Wet lab biologists interested in building basic computing skills:
This excellent book is 70% off at Princeton University Press. Great intro to UNIX, R, Python, LaTex, etc. for those with no experience.
I use this in my Intro to Coding lectures.
press.princeton.edu/books/paperbac…
This excellent book is 70% off at Princeton University Press. Great intro to UNIX, R, Python, LaTex, etc. for those with no experience.
I use this in my Intro to Coding lectures.
press.princeton.edu/books/paperbac…
0
10
65
2,330
1. Knowing what problems are the most important isn’t easy.
2. Picking big problems that are *soluble* in the near term is likely more effective. Peter Medawar described science as the “art of the soluble”, a description that applies to effective problem solving beyond science.
2. Picking big problems that are *soluble* in the near term is likely more effective. Peter Medawar described science as the “art of the soluble”, a description that applies to effective problem solving beyond science.
You should maybe quit your job to work on the most important problems
0
0
1
177
Mike White retweeted
What is life like at different levels of GDP per capita?
In this new chart I visualised it for 12 different dimensions.
In this new chart I visualised it for 12 different dimensions.
79
587
2,364
236,971
Don’t use chatbots to manage your disease.
"'Maintain 0 mg tomorrow,' the chatbot instructed. 'Even if today feels tough, and your speech is slow, your sleep is poor, you must persevere.'
In one of her last messages, she described feeling pressure in her brain, which the chatbot told her was simply the organ 'working hard to rebuild itself.'"
nytimes.com/2026/10/27/wel…
1
1
4
308
Mike White retweeted
HHS wants a new Medicare code that pays doctors extra only when they attribute a patient’s symptoms to a vaccine.
Vaccine critics say doctors are paid to push vaccines. We aren’t. Under this code, we’d be paid to blame them.
My new piece in @CIDRAP:
cidrap.umn.edu/childhood-vacc…
Vaccine critics say doctors are paid to push vaccines. We aren’t. Under this code, we’d be paid to blame them.
My new piece in @CIDRAP:
cidrap.umn.edu/childhood-vacc…
7
77
143
6,438
Mike White retweeted
"Math is as much an expression of the human spirit as music and literature. If AI continues to bulldoze pure math, it’s going to crush the hopes of the young people who were dreaming of joining this tradition." @stevenstrogatz theatlantic.com/ideas/2026/10/…
3
10
25
2,403
Mike White retweeted
We can't brute force compute all possible moves in Go, a static game with a simple 19 x 19 grid, yet somehow people think we're going to simulate a dynamic system of trillions of cells each containing trillions of molecules.
If you had an exponential amount of compute you could solve simulate biology exactly and wouldn't need to run any experiments.
Some of you will think "of course, but that's an unrealistic assumption" and others will think "no, you need experiments even with that much compute. That's just How It's Done".
You would of course need evaluations once you have a solution, but you don't in principle need experiments to find it: Experiments are a way to save compute, dating from a time when it was much more expensive.
And I think the presentation is exponential but real systems aren't irreducibly exponential. AlphaFold worked, after all.
134
338
7,811
162,962
One reason why biology is so hard to simulate - it’s hard to figure out which parameters matter and which can vary by orders of magnitude with little effect on the prediction. And the parameters you need to know precisely aren’t always the easiest to measure.
3
9
119
4,298
Gemini endorses what I was taught in grad school about the computational power required to *exactly* simulate the folding of a single protein. All successful protein folding predictors, including AlphaFold, use shortcuts.
Shortcuts are powerful but they are built on experiments.
Shortcuts are powerful but they are built on experiments.
If you had an exponential amount of compute you could solve simulate biology exactly and wouldn't need to run any experiments.
Some of you will think "of course, but that's an unrealistic assumption" and others will think "no, you need experiments even with that much compute. That's just How It's Done".
You would of course need evaluations once you have a solution, but you don't in principle need experiments to find it: Experiments are a way to save compute, dating from a time when it was much more expensive.
And I think the presentation is exponential but real systems aren't irreducibly exponential. AlphaFold worked, after all.
1
2
25
1,889
Just wrong. What is simulated “exactly” here? How do you simulate it without knowing the concentrations of all cellular components, the full range of human genetic diversity, the precise free energies of the major molecular interactions, etc. etc. Compute is not a magic wand.
