“Three profoundly destabilizing scientific ideas ricochet through the twentieth century, trisecting it into three unequal parts: the atom, the byte, the gene.”
― Siddhartha Mukherjee, The Gene: An Intimate History
automated wet labs are more like semiconductor foundries than neoclouds. TSMC maintains their value by doubling what they can do every two years, making them incredibly difficult to catch.
for automated wet labs to have foundry-like economics, they’ll also need to be “hard to catch.” this might be possible through assay reproducibility, lower error rates, and integration of instrumentation at breakneck speed.
exactly. it is probably a mixed bag though. there will be areas of biology where additional data combined with AI will make rapid advancements. there will be other areas where this approach reaches a local maximum and progress stalls, after which new tools will need to be developed to push things forward. new tools or “metrology” will be a governor.
🧵 Biohub, @ENERGY, @NIH, and new funding partners today announced a major expansion of an international effort to generate and make accessible the data enabling predictive AI models of biology. Together, the organizations are investing $1.8 billion in funding, data, computation and new measurement technology - largest coordinated commitment to map the biology of the cell to power development of digital models of biology. https://t.co/PaCs7ZV2Fg
Many drugs work by binding to a specific target in the body and blocking or changing what it does. An important first step in the drug development process is designing a molecule that can bind tightly to its target. Traditionally, that's meant weeks or months of expert work per target, sifting through a large number of candidates to identify the few that work.
We wanted to test if Claude could successfully design novel protein binders from scratch (also called de novo design). With a protein design prompt written by a human expert, Claude autonomously designed protein binders against 14 out of 15 targets.
We then worked with Adaptyv Bio and Twist Bioscience, who independently built and tested the proteins Claude designed.
yup. there will eventually be some real good capital allocators from outside of traditional biotech, Bezos / Elon’s of biotech, and who will be able to successfully break the culture and rebuild it.
automating more is necessary but not sufficient, autonomous labs are an abstraction layer - China will have them, and the most automated companies in the world (TMSC, Tesla, Amazon) still have the most brutal work cultures, it’s what enabled them to be the most automated, not the other way around.
didn’t think that anyone would use Amazon Bio Discovery – but it looks like there might something useful there as part of a larger cooperation between a pharma like Novo Nordisk and Amazon. since we are talking Novo i expect to see some followers signing up eventually.
Our strategic partnership with Novo Nordisk expands today as they announce that they've chosen AWS as their preferred cloud provider and strategic AI partner to help accelerate drug discovery.
We will combine Novo Nordisk’s scientific expertise with AWS’s AI tools, including Amazon Bio Discovery (biology AI models to design and test drug candidates) and Bedrock AgentCore (infrastructure for building AI agents that run real workflows), and also stand up a co-innovation hub in London where their scientists and our engineers work side by side.
We’re proud of our broad partnership across the two companies, which already spans AWS, Amazon Pharmacy, and One Medical.
I’m optimistic about how we’ll use AI together to speed up drug discovery, and what that will mean for patients. https://t.co/CLq0bMfvs3
to scale a robotics business, build out a competitive reagents catalog, and do even a fraction of the dozen or so other things @gingko was supposed to do, you need much more than some lab automation that runs 24/7.
you have compared @ginkgo to companies that have built amazing businesses but also have notoriously brutal work cultures eg. Fabs (TSMC), AWS (Amazon), Starlink (SpaceX / Elon inc.).
@ginkgo has always had lab automation yet you have struggled to scale business that you should have been able to do a tour-de-force in. maybe it isn't just interest rates, or vc’s, or china policy.
To figure out where to go, bio needs a few hard-core capitalists. That’s what propelled the computer and internet industries into what they are today. Bio doesn’t need another article analogizing itself to Waymo, AWS, Starlink, semiconductor foundries, or fucking Michelin tires. Bezos, Musk, Jobs, and Morris Chang didn’t need a Media Lab to figure out what products to build.
I’m convinced that culturally bio is incapable of producing the kind of companies needed to build the product layer, or even the “foundry” (in your words). My bet is that over the next decade, outsiders will get interested in bio and be able to break the culture, and build some massive companies.
I’ve always believed the No.1 application of AI should be to improve human health.
That work started with AlphaFold, and now at @IsomorphicLabs with the mission to reimagine drug discovery and one day solve all disease!
We are turbocharging that goal with $2.1B in new funding.