Steven ten Holder
@steventen
building zeroshot biolabs // prev. co-founded @acornbiolabs, YC 2016
What does a transformer model see when it’s given 512,000 zebrafish-embryo-cells-worth of scRNA data?
Chris is awesome and Waves is such a cool idea. As an amateur cinematographer, I know nothing beats being present in a magical moment. Capturing it with Waves is the next-best-thing.
Announcing Waves, a new Camera Wearable Share anything – hands free, invisible, all day We're working with the designers of @GoPro and are backed by @natfriedman, @aigrant, and @southpkcommons. Our all-Waterloo team has been building in stealth with frequent trips to Shenzhen,…
Same effect between great and mediocre entrepreneurs/startup founders.
Something I told 16 yo: In the near term, AI will increase variation in the returns for work. For example, mediocre programmers are now finding it hard even to get hired, but great programmers are making more than ever.
The unspoken open question in theoretical biology right now should be whether biology as a whole will become 'fully modellable' by larger AI systems. Are the patterns of cells reducible to data AI can 'world-build' effectively with? bio-LLM scaling laws soon will tell.
As an adult aiming science toward real outcomes, I think boring/painful execution is necessary to reach new capabilities you can then play with. The trick is to remember to flip into explore-for-fun mode at each new precipice. You can't just play and be effective, tho.
There's nothing wrong with play! As Feynman said & Shannon would contend, you can only hear your genius when you're at play.
Science! Spreading a little science-romanticism here and there ain’t so bad. 🔭 Thanks @cygnetcommitt @NeilMitra_ and @raenytuesday.
We went deep on why and how we do science @v2_vancouver last night. Thanks for bringing so much curiosity and engagement! So fun to collab with @steventen on the talk and transmit some Feynman (🐐) wisdom. @raenytuesday brought it all together, and big hbd to @NeilMitra_ !!
I'm genuinely starting to feel like: 1. I tell my AI that I'd like to achieve complex thing, X. 2. It understands better than I do how to achieve X. 3. I let it tell me what to do to achieve X. On the one hand, I'm happy that we can achieve X. On the other hand..
I’d LOVE this, but for 2001: a Space Odyssey. I currently wait ~2 years between viewings for the same reason:
This will sound crazy, but it's a serious question. Could a hypnotist make me forget the plots of all the Sherlock Holmes stories? I just reread them all, and don't want to wait another 20 years before I can reread them again. I would pay to speed this up.
if not punishing, definitely not celebrating ambition. artists largely inform the culture, and I believe there aren’t nearly enough ambition-supporting works being made like there used to.
The new episode is out! 🎙️ In this episode, @steventen of @AcornBiolabs explores the possibilities beyond human experience, discussing how Neuralink and advancements in AI could change the way we think about our existence. buff.ly/41wRP4f #FutureOfHumanity
This wont age well 🤭
Stargate's cancer mRNA vaccine is impossible to achieve, and AI research on mRNA vaccines is almost useless.
Agreed. I bought @Ginkgo stock a couple years ago. Has performed terribly. I’m holding.
Prediction: Gene editing will be huge. This is just Internet stocks in 2000.
energizing, optimistic, high-agency Canadian spirit:
the 2025 World's Fair will be held in a waterloo hockey arena excited to share a bts look at what folks have been cooking !! see you at symposium :)
So beautifully put 🫶🏼
Why desire the radical extension of life? Because everything valuable requires time. Time to write, to read, to create and refine work that leads us to deep meaning. Time to learn multiple languages, to peel back the endless layers of our identity, to reinvent the best version…
Ask o1 pro about this and it seems at direct ods with the fundamentals of what makes an LLM powerful. HOWEVER ‘anchoring’ classical compute ‘truisms’ throughout an otherwise creative LLM core would be amazing for bio
i think i’m far better at playing the piano after a long warm bath