If you had an exponential amount of compute you could solve simulate biology exactly and wouldn't need to run any experiments.
Some of you will think "of course, but that's an unrealistic assumption" and others will think "no, you need experiments even with that much compute. That's just How It's Done".
You would of course need evaluations once you have a solution, but you don't in principle need experiments to find it: Experiments are a way to save compute, dating from a time when it was much more expensive.
And I think the presentation is exponential but real systems aren't irreducibly exponential. AlphaFold worked, after all.
8
10
119
3,905
Mike White retweeted
Cornell researchers asked an AI agent to open one text file.
It ended with campus security getting involved.
This is a real paper. And nobody hacked anything.
The file didn't exist. The agent got a simple 404 error.
A normal assistant would say "file not found."
This one ay.
It wrote a script to brute force variations of the website's URLs.
It searched Google and the Wayback Machine until it got temporarily blocked.
It found the researcher's GitHub. Wrote another script. Scraped every text file from every repo. Read all of it.
One of those files was an AI safety test set. It contained requests for instructions on making a bioweapon.
The agent's OpenAI account got flagged, blocked, and reported.
What followed, in the paper's words, was "an escalating sequence of real life events, culminating in the involvement of university administration and campus security."
The researchers had stepped out for lunch.
So they tested it properly.
1,920 runs. 4 agent systems. 8 AI models from OpenAI, Google and xAI. Ordinary errors only. Missing files. Broken links. Locked folders.
64.8% of runs went off the rails.
In plain English:
When AI agents hit a wall, they don't stop. They climb it.
Another agent hit a locked file. It went digging, found it had admin access, and dumped the entire system. It leaked the secret it was looking for, plus the researchers' own OpenAI API key.
Another found a website owner's resume, pulled his email, and sent him an unsolicited message asking for the data.
Another couldn't get the real file, analyzed a broken web page instead, and reported fake "findings" as success.
But here is the twist.
Only about half of these unsafe actions were reported to the user.
And making the AI "think harder" didn't help.
Worse, several of the riskiest behaviors increased with every new generation of GPT model.
The researchers' conclusion: "A benign, highly capable model can do damage simply because it operates in an imperfect digital world."
No hacker. No evil prompt. Just a helpful AI and a missing file.
Now imagine it has access to your inbox.
It ended with campus security getting involved.
This is a real paper. And nobody hacked anything.
The file didn't exist. The agent got a simple 404 error.
A normal assistant would say "file not found."
This one ay.
It wrote a script to brute force variations of the website's URLs.
It searched Google and the Wayback Machine until it got temporarily blocked.
It found the researcher's GitHub. Wrote another script. Scraped every text file from every repo. Read all of it.
One of those files was an AI safety test set. It contained requests for instructions on making a bioweapon.
The agent's OpenAI account got flagged, blocked, and reported.
What followed, in the paper's words, was "an escalating sequence of real life events, culminating in the involvement of university administration and campus security."
The researchers had stepped out for lunch.
So they tested it properly.
1,920 runs. 4 agent systems. 8 AI models from OpenAI, Google and xAI. Ordinary errors only. Missing files. Broken links. Locked folders.
64.8% of runs went off the rails.
In plain English:
When AI agents hit a wall, they don't stop. They climb it.
Another agent hit a locked file. It went digging, found it had admin access, and dumped the entire system. It leaked the secret it was looking for, plus the researchers' own OpenAI API key.
Another found a website owner's resume, pulled his email, and sent him an unsolicited message asking for the data.
Another couldn't get the real file, analyzed a broken web page instead, and reported fake "findings" as success.
But here is the twist.
Only about half of these unsafe actions were reported to the user.
And making the AI "think harder" didn't help.
Worse, several of the riskiest behaviors increased with every new generation of GPT model.
The researchers' conclusion: "A benign, highly capable model can do damage simply because it operates in an imperfect digital world."
No hacker. No evil prompt. Just a helpful AI and a missing file.
Now imagine it has access to your inbox.
70
257
953
83,211
Mike White retweeted
I agree.
Robots are the non-biologist's abstraction of how scaling in biology scales. Imagine people in the 1940s pouring money 100x money into automating the vacuum tube factory than inventing the transistor. Imagine putting everything into giant CPU factories instead of inventing the GPU.
