Sabitlenmiş Tweet
Welcome to the scientific revolution. 100s of robots. Zero coffee breaks. America’s largest autonomous lab, open today.
English
Michelle Lee
1.7K posts
Michelle Lee
@michellearning
Founder & CEO of @medra_ai. Let's eradicate disease. Prev. phd @StanfordAILab, robotics @nvidia, propulsion @SpaceX. Loosely hinged.
San Francisco, CA Katılım Ocak 2018
1.2K Takip Edilen9.1K Takipçiler
@michellearning And…. Analysis, data mining, storage and triaging. More data = more problems. Merge transcript with some Proteomics’s….
English
High throughput is not the bottleneck. Nicer UI and 5% more optimal dynamic scheduling is not the bottleneck. Natural language interface is not the bottleneck.
Running more complex, informative, decision-relevant experiments and generating net new data at scale is.
Ron Alfa@Ronalfa
big pharma has been running cheap scalable high throughput experiments with robots for decades, more cell titer glow was never the problem
English
@JeremyBlitzer @Ronalfa 100% agree. Clinical trials are the most informative, complex, and decision-relevant experiment we can run!
English
@Ronalfa @michellearning I really believe the greatest bottleneck is in clinical development: right patient to right drug.
English
@SRsrivatsan @jrkelly What would it take for a solution to be out of the box?
English
I'm not opposed to the 96-well plate format — there aren't tools more versatile than plate readers, for example. My point is that your tools, chained with arms like Medra's, are enabling (like pipettes) rather than an out-of-the-box solution. We still need to get these into the hands of many kinds of biologists to develop the assays at scale.
Wrt high value experiments for AI models of bio, I'm bullish on scaled single-celled organism experiments (e.g., bacteria, archaea, protists). There's a ton of alpha here for diverse training data, and it's been largely ignored because it's outside the purview of traditional biomedical science.
English
Michelle Lee retweetledi
The experiments we need for a general bio model are diverse and most are incompatible with a 96 well format.
RL-scale data for bio is going to come from the design of new assays that combine automation and multiplexing to generate multimodal data.
Samuel Spitz@samuel_spitz
Anthropic / OAI should acquire Ginkgo Biosciences Their automated labs unlock verification/RL for bio. Why hasn’t this happened already?
English
Run autonomous experiments because the data is valuable not because it’s easily automatable.
Sanjay Srivatsan@SRsrivatsan
The experiments we need for a general bio model are diverse and most are incompatible with a 96 well format. RL-scale data for bio is going to come from the design of new assays that combine automation and multiplexing to generate multimodal data.
English
@jrkelly @SRsrivatsan Nothing is “wrong” with SBS format. But science is more diverse than one standard format. I’m sure Ginkgo’s scientists are also using flasks, microcentrifuge tubes, conical tubes, in their day to day science work
English
Yeah of course, multiplexing always a good idea — but some standard container format to move atoms across various lab devices seems like a good idea. What’s wrong with SBS format? Many analytical devices, centrifuges, thermocyclers, etc etc, are already compatible with it.
Or you have another format you like ? Or you think we need a bunch of different ones ?
English
Michelle Lee retweetledi
love this. im funding:
- Scale for mouse models
- Palantir for long-term biobanks
- Vercel for cell-free expression
- Figma for synthetic promoters
- Cursor for peptide design
Mgoes (bio/acc 🤖💉)@m_goes_distance
biotech ideas we're funding rn: - Stripe for gene therapies. - GitHub for wet labs. - Cloudflare for clinical trials. - Plaid for biomarkers. - Bloomberg for biological age. - Costco for GMP manufacturing. - Anduril for biosecurity. - NVIDIA for programmable biology. - Cursor for molecular biology. - A "WHOOP for cells," not steps. If infrastructure precedes the science, biotech accelerates 10x faster if youre backing this wave, i'd love to chat.
English
working on it 🫡
Shahin Farshchi@Farshchi
@wsj: The world’s top drugmakers—operating globally and mostly headquartered in the U.S., Europe and Japan—booked $68.7 billion in deals from Chinese companies last year, more than tripling the $22.2 billion total of 2024, according to market tracker DealForma @medra_ai will bring that spend back to the US with its incredible autonomous lab, by helping its partners develop drugs better & faster
English
I forgot to get snacks today but Anya one of our robotics engineers decided to surprise the weekend crew with a box of cookies 🍪 she’s not even working in the office today 🥹🫶
English
Today’s snacks: super sweet strawbabies from the farmer’s market 😋 blowing the minds of our out-of-state summer interns
English
Rule at Medra. If you work weekends you’ll be well fed. Dandelion pastries for breakfast and Fourth of July Chinese bbq for lunch 😋
Science doesn’t stop, neither do we! 🚀
English
When you’re tired it’s easy to feel cynical about technology. Then you read something that makes you go HELL YEAH I want to go build cool shit. 🤘
Andreas Kling@awesomekling
If you're a married programmer and your wife isn't using at least one of your custom apps daily, you're missing out on one of the most rewarding experiences in programming. Build wifeware!! 🤓
English
@DylanQian lol can't wait to meet these people when they magically show up monday xD
English
people who get frustrated at outcomes and immediately channel it into making it better is such a strong signal for hiring esp when they can’t help themselves
English
@technooptimist1 @chamath Until death, all defeat is psychological
English
Chamath Palihapitiya @chamath "Bro, there is no failure."
"That is what other people think of you to keep you down. There's no failure. If you were left on an island and it happened, you would just brush it off and move on. It's everybody else that you think is judging you. But then, here's the secret. They don't give a fuck about you. They are living their own lives. So it's your perception of what they think of you."
"There is no failure. There's do and learn. Do and learn. Do and learn. Do and learn."
@StanfordAILab
English
@DylanQian lowkey more like Medra Gen Z Slang Class in session, something something goated
English
Overheard from a group of Medra interns: “Why is being a sweat a bad thing? You have to question all beliefs.” 😆
English
You, yes you, let’s go all in and eradicate disease
Adil Mania.@adilmania
nothing makes me more furious than watching generational talent wasting time on B2B SaaS shit instead of going all-in deep tech or consumer.
English
How do we make experimental velocity the metric that determines which AI-for-biology models have the greatest impact?
scaling data < scaling AI-guided experiments
I'm at ICML 🇰🇷, let's chat!
English