I've had many engineers ask me why its worth their time and effort to learn biology in response to this post. Why should they be excited? We are poised for a revolution in biotech that will be uniquely enabled by computers. Convince yourself by digging into the examples I link:

These products will behave far more like a collective, shared memory of scientific details *usable* in context than other things you are used to

📊 Private equity continues to deliver some of the strongest long-term returns for retirees. New AIC research shows 13.5% median annualized returns over 10 years—benefiting 34M teachers, firefighters & public servants. See the full data.

An intuition about agents in comp bio you should really internalize: a system need only see a specific problem - task in context of disease, tissue - a single time before it can be deployed + replicated to tens of thousands of scientists

one of the primary reasons i stopped working in bioinformatics was because i was frustrated by how bad, painful, and straight up broken the tooling for biology was. tools were unintuitive, documentation was non-existent - it always felt like trying to dig through a mountain with

Today, we’re excited to introduce Polyphron ( https://t.co/yFz5TgYN8N), a compute-first bio company building the world’s first personalized tissue foundry. Medicine today can influence biology, but it still cannot rebuild the tissues that fail. We aim to change that. Polyphron

Mind-blowing scoop this morning on Polyphron. They're building human tissue - ie parts of organs - to replace diseased and damaged parts of the body. @eryney_ok delivered the goods on this one https://t.co/JD4wOBEUeO

Learned a lot about biology specific agent engineering in the past month building and saturating our benchmarks (eg. what is the analogue of a “code diff” for spatial analysis). Working on more educational material on these topics in the coming weeks https://t.co/1YTzrJnaye

Released concrete spatial flows across diverse tissues, diseases + kit technologies. Ex: > count antral or atretic follicles across ovulation timepoints in mice > detect spatial niches in Alzheimers like GFAP+ reactive astrocytes and HEXB+ activated microglia > recover cortical

Benchmarks across diverse tissues + diseases. Intermediate artifacts (plots, statistics) surfacing trusted controls to prevent errors from compounding. Transparent decision making. - https://t.co/lB6KeSpK5z - https://t.co/ezA1ILSGYw -

Reliable agents in science require focus. Modern assay data often the distillation of weeks of work and tens of thousands of dollars in reagents, labor, machines. Tools that steer raw data to scientific decisions must be reliable. Eg. code diffs allow engineers to inspect and

When your business spans factories, languages, currencies, and continents, visibility is everything. Priority ERP gave Arkal a single, unified platform to oversee performance across all sites, consolidate financials, and maintain full control over every operation.

The path ahead is a long tail of soluble work: continued progress on benchmarks, purpose built infrastructure and technology specific heuristics encoded into prompts/evals. We’ve learned a lot about "biology specific" agent engineering in the past month building and saturating

Our focus is on Takara Seeker, Vizgen MERFISH, AtlasXOmics DBiT-seq, 10X Xenium and 10X Visium. Few notes: - You can try to use your own data, and the system will likely generalize accurately on some tasks, but we provide no guarantees on performance quite yet. - The sandboxes

Launching a public agent sandbox for spatial biology. Five demo flows tailored to specific kits/machines Try it now: https://t.co/6lCf97Q5xi This is a shippable intermediary towards reliable and widely deployed agentic systems used to make expensive scientific decisions.

There's unreasonable wisdom in "be super technical and follow your curiosity to build things of value for the world" as a guiding framework for a meaningful life. Institutions and ideologies rise and fall, but creation in the pursuit of understanding Nature will always be the

We don't know what most microbial genes do. Can genomic language models help? there's only one way to find out! this is a 1 hour and 42 minute interview with an MIT professor (the famous @Micro_Yunha) chatting about these questions, her work in solving them at @tatta_bio, and

Spatial biology agent sandboxes scoped to public datasets for Takara Seeker (3x3 + 10x10 mm), 10X Visium + Xenium, Vizgen MERSCOPE and AtlasXOmics DBiT-seq launching this week. The real work began after launching this POC 5 weeks ago - decomposing practical analysis into

A new way to watch UFC

Simple result in Nature shows flow doesn't just destroy spatial context but tosses out important biology. Activated T cells are usually physically stuck to tumor cells and gating pipelines toss them out. Up to 91% of T-cell/tumor clusters gone. A huge fraction of clinical IO

If you think about it, machine guided data analysis, especially in biology, likely the next frontier after agentic SWE. Verifiable tools that help scientists with strong existing understanding of the domain do work with higher quality + speed (raise the ceiling not the floor)

Working on a video series digging into the internals of spatial biology datasets and specific ways they strain hardware + browsers. Intended audience is bio curious systems people but anyone interested in computing x bio (scientist comfortable looking at code) welcome

If you work regularly with molecular data, but don't prep it yourself, highly recommend reading the user manual for a kit you interact with cover to cover. Every little step is there for a reason + will teach you interesting biology. Small deviations at the bench often perturb

happy thanksgiving you filthy animals