Seedling — early notes, still formingThe Compute Wall in the Garage
Before Lattice Protocol: a wall of dumpster-dived computers in a Palo Alto garage, quietly folding proteins for disease research. The instinct never changed.
Journal
Research updates, explainers, community reflections, and the occasional dispatch from the rare disease frontier.
Growth stage: Seedling — early notes, still forming still forming Budding — developing, revisited developing, revisited finished essays carry no glyph
Seedling — early notes, still formingThe Hardware-Grant Hackathon
You couldn’t buy the chips yet — you wrote a grant for early access. Then a roomful of builders aimed the same tiny AI processor at wildly different problems.
Budding — developing, revisitedHow a Computer Learns to Read a Molecule
A molecule is a graph — atoms for dots, bonds for lines. Once you see that, a drug-discovery model called the D-MPNN makes sense. We built it into DeepChem.
Budding — developing, revisitedTeaching AI to Speak Protein
If a protein is a sentence, a model like the ones that translate French can learn to read it — and then write it. That’s where new medicine begins.
The Exploding Lego Set
Classical docking searches for where a drug binds; diffusion models imagine it — by running an exploding Lego set backward. On the leap, and its plumbing.
You Don’t Look Sick
“You don’t look sick” is the kindest thing people say to me, and the hardest. It’s true — and it’s also the whole problem, with my disease and with the system built to treat it.
Happy 4th — Thank a Rare Disease Patient
The enzymes digesting your Fourth-of-July cookout are winning rolls in a billion-year genetic dice game. Some people carry the rare rolls — here’s why they belong at the table too.
The Paper Behind Lattice
Before Lattice Protocol could work, someone had to answer an awkward question: how do you trust that a GPU you don’t control actually ran your model? That answer became my first paper.
What If DNA Could Sing?
Not the sound of the waves — the genetic symphony of the thousands of organisms living inside a coral reef. The World Genome Symphony is being built to make biodiversity something you can hear.
Who Owns the Data?
When a teenager sequences the DNA in her school garden, who owns what she finds? Most platforms never really answer. The World Genome Academy is being built to answer it in the architecture itself.
Building a Personal Brand Content Engine with aDNA
I needed a system that could produce on-brand content across six platforms without losing my voice. So I built one — using the same knowledge architecture I use for federated science.
The Invisible Web Under Your Feet
A single gram of soil contains tens of thousands of microbial species. We’re putting the tools to map them in the hands of high school students.
What DeepChem Taught Me About Mentorship
Over multiple Google Summer of Code cycles, I mentored more than ten PhD and postdoc fellows on projects from neural ODEs to molecular property prediction. Here’s what I got wrong first.
Why I Still Don’t Know What Causes My Disease
After years of testing, imaging, and specialist visits, the honest answer is still ‘we don’t know.’ Here’s why that matters — and why it’s not the end of the story.
Church for the Curious
Every Sunday at The KINN in Venice Beach, we hold what I call Church for the Curious. No prerequisites. No credentials required. Just show up with a question.
Proteins Are Not What You Think They Are
You were taught that proteins are building blocks. They’re actually self-folding electric motors, linguistic structures with syntax, and the most sophisticated nanotechnology ever produced.
No entries in this pillar yet.