David P. Roberson, PhD

#lovepeople #addresspain #processtrauma #healaddiction #pursuejustice #welcomerefugees #embraceimmigrants #techbioentrepreneur #harvardneurosciencephd

[Dario Amodei @DarioAmodei]

Today we announced the Claude-led discovery of a molecular machine that we suspect could represent a new gene editing mechanism. Its precise function, biotechnological utility (if any), or level of significance is not yet clear, but at minimum it is work I would have been proud to do as a PhD student. The work was done mostly, though not entirely, by Claude: our life sciences team suggested a broad area of research, Claude read through the literature and a bunch of genome data and discovered something interesting, then Claude proposed experiments to verify the discovery and our team carried them out.

It’s easy to dismiss this as a one-off or curiosity, but we’ve repeatedly seen a pattern where AI performance in new intellectual domains goes from weak to superhuman in a matter of a few years. In 2023 models struggled to do math at the level of an average high-school student. In 2024 they started to do well on math competitions for the best high-schoolers in the country, in 2025 they started to solve minor open problems, in early 2026 more significant open problems, and in late 2026 they are beginning to solve the top few open problems in all of mathematics. We believe AI for biology is on a similar exponential trend.

The main difference between biology and mathematics, of course, is that math can be done purely theoretically, while biology requires experimentation. Some have used this to draw the conclusion that AI’s utility in biology will be limited. We think this is wrong. As we’ve demonstrated today, humans can collaborate with AI to perform the experiments, validate key results in a few weeks and, if necessary, work with the AI to iterate on what they find. Eventually it may even be possible for Claude itself to safely perform the experiments by autonomously controlling lab equipment, with appropriate safeguards in place, but we aren’t doing that today (our lab is also a BSL1/BSL2 facility that doesn't handle materials dangerous to humans).

More broadly, biomedical advancement has many stages — from fundamental biology discoveries, to translational research, to drug discovery, clinical trials, and finally the actual delivery of medicines and health care to patients. We are also interested in these later stages, but even simply accelerating the first stage of fundamental biological discoveries has the potential to speed up and broaden the entire pipeline. Improving our understanding of biology and sharpening biologists’ tools can drive forward all of the later stages, for example by identifying new drug targets, finding new therapeutic modalities, allowing for more precise measurement, and speeding up the experimental loop which itself further accelerates our understanding of biology. This will not in itself speed up clinical trial times, but if it succeeds it could greatly increase the number of promising candidates that go into the pipeline — an increase in throughput even though latency remains.

In Machines of Loving Grace, I wrote about AI’s potential to “cure most diseases in 5-10 years” — a goal that sounds impossible, but one I believe is just barely possible if AI is applied to every stage of the pipeline. The first step is showing that AI can first help with, and then drive, biological discoveries.

Claude’s discovery is the latest in a line of related prior work that goes back decades, beginning with systems like CRISPR, and continuing with discoveries like the bridge recombinase and VIPR in the past few years. Recently, there has been heightened interest in systems based on reverse transcriptase (RT) enzymes, the enzyme underlying the system Claude identified. And most recently, a Stanford team working independently described a novel RT system with an associated non-coding array that is in some ways similar to the one Claude found, though they are distinct systems that evolved independently from each other. I believe that we’re at the very beginning of finding such systems and developing them into powerful tools for biotechnology.

I’m proud of the resources Anthropic has invested in accelerating the public benefits of AI through the life sciences, and we’re aiming both to grow our life sciences team and to work with other scientists to extend this approach to a broad range of problems. If you have a proposal for a research collaboration or are interested in joining our life sciences team, please reach out.

[Samuel Hume @DrSamuelBHume]

This is cool — of the ~20,000 human proteins, 610 are the targets of approved drugs:

[Rebecca Neeser @RebeccaNeeser]

Excited to see this out! Coming from drug discovery into ML, evaluating generative molecules is one my favorite parts. With this model, finding flaws got noticeably harder and critiques had to become much more nuanced. That’s a massive step forward.
Congrats to the team!