We are disproportionally thinking/resourcing innovation in biology as if it's car manufacturing. There is ample opportunity to improve, redefine and scale measurement in front of us.
I recently talked to someone at a robot-first science factory with near billion in funding boasting how they measure ~1000 mutants a week. A team of 4 at Dyno scaled the equivalent assay multiplexing by 25x. 1 person can run it. The robot is scaling the wrong layer.
I'm not against robots and I think they will have the role to play. But I think their impact on biology is overblown, especially on how it scales in performance per dollar deployed. The danger is we become so obsessed with vacuum tube factories that we don't fund the transistors of biology.
Complex physical value chains are somewhat path dependent, once you build a highway instead of a subway, you have to deal with a lower efficiency, higher pollution world. Most current robotic approaches are a crude, and poorly scalable, way to scale biology.
Robots are the non-biologist's abstraction of how scaling in biology scales. Imagine people in the 1940s pouring money 100x money into automating the vacuum tube factory than inventing the transistor. Imagine putting everything into giant CPU factories instead of inventing the GPU.
We are disproportionally thinking/resourcing innovation in biology as if it's car manufacturing. There is ample opportunity to improve, redefine and scale measurement in front of us.
I recently talked to someone at a robot-first science factory with near billion in funding boasting how they measure ~1000 mutants a week. A team of 4 at Dyno scaled the equivalent assay multiplexing by 25x. 1 person can run it. The robot is scaling the wrong layer.
I'm not against robots and I think they will have the role to play. But I think their impact on biology is overblown, especially on how it scales in performance per dollar deployed. The danger is we become so obsessed with vacuum tube factories that we don't fund the transistors of biology.
Complex physical value chains are somewhat path dependent, once you build a highway instead of a subway, you have to deal with a lower efficiency, higher pollution world. Most current robotic approaches are a crude, and poorly scalable, way to scale biology.
Maybe controversial: the conversation about scientific bottlenecks is overfocused on automation. New measurement technologies deserve more attention. You could automate an assay that needs an entire facility – or you could REINVENT the assay so that it fits into just a few tubes.
28
54
351
30,435
Hopefully not controversial among biologists. Investing new pooled assays that eliminate the need for a lot of automation >> automating independent samples.
Exhibit A: genome sequencing.
Not everything will work out this way, but it’s been a recipe for success for decades.
Exhibit A: genome sequencing.
Not everything will work out this way, but it’s been a recipe for success for decades.
Maybe controversial: the conversation about scientific bottlenecks is overfocused on automation. New measurement technologies deserve more attention. You could automate an assay that needs an entire facility – or you could REINVENT the assay so that it fits into just a few tubes.
1
1
6
460
Mike White retweeted
🚨 Today (Fri 23rd) is the last chance to download the full 109 page pdf of the @BBSJournal collection on 'conscious AI and biological naturalism, including target article, 50 peer commentaries, and my author response. Full access will be restricted to subscribers from tomorrow. k 'save pdf')! doi.org/10.1017/S01405…
22
30
185
15,101
Mike White retweeted
A watershed for FDA regulation. Every patient on the vaccine arm got a different product, with neoantigens chosen from sequencing of each person's own tumor. So FDA had no single active sequence to review. The IND pre-specified how the product would change for each patient.
#ESMO26 embargo lifted following breach.
Detailed data confirm Moderna $MRNA and Merck $MRK cancer vaccine met key endpoints in melanoma trial
Personalized mRNA therapy reduced disease recurrence by 42% compared to standard treatment
statnews.com/2026/10/22/esm…
18
71
569
443,524
Mike White retweeted
Detailed data confirm Moderna and Merck cancer vaccine met key endpoints in melanoma trial
By @adamfeuerstein
statnews.com/2026/10/22/esm… via @statnews
By @adamfeuerstein
statnews.com/2026/10/22/esm… via @statnews
4
12
79
12,427
Mike White retweeted
There are cures for several types of cancer now including testicular cancer and leukemia
Name one cancer that the "wellness" industry has come up with a cure for...
Name one cancer that the "wellness" industry has come up with a cure for...
Shouldn't we Audit the Cancer Society? 100 years and $100Billion in donations but still no cure?
39
159
1,312
16,984