AGENT - The Virtual Biotech is happening January 9-10 in San Francisco, ahead of JPM Healthcare Week.
It’s a two day build weekend for AI tools across drug discovery, clinical development, and regulatory approval. No AI experience required, and solo applicants will be matched with a team.
We’re sponsoring through Strand AI and will be there in person. Would love to see y'all there!
Applications close December 11: luma.com/3qpuutkz
1
57
there's something impalpable about this new world where everything you could dream of is at your fingertips. creations can be summoned upon a whim
but the systems put in place before this new world impede us. and they're deeply seated, hard to shake off. the world can't catch up
maybe tomorrow, we can go a bit further?
1
1
70
you can dream up a city on the train ride home and walk its streets under two moons before you reach your stop.
i wonder how many of its roads would still bend around walls we never put there
maybe tomorrow we go a little further.
1
19
you should be hacking yourself
github.com/BaaRai/ARTEX
The #FBI issued a FLASH advisory warning that cybercriminals and state-aligned hackers are exploiting ARTEX, a super intelligence-powered penetration testing framework, to automate attacks against businesses and organizations worldwide. ARTEX can accelerate intrusions from initial targeting to privileged access in under three hours. It was used in the October ConnectWise supply-chain compromise, which exposed Microsoft 365 customer environments and involved attempted authentication against more than 2,250 downstream tenants.
The FBI is urging businesses and organizations to strengthen access controls, monitor suspicious automated activity, and report potential compromises to ic3.gov.
Read more here: ic3.gov/CSA/2026/261113.pdf
83
a huge step forward to making medicine a verifiable domain
Today, we’re introducing the Nolla Visual Medicine Benchmark. We’ve been using it internally for the past year to evaluate models for patient care.
Over that year, we’ve seen frontier models make rapid progress. Now we’re sharing the results comparing frontier model performance with human performance.
1
154
So bullish on these two. They're building the next generation of drug delivery
Today we are launching @BioDelivera a programmable drug delivery platform that delivers therapies only to diseased cells enabling cures for every disease. 🎯🧪
Less than 3% of an injected therapy reaches diseased cells. The rest ends up in healthy organs. At @BioDelivera we want to change that, our mission is to make every therapy accessible to patients.
Happy to build with @ikeda_trashi this exciting technology!
Reach out at [email protected]
2
5
363
עודד | oded retweeted
futurebio.xyz is back in SF bringing you the second edition of re:AGENT w/ @gxl_ai + @AnthropicAI
48 hours to build the infrastructure for fully agentic drug discovery, development and commercialisation
imo one of the most talent-dense rooms in the Bay
link below
As your surface area grows (by way of agents), so does the attack surface!
Any and every package you use is an attack vector! Be wary!
North Korea was committing to our GitHub under my name for months without us noticing.
Embarrassing to say, but it's worth a PSA for anyone building with AI right now.
One of my engineers was moving fast on a build. His AI agent found a project on the internet that helped with the task.
It either picked it on its own or recommended it, and nobody really cared. We just went with it. He built it on his Mac.
That one build tampered with his GitHub Desktop app and stole the token that lets anyone act on our behalf.
(GitHub is the Google Drive for developers. It's where almost every piece of software you use lives.)
From there they re-created our real commits. Same names, same dates, one thing added: a line of malware at the end of a file nobody ever looks in, hidden behind a trail of blank space.
You could open the file, look right at it, and see nothing.
We cleaned it up in September. It came back in October. Then November 2. Then November 6: 153 force-pushes in 23 minutes.
We'd swap our tokens and it still came back. We scanned the Mac and it said clean. It wasn't. The scanner tripped on one small detail: the space in GitHub Desktop app filename. Once we found it, we were good. It wasn't easy at all.
Luckily our systems held. The biggest wave never reached production. Not a client's app, not ours. Main branches block force-pushes, production only ships from them, every push gets scanned, real secrets only live in production. We built that for a day like this.
But it was a close one. It's my fault. I didn't put enough safeguards around my engineers and what their agents were allowed to do. Moving fast vs moving properly. It's the risk you take. You need good systems underneath it.
This matches public reporting on PolinRider, which researchers tie to North Korea. Hundreds of repos hit. We're just embarrassed enough to write it up.
Vibe coding or letting agents run things you don't fully understand? Run routine checks. Talk to someone who builds systems that hold up, because attackers are using the same tools we are.
Know what you're running. Especially when you didn't pick it.
The full timeline, a 30 second check, and a prompt you can paste into your AI to audit yourself: github.com/anurieli/polinrid…
PS. Yes, I just warned you about random repos and sent you to a random repo. Read it before you run it. That's the whole point.
2
5
166
We're sponsoring re:AGENT II: The Virtual Biotech! It's a two-day hackathon weekend on January 9-10 in SF, the weekend before JPM Week.
Apply here: luma.com/3qpuutkz
OpenAI needs @MomenticAI
1
119
i'm watching the match and then i see this... @moremelessyue cmon was this really the best use of funds?
3
171















