Matthew M.Y. Mo

AI & MLE in TechScience TechHealthBio, 1X FE
Curious biosciences builder
Stay hungry (16-8) & executing

[Matthew M.Y. Mo @mattmo_668]

I'm representing a VC looking to fill the role Forward Hiring Founder.

AI is moving so fast that as a founder if you're still hiring for people to fit a square-boxed role you're stepping yourself in the foot.

The founder needs a keen sense on finding the people who're perpetually curious, frequently engaged in seemingly unconnected activities that have nothing to do with the role or hyper multi-tasking. E.g. Skiing during meetings are highly encouraged.

The founder should lead with humor as the principle culture. It is expected of him or her to have a cracked sense of humor, because no one wants to work in an environment with stale humor. With the given leadership, it'll nurture a team that is not only very creative, perceptive and would never need a joke explained.

That is the hallmark of the companies they would like to see built

[george church @geochurch]

https://t.co/3BWIq8mv2F
A Framework for Autonomous AI-Driven Drug Discovery
Douglas W. Selinger, Timothy R. Wall, Eleni Stylianou, Ehab M. Khalil, Jedidiah Gaetz, Oren Levy
@plexresearch

[Matthew M.Y. Mo @mattmo_668]

Revolutionary and domain crossing drug discovery tools are exploding everyday, almost feels like one can run a drug-discovery lab sending data to CloudLab from a garage in about 5 years

[antisense. @razoralign]

TransXplorer: An automated translational discovery platform for RNA-seq data https://t.co/syz94jg6Nd

[Le Cong@Stanford, AI+Bio+Gene-Editing @lecong]

Can we program cells like computers — using RNA?

Two years ago, our group trained the first language model to decode the regulatory grammar of 5′ UTRs in mRNA, published in Nature Machine Intelligence.

Today, we’re excited to share the next step, also in Nature Machine Intelligence:
“Programmable RNA translation through deep learning-driven IRES discovery and de novo generation.”
We built an AI engine to discover, predict, optimize, and generate IRES elements — RNA control modules that regulate translation initiation.

This brings us closer to programmable RNA systems that control when, where, and how strongly proteins are produced inside cells.

AI is no longer just helping us read biology.
It is beginning to help us write it and harness it.

The future of computing may not only run on silicon — it may also run inside living cells.
#AIForBiology #LLM #AI4S #AI #RNA #MachineLearning #Bioengineering

[Matthew M.Y. Mo @mattmo_668]

When Prof. Katerina Fragkiadaki (best RL professor one could hope for) taught about GODA RL, It is clear that it's a match in heaven with Biology.
Functional semantics. Lab tech (like FACS) are always constraint by temporal sparsity, multiplexity. Impactful technique!
#ML #RL #Laboratories #Bio

[Seohong Park @seohong_park]

What if we represent a state as a "list" of similarities to all other states? In our recent ICLR paper, we studied this "dual" representation.

Come visit our poster at #4608 10:30a-1p on Fri (morning, 2nd day)!

Paper: https://t.co/zYKFjyO0i4
Blog post: https://t.co/lw1Port5k6

[Matthew M.Y. Mo @mattmo_668]

From gene cards to cell cards. The days of collecting all the Pokemon Cards in Bio is coming 🧫 !

[Katy Beckermann @katy_beckermann]

Single-cell survival analysis is here, and it's smarter than pseudobulk.

scSurvival (Cancer Discovery 2026) uses AI to score risk cell by cell:
🔴 SPP1+ macrophages = high hazard
🟢 TCF7+ stem-like T cells = protective
📊 C-index 0.812 in melanoma ICB cohort, open source

The next biomarker isn't a gene. It's a cell state.
Congrats to @ZhengXiaCompBio on this work!
https://t.co/6QTYfovC3h
#scRNAseq #oncology #AIoncology

[Matthew M.Y. Mo @mattmo_668]

At CMU, I got my worst grade ever by a class in "Quantitative Biology". Fooled by the class name but should've known better from Dept. of Biology.
Lots of vocabularies for each MOA-phenotype set-up
Real lesson: Don't bring a knife to a gun fight
#funny #biology #education

[Matthew M.Y. Mo @mattmo_668]

@RapaNews @DannyArends @Nature Soma loci don't meet the exact definition of antigonistic pleitropy where it happens on the same allele though IMO
While the effects of the disposable soma demonstrated an antagonistic trade-off through allocation, it doesn't share the same mechanism as a classic William AP?

[Matthew M.Y. Mo @mattmo_668]

I get a tier 2 plan on Kimi along with 2 Claude Max sub because of scientific reasoning.
A longitudinal AGING data paper to multi-agent experiment planning. Right directions:
1. Intent
2. data correction
3. Subdomain X properties
4. Interventions
#Kimi #AI4Science #aging #bio

[Danny Arends @DannyArends]

Our paper in @Nature today 🥳 We tracked 6,438 mice from puberty to death and mapped the genetics of *when* you die, not just whether a gene associates with lifespan.

https://t.co/EoeexqJoHk

59 loci. Two decades of data. Thread 👇

#Longevity #Aging #Genetics #Healthspan

[Matthew M.Y. Mo @mattmo_668]

It's always been a dream to keep intensely building yet stay curious w/ research. Now I can test an idea over a weekend.
Check out contextual multi-agent active reinforcement learning a perturb-seq bio-analysis! you CAN do LLM "uncertainty sampling"
https://t.co/OQnO7mI14W

[Matthew M.Y. Mo @mattmo_668]

The way I work now is before I sleep I hand off a vision prompt and let multi-agent work on it for hours to complete the development lifecycle including the audit.
It forces me to spend a lot more time on the first step with the planner to clarifying all reqs. Win-win. Do & learn

[Matthew M.Y. Mo @mattmo_668]

@google https://t.co/fYyQyC0QWM
This opens an exciting arena of doing active learning by directly using LLM as acquisition function. Do we need to stress test explainability to sufficient degree that meets scientific rigor?

[Matthew M.Y. Mo @mattmo_668]

I suppose I'm not the only buying multiple Claude subscriptions #Claude #Outage

[Matthew M.Y. Mo @mattmo_668]

Remeber Einstein once said "Intuition is the highest form of intelligence" and it cannot be more true in the age of AI.
https://t.co/2CS1DLEPKq
It would seem even proofing theorems benefit from an AI system that verify intuition.
Explainability-first is how I can learn with AI

[Matthew M.Y. Mo @mattmo_668]

@lupantech @james_y_zou Can you give a brief overview on why multi-agency make for a particular functional appeal in microbiome r&d? As opposed to other form of biology research

[Zaixi Zhang @ZaixiZhang]

https://t.co/fe5Alzz5I7 is a platform, a library, a dataset, and an ecosystem on Claw for Scientific Research

technical report and survey: https://t.co/tRY5JGfn6y