Vitaliy Duvalko

Preclinical evaluation isn’t ready for the AI drug boom. We design drugs faster than we can test them. 90% entering clinical trials still fail, even after promising preclinical results. We’re fixing that: 3D-bioprinted human tissues, on demand, built to de-risk your drug.

@nisargxj @NatalieLeviii @obotiq Very very cool

We built a robot to support caregivers, reduce burnout, and enhance care

@daleverett Wow crazy.

@humfhuang We should meet up

I just tried matcha at Binge Coffee House in Berkeley 8.5/10, def recommend📈 if you also love boba + build cool things, we should connect!

Preclinical evaluation isn’t ready for the AI drug boom. We design drugs faster than we can test them. 90% entering clinical trials still fail, even after promising preclinical results. We’re fixing that: 3D-bioprinted human tissues, on demand, built to de-risk your drug.

@parmita Let’s connect. Are you in the bay?

You should all really think about how bad things had to get for me to burn bridges and publicly come out completely against Anthropic. So read this whole post. I do not care if it is long and it is 10 PM PST. THIS ACTUALLY MATTERS FOR YOUR HEALTH AND YOUR SAFETY. IN TEN YEARS. I will be entering one of the most consequential fundraises of my life and I called the company that Karpathy joined a 'demon cult' on main. Not because I am irrational. Because I GIVE A FUCK. So hear it from me: I care about the world winning. I will NOT shut up. You cannot get me to do that. Because we will NOT let biology DIE. I wrote a 71-page document, sent to major orgs in government RIGHT NOW, and to be released to the public after their review and at the right time. I document the serious concerns no one else seems to understand here. The short of it is THIS: there are TWO versions of the world forming under the new Biological Power. One concentrates power whilst - and this is important - absolutely DESTROYING biological progress, it HURTS AMERICA. The other does not - it leads to biology, and humanity, flourishing in ways most AI engineers cannot even foresee. I will do ANYTHING to prevent the former. I mean this. I will NEVER let the former happen. Anthropic is trying to own the future of AI in medicine while actively limiting who gets to generate the ideas that future actually depends on. They WILL lose if they do not change their ways. Because it is either that, or it is everyone else winning. And everyone else WILL win. They want the models and the credit. They just don’t want the people doing real biological measurement to move without their permission. This will NEVER be safety. EVER. This is the early formation of a control point. I will yap and yap and build and build until that control point is squashed completely, so help me God. Biology doesn’t live in papers or sequences. It lives in cells that change state under stress, adapt, resist, or die in ways no static dataset has ever captured. The next generation of biological models will be trained on living response or they will stay expensive guessing machines. We hold the IP for the provably right technology that is getting there. We will work with everyone else, unless they stop the safety-capture. My company is spearheading a big change. I care who this goes out to. And we have over two years of building that has led to this exact moment. At Precigenetics we’re building the measurement layer that actually sees how living human cells behave over time. While some labs fight over who controls the model, we’re focused on the part that decides whether any of it is true: the living system. Frontier-lab CEOs, investors, and my own investors, are following this conversation I am having. I will do it with no fucks given about what it costs socially. That is all GIBBERISH. Biology is not, and it is not to be weopanized. My company will win one way or another. While others try to gatekeep the biology, our progress will say more than any safety slide ever will. If we scale the way we aim to, we will be part of what decides how AI-driven medicine can move. Everyone else will be training on approximations. We’re not waiting to be approved for Mythos. We matter, and we know it. But above that: biology matters, biologists matter, And the only ones having access to frontier intelligence will NEVER be the ones on Anthropic's payroll. Hell, we build frontier eyes, and we will build frontier brains, if we are forced to do it. I rest my case.

@BrunoMenarim True

@vitalduval If preclinical models were appropriately chosen, studies involved a veterinary pharmacologist we would be a lot more successful

@ahengyuc World changing.

Adam Biotech is officially launching! Learn more by checking out our website: adambiotech.com

@AnselDias @fdotinc @heyato_ai It’s been great seeing the progress here.

My first robot deployment today at the @fdotinc garage with @heyato_ai Wanted to demo finished print removal for continuous production, but instead demoed something a lot more valuable, print failure removal. Continuous production makes more profit, but failure removal saves you loss. I will not say that it failed because it was a Creality (it did), I'll say it would be impossible for a print to fail with @AutoFAB_inc

@humfhuang 🔥

@vitalduval yes

Claude Fable access just got extended, so I used it to one-shot CAD parts and assemblies And I'm not an engineer / don't know how to CAD. Here's how it works: > Describe part/assembly in natural language > Fable writes build123d code > OpenCASCADE generates STEP + STL > Validates CAD model, retries on failure DeriveCad lives in your terminal. No manual sketching. Try today for free: pipx install dcad github.com/HumfDev/derive…

@ahengyuc @Adam_Biotech 10/10

@vitalduval @Adam_Biotech Whoever took the pic did a really, really good job!

Coming soon… @Adam_Biotech

@jainavx @nashpateee Founder moment

@nashpateee Im going to eat glass and stare into the Abyss

are you employed? if yes, what do you do?

@MenarimBruno Yes, definitely reaches past just brain organoids.

@vitalduval Same issue for all organoids, in addition to the lack of immune cells native to the tissue....

Brain computing through organoids has a fundamental problem: Past a certain size, organoids develop a necrotic core where nutrients can’t diffuse into the center. So the cells die. They’ve tried to fix this by adding blood vessels inside the brain organoid. But these are self-assembling - you can’t place them in specific locations. So bioprinting is a much more promising solution - Print brain tissue and vascular networks at the same time. The brain tissue can be far larger, be made much faster, and you have control over the placement of each blood vessel and neuron cluster.

@jarvisnuss Right! The process is much more controllable than growing organoids.

@vitalduval Exactly. Diffusion is the hard governor. The moment the tissue becomes computationally interesting, the center starts dying. Vascularized bioprinting gives the problem coordinates. Nutrient flow, vessel placement, neuron clusters, failure modes you can actually engineer.

@ahengyuc I can attest to the 50h.

I have a lot more respect for jewelers after spending approximately 50 hours the past week electroplating (and not succeeding).

Exactly. I see a vision where: 1) AI creates a new drug 2) AI identifies the key risks of that drug. Maybe it’s an immune reaction or endocrine 3) AI designs the optimal tests to mitigate these key risks. Perhaps it’s organ-on-a-chip with specific cell concentrations generated by AI, or a bunch of bioprints with each one designed to test some key risk