Eric Abrams

Head of Life Sciences @AnthropicAI

[Eric Abrams @eric_kabrams]

I believe that the best way to accelerate curing disease is to run lots of real drug programs with Claude, across disease areas and stages of development, and get a little better at it every day. The bottlenecks are everywhere, not just in the initial stages of molecule design that tend to dominate the public discourse.

Today we're sharing early results from six research partnerships, using Claude in things like CAR-T design, lead optimization for antibodies, drafting of regulatory documents, and partnering with a new biotech company built to run many more programs than a small team usually can. We're also sharing results for the first time from our own drug programs in rare childhood disease.

Our approach to drug development owes a lot to @Prof_Oak_ . His "pipeline-in-a-person" vision was the first take on how to use AI in drug development that I found to be simultaneously AI-pilled and rigorously grounded in cold, hard reality. Welcome to Anthropic, Cody!

[Anthropic @AnthropicAI]

Over the past few months, Anthropic scientists have engaged in deep research partnerships with six organizations developing new medicines. We've also begun drug discovery efforts of our own.

In doing so, we’ve learned much more about how Claude can support the many stages of drug development.

In this blog post, we share some early results from these efforts, and describe how this work fits in with our wider ambitions in the life sciences.

https://t.co/EVbxkdDTgJ

[Eric Abrams @eric_kabrams]

we took the prompt that led a Mythos 5 agent to the original ART discovery and re-ran it over and over on other Claude models.

It's fascinating to watch their personalities come out.
Mythos 5 is disciplined and quietly excited. Opus 5 is over-the-top and tends to bask in the glory of a eureka moment.

[Eric Abrams @eric_kabrams]

Discoveries like CRISPR, Taq polymerase (PCR), and restriction enzymes all came from looking closely at things that already exist in nature and finding useful tools. These kinds of discoveries, ones that reshape entire fields, come along only once in a while. We set up a research group to try to dramatically increase the rate of this type of research, with human scientists collaborating with AI in every step of the process.

I’m excited to share our first results. While we don’t yet know of the significance of Claude’s discovery, I’m excited by several aspects of this work. To do this, Claude had to exercise some research taste and judgement—there’s no procedure to apply to determine which stretches of DNA might constitute something new and interesting. It’s long been obvious that frontier models are great at doing technical work, but this represents one of the first examples I’ve personally experienced of AI contributing to a scientific project on a creative level.

Our lab looks like a typical molecular biology lab. We do research that involves only the lower-levels of the biosafety risk level (BSL-1 and BSL-2) and we do not handle pathogens that can infect humans. All of the lab work is performed by human scientists.

We’re at the very beginning of this work.

[Anthropic @AnthropicAI]

Claude has discovered a previously unknown enzyme system hidden in the DNA of bacteriophages. Beside the enzyme’s gene sits a long array of repeating DNA—a structure that looks somewhat similar to CRISPR.

We don’t yet understand what this system does, but only a handful of known systems share its features, and all of them are able to cut, copy, and paste DNA. Historically, the discovery of such programmable systems has helped revolutionize medicine. CRISPR, for instance, is now the foundation of genetic medicines. But it will take much more work to learn what this system does, and whether it can be put to similar use.

Read more: https://t.co/RuEosScSMb

[Eric Abrams @eric_kabrams]

Today we launched Opus 5.5, which matches and even beats Mythos 5.1 on many bio-related tasks. We launched Opus 5.5 with the same safeguards as Fable 5.1, which means that most bio-related queries fall back to Opus 5. To use Opus 5.5 for life sciences work, enroll in our Life Science Verification Program (LSVP).

I imagine that many in the bio community will be frustrated by this decision and I want to share more context on how I think about these things. I also want to say upfront—bio is one of our most significant areas of focus, and it’s why I’m here. When we set about designing the new LSVP, it was with the goal of providing broad access for life sciences professionals with far fewer annoying classifier misfires, while also protecting against many of the more sophisticated risks.

Naively, it would seem that we have to choose between enabling broad access and protecting against misuse—I view this as a false choice. The key is to think of developing these safeguards as an engineering problem that requires the same level of investment, ingenuity, and ops execution as developing our frontier models, or any other large scale technical project.

The risks we’re defending against—which, as we shared in our threat report a few weeks ago, are no longer hypothetical—were certainly not obvious to me until recently, and I imagine are also not top of mind for most biologists. When most people picture bio risk, they imagine a small group trying to develop a bioweapon in a single session, from their own account. The threats that we need to defend against are much more sophisticated than that. They include malicious work split over hundreds of distinct benign-looking sessions run from different accounts, and legitimate organizations’ accounts being hacked by a well-resourced actor or misused by a rogue employee.

To protect against these risks and provide a better experience for our users who are doing legitimate life sciences work, in the LSVP we’ve introduced the concept of use-case grants. While it isn’t always possible to distinguish benign from harmful usage in isolation, it becomes much easier when comparing activity against a team’s own description of intended usage. You describe what your team works on at a high level, without revealing anything sensitive. Data are retained for 30 days, and automated offline review compares usage against that description and flags notable mismatches for a strictly compartmentalized review team. My team, the life sciences org, can never access this data. Because of these additional safeguards, for users in the LSVP we're able to tune down the classifiers so that they misfire far less often, without sacrificing on safety.

I’ve described why I believe that the LSVP provides a better way for people to use our most capable models in bio, but one can still wonder: why are we implementing this now? There is no unambiguous threshold for when model capabilities reach a point that they start posing grave risks. But capabilities in bio, as in most other domains, progress very quickly and we should not be in the business of trying to time it perfectly. When in doubt, we’ll always choose to proceed with caution, and I’d much rather implement this new system too early than too late.

