Synthetic Biologist - Foundry Theorist - American Wetware - stay humble biodesigners
Cambridge, MA
Joined June 2014
- Tweets14.3K
- Following444
- Followers18.6K
- Likes50.9K
I don't think the interesting question is "can we cure all diseases in 10 years?" because we can't
The interesting question is "10 years from now, will Bio-AI still be accelerating so quickly that failed predictions about today's tech don't matter?"
Lots of people are pointing out that you pay a price for over-hyping a technology. Historically, this is true. The peak of inflated expectations is followed by the trough of disillusionment.
But that model applies to a stable technology: CRISPR or RNAi, for example. These were powerful tools with a wide, but defined, scope. The CRISPR of 2026 would be recognizable to a CRISPR researcher in 2016.
AI does not seem to be like that. AI accelerates. The AI tooling of 2036 won't look anything like the tooling today. It will be more powerful to a degree that is hard for me to comprehend.
By the time Dario's predictions about curing all diseases are false, AI will be in effect, a new technology. There will be no time to be disappointed.
I'm not sure I'm fully convinced by the above logic. But I assume some version of it is behind the choice of Dario and other neolab leaders to let rip on hype. The acceleration of AI means they don't expect to pay a price for being wrong.
Jake Wintermute 🧬/acc retweeted
Love the energy of "Become the science funder you want to see in the world"! Cool program by an awesome team backing people with money to give out as fast grants
Field Builders make the fields that make the future. More people should consider the role.
We’re backing at least 50 Field Builders to launch grant programs in emerging scientific fields. Each Field Builder will award $70,000 to researchers they recruit and receive a $5,000 honorarium for leading the work.
Applications for the Field Building Accelerator are open now through December 15, 2026: fieldbuilding.org/
Jake Wintermute 🧬/acc retweeted
Field Builders make the fields that make the future. More people should consider the role.
We’re backing at least 50 Field Builders to launch grant programs in emerging scientific fields. Each Field Builder will award $70,000 to researchers they recruit and receive a $5,000 honorarium for leading the work.
Applications for the Field Building Accelerator are open now through December 15, 2026: fieldbuilding.org/
The life of a mercenary biohacker isn’t easy
Sure, traveling to elite labs to work on biological millennium problems might seem very sexy and mysterious and cool - and it is
But sometimes I think I might love the Bio-AI game too much. It haunts me in a very sexy and mysterious way
The pangi or football fruit of indonesia must be boiled and fermented for 40 days to remove hydrogen cyanide before it can be eaten
I bet 2-3 clever gene edits could bring cyanide-free pangia to the world
en.wikipedia.org/wiki/Pangiu…
Jake Wintermute 🧬/acc retweeted
Happy Four Seasons Total Landscaping day to all who celebrate
1. Buy the patents for this peptide vape
2. Paint it black and slap some LED gamer lights on it
3. Swap out the insulin cartridge for the latest trendy peptide
4. Cyberpunk billionaire
Pfizer lost over a billion dollars on their inhaled insulin product, Exubera. There were a bunch of problems with it (most importantly that it cost twice as much as other forms of insulin) but the funniest was that patients really didn't like using it in public.
You popped a cartridge into it, the clear chamber filled with white "smoke", and then you cleared it one breath. It looked exactly like you were smoking weed out of your Pfibong.
There's still an inhaled insulin product on the market, Afrezza, which looks like a weird vape.
Jake Wintermute 🧬/acc retweeted
Nice DeepSWE score! Does it survive a lowercase filename?
Introducing MaxxBench: 89 DeepSWE tasks with small changes to the instructions or starting code. Same engineering problem, slightly different requirement.
Models clustered around 74% on DeepSWE end up 23 pp apart.
Good at coding, or good at DeepSWE? 🧵
Jake Wintermute 🧬/acc retweeted
i went to slavoj žižek’s talk in sf last night and this is what it taught me about b2b saas:
- the old world dying doesn’t necessarily mean a better one is coming. things could get much worse
- hope is an excuse to wait for someone else to do what you should be doing
- despair is good because it forces you to act
- dystopian fiction lets us secretly enjoy the horrors we pretend to condemn
-covid lockdowns were great because nobody bothered him
- creative office jobs are more exploitative than factory jobs because they demand your entire personality
- lying flat and refusing to participate in capitalism can be more revolutionary than protesting (china’s tang ping movement)
- ai executives warning us about the dangers of ai are hypocrites
- the best part of oppenheimer was florence pugh naked
- neuralink could eliminate the uncertainty and awkwardness of flirting by letting us know instantly whether someone wants to have sex with us, turning seduction into something mechanical
- the future of intimacy is two people letting their sex machines have sex while they have a nice conversation
- pursuing knowledge is an ethical duty, even if the consequences are catastrophic
- humanity might eventually create god through ai
- the disappearance of neighborhood cafés, bookstores, and public squares is destroying social life
- falling in love comes before discovering why you love someone
- marriage is more romantic now that it’s economically unnecessary. getting married is telling society to fuck off
- leaders should retire after ten years because they get too comfortable
- just as hegel said the peloponnesian war happened so thucydides could write about it, san francisco was built so hitchcock could film vertigo here
This is so funny to me like a jurassic park situation they should declare this guy innocent by reason of fun
3
49
2,703
Jake Wintermute 🧬/acc retweeted
I've joined Sentinel Bio to lead their work on synthesis screening. This means I'm taking responsibility (insofar as I can) for getting to the 2030 victory conditions published earlier this year. These are ambitious goals, and 3 years is both a short and long time: let's do this.
I appreciate all the biosecurity people who engage earnestly with my snark. I believe you are working hard on a hard problem and that you care
I will never stop snarking
This is an entirely new modality
It sidesteps many hard problems of small-molecule drug discovery
It is has a clear path to AI-driven improvements in cell targeting and receptor design for dozens of applications
Also... IT WORKS IN FREAKING HUMANS
If AI twitter actually knew anything about biotech instead of just huffing every retarded hypepost this would have a million engagements
Dreadds work in humans nejm.org/doi/full/10.1056/NE…