[David P. Roberson, PhD @dproberson]

Both of you are scientific heroes of mine - not only for your discoveries, but also the way you use the platform your work has earned you to speak up and speak out. I’ve seen both of you do this consistently, even doubling down in the line of fire. Keep pushing scientific boundaries and breaking societal barriers, brothers. You’re making a difference!

[Jason Kelly @jrkelly]

Awesome work by @ginkgo team adding 50 more robots and lab devices to Nebula, our autonomous lab. This thing is insane.

You can price out and order your favorite experiment here:
https://t.co/uycfUaGYu0

Or happy to build an autonomous lab for you at your site!

[Jaynit @jaynitx]

Marc Andreessen on Elon Musk: "Every week he identifies the biggest problem and fixes it. That's 52 problems solved per year."

"He has an operating method that is very unusual by modern standards. I'm not aware of another current CEO who operates the way that he does.

And I think probably the single biggest question in all of business right now is... why don't more CEOs operate the way that he does?"

Andreessen explains:

"If you go back in history, you find characters more like him. The industrialists of the late 1800s, early 1900s... Henry Ford, Andrew Carnegie, Thomas Watson who built IBM."

On what makes Musk different:

"The top-line thing is this incredible devotion from the leader of the company to fully, deeply understand what the company does. To be completely knowledgeable about every aspect of it. To be in the trenches, talking directly to the people who do the work. Deeply understanding the issues. Being the lead problem solver in the organization."

The method:

"Basically what Elon does is he shows up every week at each of his companies. He identifies the biggest problem the company is having that week... and he fixes it. Then he does that every week for 52 weeks in a row. And then each of his companies has solved the 52 biggest problems that year."

On everyone else:

"Most other large companies are still having the planning meeting for the pre-planning meeting for the board meeting for the presentation... with the compliance review and the legal review. It's this level of incredible intellectual capability coupled with incredible force of personality, moral authority, execution capability, focus on fundamentals... that is just really amazing to watch."

On why top talent wants to work with him:

"The side effect is he attracts many of the best people in the world to work with him. Because if you work with Elon... the expectations are through the roof in terms of your level of performance. He is going to know who you are. He is going to know what you've done. He is going to know what you've done this week. He is going to know if you're underperforming. And he may fire you in the meeting if you're not carrying your weight."

But for those who match his commitment:

"If you are as committed to the company as he is and working hard... many people who have worked for him say they had the best experience of their lives."

On delegation... and the bottleneck:

"Most CEOs have a problem knowing when to delegate. The Elon method is a little bit different. He actually delegates almost everything. He's not involved in most of the things his companies are doing. He's involved in the thing that is the biggest problem right now... until that thing is fixed. Then he doesn't have to be involved anymore. Then he can go focus on the next thing that's the biggest problem."

Andreessen uses a manufacturing analogy:

"In any manufacturing chain, there's always a bottleneck. Something keeping the line from running the way it's supposed to. Sometimes the bottleneck is at the beginning... we can't get enough raw material. Sometimes it's at the end... we don't have enough warehouses.

Or it might be somewhere in the middle. Whatever the bottleneck is... is holding everything up. Job number one is to remove that bottleneck and get everything flowing again."

Musk universalized this:

"He looks at every company like it's some sort of conceptual assembly line... sometimes a literal assembly line making cars and rockets. Any given week, there's guaranteed to be one main bottleneck. One thing holding people back."

The resolution:

"I'm going to micromanage the solution of that. I don't need to manage everything else... because everything else, by definition, is running better than that. So I can go focus on that."

On going directly to the source:

"When he identifies the bottleneck, he goes and talks to the line engineers who understand the technical nature of the bottleneck. If it's people on a manufacturing line, he's talking to people directly on the line. If it's a software development group, he's talking to the people actually writing the code."

What he doesn't do:

"He's not asking the VP of engineering to ask the director of engineering to ask the manager to ask the individual contributor to write a report... to be reviewed in three weeks. He doesn't do that. He goes and personally finds the engineer who actually has the knowledge about the thing. Then he sits in the room with that engineer and fixes the problem with them."

Andreessen explains why this inspires loyalty:

"The technical people who work with him are like... wow, if I'm up against a problem I don't know how to solve, freaking Elon Musk is gonna show up in his Gulfstream and sit with me overnight in front of the keyboard or in front of the manufacturing line and help me figure this out."