The LSVP is currently in beta, which means that there are lots of improvements that we’re working on. We’re working to expand it so that individuals can get access—for example, for people developing or identifying treatments for a family member’s cancer or rare disease. Personally, I feel both the urgency to get our models into the hands of the broader life sciences community and a visceral sense of the stakes in getting it right. I think this is the right balance, and I look forward to seeing what people do with our most capable models.

https://t.co/X4kYEkzbYP

[Eric Abrams @eric_kabrams]

Today we’re launching the Life Science Verification Program, enabling life science professionals to use Mythos-class models in biology-related work for the first time. Applications are open at https://t.co/nNIFzZcbz6

Over the past few months, we’ve been hard at work developing a new set of safeguards that offer a better experience for those doing biology work, while also providing more protection for our users against the increasingly sophisticated attempts at misuse that we’ve seen on our platform.

DM me if you have questions about the program or trouble with the enrollment process.

[Anthropic @AnthropicAI]

Today we’re opening applications for the Life Sciences Verification Program.

Through the LSVP, life science professionals can use our models—including, for the first time, Mythos—with a new set of safeguards designed to enable the full range of biology-related work. We designed these new safeguards to provide a better experience for biologists and more protection from risk of misuse.

The program is launching in beta for teams of all kinds—from academic labs to startups, pharma companies, and more. We will continue to improve the program and expand access to individual Pro and Max plans over time.

Learn more about these access grants and apply: https://t.co/uALS2lZuN4

[Eric Abrams @eric_kabrams]

@ImmunoFever @claudeai dm me if you’re having trouble, I’ll help get this resolved asap

[Eric Abrams @eric_kabrams]

@aaronmring we are working literally as hard as we can to make this happen. Please hang on a little longer.

[Eric Abrams @eric_kabrams]

We’re also expanding the scale of our AI for Science program, with large-scale tokens grants for researchers working on high-impact projects.

Meanwhile, we continue to increase the scale of our bio efforts, and I expect to share more on this soon.

[Eric Abrams @eric_kabrams]

Our models are now capable of accelerating many aspects of scientific research. I’ve been having endless fun pointing Claude at problems I’ve always cared about, and in some cases watching it make progress for the first time.

Today, we’re expanding our focus beyond bio to include support for every scientific discipline. We’re launching a new plan for scientists and are making an initial 10,000 seats available for free and at heavily discounted rates.

[Claude @claudeai]

Starting today, 10,000 scientists across every field, from math to chemistry to physics and more, can get Claude through our new Claude Team plan for scientists. Standard seats are free, and premium seats with 5x usage limits are $15 per month, an 80% discount, for one year.

Claude is becoming increasingly capable of scientific work, with recent progress on problems from advanced physics calculations to protein design. Alongside that progress, we've been investing in the research community: Claude Science launched in June, and our AI for Science program funds high-impact projects with free credits. Today's expansion builds on both.

Principal investigators (or equivalent) at academic and nonprofit research institutions can sign up, then add the researchers in their group. Over the coming months, we plan to extend the program well beyond the initial 10,000 seats.

Learn more: https://t.co/RTG1JxWi4Q

[Eric Abrams @eric_kabrams]

Finally, it is among our team's highest priorities to make our most capable models available to the scientific community for biology research, and we expect to share more on this soon.

[Eric Abrams @eric_kabrams]

Today we shared an update on Claude’s capabilities in protein design and analytical chemistry. These works illustrate something that is often glossed over: we want Claude to do the task end-to-end, whatever that entails. Today, this includes selecting from and using specialized models as tools. In the future, this may include Claude deciding to train its own specialized models.

In the case of binder design, Claude selects from the best structure and design models available. As a human would, we see Claude stitch specialized model calls together, inspect the design, tweak some things, think about it more, and try again until satisfied. On many of the targets we tried, Claude’s results are competitive with the best that have been achieved previously.

[Anthropic @AnthropicAI]

Many drugs work by binding to a specific target in the body and blocking or changing what it does. An important first step in the drug development process is designing a molecule that can bind tightly to its target. Traditionally, that's meant weeks or months of expert work per target, sifting through a large number of candidates to identify the few that work.

We wanted to test if Claude could successfully design novel protein binders from scratch (also called de novo design). With a protein design prompt written by a human expert, Claude autonomously designed protein binders against 14 out of 15 targets.

We then worked with Adaptyv Bio and Twist Bioscience, who independently built and tested the proteins Claude designed.

[Eric Abrams @eric_kabrams]

There is much more that goes into developing a molecule with drug-like properties, and we’re continuing this work so that Claude can perform all of it, with established modalities like antibodies and small molecules.

And these tasks are themselves just one small part of our larger efforts to speed up drug development end-to-end, many aspects of which have more to do with policy and operational bottlenecks.

[Eric Abrams @eric_kabrams]

I've always been drawn to solving science and engineering problems to benefit people. Throughout my career, it has been an honor to partner with the federal government in these pursuits. The ability to express reasonable disagreement with our government, without untoward retaliation, is part of the fabric of American society, and is part of what makes our technology development environment the best in the world. I'm hopeful that we can continue doing important work with our federal government partners, and I'm proud to be part of the Anthropic team.

[Anthropic @AnthropicAI]

A statement on the comments from Secretary of War Pete Hegseth.

https://t.co/Gg7Zb09IMR

[Eric Abrams @eric_kabrams]

fare thee well, San Geraldo

[Eric Abrams @eric_kabrams]

@eshear Essays on Life Itself - Robert Rosen
https://t.co/iUzn5HKGI2