Mary Wang

Our latest FRO is a massive telescope! They just launched a pretty sweet website too.

"How does a child learn about the world so much, from so little, so quickly? Why does it take so much more data to train today’s leading AI models? How do you grow a mind?" essentialtechnology.blog/p/the-evolutio…

It’s been a privilege to partner with ARIA. We’ve (@Convergent_FROs) gotten to really hone our approach to looking for and incubating FROs as part of our work in the UK.

Let’s go!

Congrats @EvEBiotech!

Cool work on plastic degrading microbes from @CultivariumFRO blog.cultivarium.org/p/made-from-pl…

Applications for the 2026 cohort of the Brains Research Accelerator are now open! 1/

E11 is lighting up the brain. Amazing work!

Come meet our residency teams! Register: venturecafelondon.org/event/venture-…

London friends—join me on Oct 9 at Venture Cafe London for the launch of Convergent’s UK FRO Founder Residency! I’ll be chatting with ⁦@Pranay_Shahh⁩ ⁦@G_T_Heller⁩ & ⁦@Phil_Budden⁩ about engineering catalytic breakthroughs. Register ⬇️

One of the coolest programs at ARIA. Definitely apply!

Sharing some updates on our ARIA partnership and FRO Founder Residency!

This June, we kicked off our first-ever UK FRO Founder Residency as an Activation Partner for the United Kingdom’s Advanced Research and Invention Agency (@ARIA_research ).

💡 What if the cures for chronic disease are already hidden in our immune systems? Imprint is decoding the immune archive using ML + experimental tools. Our @TEDTalks talk just dropped. Watch here 👉 go.ted.com/beckbrachman @Convergent_FROs @BeckBrachman #Immunology #Biotech  #ML #FutureOfMedicine #HealthTech

@owl_posting I definitely can relate to this at the Rothko Chapel in Houston

i saw a Rothko in person while visiting the Cleveland Museum of Art a few months back, and it was a captivating experience. some people don’t get the hype over Rothko, but they also don’t usually get that the most important thing in art boils down to one thing: how big it is. the literal physical size. this isn’t a metaphor, that’s all there is to it. the average size of a Rothko is 5” x 4.3” feet. the largest one exceeds 9 feet tall. some people will tell you that no, it’s actually the immaculate color theory, or the way that he used egg whites in the paint thinner, or that he somehow infused the whole artwork with his profound grief this is all, of course, nonsense. it ignores that we spent our generational past being wary of creatures the size of buildings and rival tribesman with hands big enough to pop our skulls. and it ignores our present, where we spend our earliest years around entities unfathomably large, the biggest thing we have ever seen, yet so consistently gentle in how they wield that size this is all to say: when i see a Rothko, the sheer breadth of it fills me with terror, like it’s going to swallow me alive. but then nothing happens at all, it just patiently sits there, waiting. so then i feel grateful almost nothing else in life treats us this well, so we feel deeply emotional connected to whatever does

What does it mean to do high impact science in an age of AI advancement? Some thoughts on the future & the next chapter of Convergent’s research priorities.

A cell is a vibrating, densely packed bag of molecules. The ratios of these molecules are not balanced, however. A typical HeLa cell, for example, has about 20 times more protein than DNA by mass. Why, then, are the methods available to study proteins so much more expensive, and less scalable, than those used to study DNA? Human cells make thousands of different proteins using 20 amino acids. Those amino acids can be arranged in a staggering number of combinations to make everything from small, spindly structures like insulin (51 amino acids) to globular giants such as titin (34,350 amino acids). This chemical diversity is what makes proteins so adaptable and infinitely useful. Ideally, scientists would be able to study the proteomes of thousands of cells in one experiment! And cheaply, too! But in reality, existing proteomics technologies can only be used to analyze about a dozen cells at once — a paltry amount. Modern proteomics is majorly bottlenecked, in other words. And it's a shame, because many of the diseases we care about are influenced by proteins. Protein alterations — like phosphorylation and ubiquitination — can change a protein’s function in ways that are invisible at the level of DNA or RNA. Diseases like Alzheimer’s are partly driven by these changes. And it is exceptionally difficult to study the brain, with its billions of cells, by looking at the proteins of just a few cells each day. For my latest essay in @AsimovPress, I explain how a non-profit organization, called Parallel Squared Technology Institute, is trying to solve this problem. Their goal is to make proteomics as scalable and accessible as DNA sequencing. And remarkably, they are working toward this goal NOT by making better hardware (aka bigger, more expensive machines), but rather by inventing better molecular barcodes, devising protocols to squeeze more data out of mass spectrometers, and writing software to handle all the data. So far, they’ve made a 9-plex barcode system (up from the state-of-the-art 3), and combined it with a time-staggered injection method called timePlex, thus enabling 27-plex experiments on existing machines. Their strategy has already boosted the number of proteomes scientists can study in a single day by a great deal. Now they want to 100x this again. The Parallel Squared team is now applying their tools to study Alzheimer’s. They’re mapping the proteomes of individual neurons taken from patients who died at various disease stages, from Braak 0 to Braak 6. By lining up healthy and diseased cells and watching how protein quantities change, they aim to flag molecular markers that appear before symptoms begin — in the disease’s quiet, prodromal phase, when treatments might still work. Thanks for reading. You can check out the full article here: press.asimov.com/articles/prote…

Congrats Eric! So excited for the BBN—oh wait, FRC Launchpad!

NEW: An exciting day at @RenPhil21. Our work at the intersection of AI and math continues to grow, thanks to generous support from Alex Gerko. Our bet: accelerate the utility of AI in discovery by investing in the tools, data, and talent that bridge domain and AI expertise.

📣 We're excited to share the new lean-lang.org! Relaunching our website was a key deliverable in our Year 2 roadmap to provide "improved navigation and access to valuable content, resources, and tools." We hope you like it! #LeanLang #LeanProver