He asks:

"If you're a normal CEO running a normal company... how can you possibly compete with that?"

On why other CEOs don't do this:

"It's the way management is taught. Most classically in something like Harvard Business School or Stanford Business School. It's management as it was developed in the 1950s, 60s, 70s... the so-called scientific school of management."

He describes it:

"Management as a generic skill you can apply to any industry. You could manage a soup company or a car company... they're kind of all the same. There's a common set of management practices. It's process. How to manage the balance sheet. How to set the review schedule for meetings. How to do compliance. How to hire and motivate executives. How to resolve interpersonal conflicts. All these general business skills."

The problem:

"Those general business skills are very useful in lots of contexts. But that training gives you none of what you need to go do what Elon does."

Andreessen concludes:

"Elon pushes as far as he can... not doing all the stuff you're classically trained to do... so that he can spend all of his time doing the things only he can do. And it turns out that has this incredible catalytic, multiplicative effect. His companies are just incredibly amazing."

[IASP @IASPpain]

Clifford Woolf, MB, BCh, PhD, will present a Presidential Lecture at the #2026WorldCongressonPain exploring spontaneous neuropathic pain as a distinct and clinically significant component of neuropathic pain. Register by 15 May for the Early Bird rates. https://t.co/TmrPDq5r17

[David P. Roberson, PhD @dproberson]

@DavidChoiMusic @brivael there will always be exceptions to the rule. it doesn’t mean the solution to the problem is invalid. it means we need more nuance to solve it more efficiently

[David P. Roberson, PhD @dproberson]

This changes so much in my world. 🚀

[Ejaaz @cryptopunk7213]

my god anthropic really going after the entire design industry:

> Claude x Blender feature automates 3D modelling.

> artists, engineers and architects who spend WEEKS creating 3d visuals can now automate the entire skill

> anthropic also integrated 50+ adobe tools today.

so many artists and designers hating on this are missing the point:

claude + creative tools will make you better at your job

artists won’t be replaced by ai. but they might be replaced by someone who learned to use these tools to elevate their skill level

unreal

[IASP @IASPpain]

Clifford Woolf, MB, BCh, PhD, will present a Presidential Lecture at the #2026WorldCongressonPain exploring spontaneous neuropathic pain as a distinct and clinically significant component of neuropathic pain. Register by 15 May for the Early Bird rates. https://t.co/TmrPDq5r17

[David P. Roberson, PhD @dproberson]

@itchdoctor 100% agree

[David P. Roberson, PhD @dproberson]

@ZacharyLevi Boom!

[Ishmail Abdus-Saboor @IshmailSaboor]

Much respect to Scott Waddell for retracting the paper after catching a major mistake which made a subset of the data incorrect. And respect to Gaby Maimon for catching the error in the raw data and informing the authors. This is how science moves forward.

[David P. Roberson, PhD @dproberson]

@cheyennemedders @TedLogan1010 As a vampire, I can confirm this is true

[Noah @NoahKingJr]

Coders in 2030 be like:

[David P. Roberson, PhD @dproberson]

@TheOhioAlien @maximumpain333 This kind of angry response is a very natural reaction to someone questioning your very identity. It’s not irrational. No doubt, for Apollo astronauts, their landing on the moon very much became central to their identity.

[David P. Roberson, PhD @dproberson]

@maximumpain333 If American astronauts never went to the moon, you can count on it that our adversaries like China and Russia would have called us out on it decades ago. Of course, the same people who think we didn’t go to the moon also think the earth is 6,000 years old and flat. 😂

[David P. Roberson, PhD @dproberson]

@JasonSynaptic wow, that’s impressive

[Vala Afshar @ValaAfshar]

Nvidia CEO: Greatness does not come out of intelligence, it comes from character.

Character is not formed out of smart people: it is formed out of people who have suffered.

[David P. Roberson, PhD @dproberson]

@mendezmendez @PennMedicine @FlyBottleEscape @Nature Congrats @FlyBottleEscape and team. Really exciting paper!