←
Amir Mitchell733 posts
Amir Mitchell avatar

Amir Mitchell

@mitchell_sysbio

Systems Biologist fascinated by AI + scientific research | AssProf at @UMassMed School | Accepted my fate as a HITL

◷ Joined July 2016▧ 201 Photos and videos
Amir Mitchell avatar
Amir Mitchell@mitchell_sysbio·Sep 27

This clip is a fantastic tool to teach students so many of the basic concepts in AI/ML (and its super catchy)

this is actually just straight up good shit. genuine art. what is happening

2:37
Amir Mitchell avatar
Amir Mitchell@mitchell_sysbio·Sep 27

Worth watching. Wondering when will my fellow scientists adopt such an optimistic view (… it will probably be a while 😉)

My god this is such a good speech that every SWE needs to hear. You know what? Every person should hear it Keep the happy memories, eyes on the reality, be excited about the future. That’s the best that anyone can do

2:49
Amir Mitchell avatar
Amir Mitchell@mitchell_sysbio·Sep 26

Still unimpressed? Remember the first AI-generated video clips: short, grainy, and full of absurd errors (six-fingered hands). Compare them to clips we can now generate! Now consider this AI-based discovery loop, what will improved loops uncover in biology within a year or two? I think ART is just the canary in the coal mine. (9/9)

Amir Mitchell avatar
Amir Mitchell@mitchell_sysbio·Sep 26

Now that we see it works, we can let AI-based swarms explore public and private bio datasets and equip them only with broad scientific motivations. If we allow them to pursue unusual observations, they will occasionally make discoveries nobody explicitly asked them to make. For the first time, we can *scale* scientific serendipity! (8/9)

Amir Mitchell avatar
Amir Mitchell@mitchell_sysbio·Sep 26

Further showing how *incidental* that finding was, rerunning the research loop 10 times more failed to rediscover ART. This isn't a BUG; it tells us something profound about the discovery route and the power of non deterministic software for discovery (7/9)

Amir Mitchell avatar
Amir Mitchell@mitchell_sysbio·Sep 26

An agent decided that this particular RT lineage was interesting enough to investigate further. The process eventually reached the raw upstream DNA, noticed an unusual repeat array, and recognized that it might represent a novel finding. It wasn’t part of the research prompt, it emerged from the exploration. That’s the part we should be paying close attention to. (6/9)

Amir Mitchell avatar
Amir Mitchell@mitchell_sysbio·Sep 26

At its core, it's a beautiful example of serendipitous research leading to discovery. The research brief (still unpublished) seemingly focused on finding new protein partners of RTs. It apparently did NOT instruct the agents to search for DNA repeats or systematically inspect the non-coding DNA upstream of RT genes. Yet the research loop went there anyway. (5/9)

Amir Mitchell avatar
Amir Mitchell@mitchell_sysbio·Sep 26

I honestly believe most PIs would say this is a brilliant/promising student, since they noticed something unexpected, recognized that it might matter, and pursued it. An AI agent did exactly that, and its not trivial at all (4/9)

Amir Mitchell avatar
Amir Mitchell@mitchell_sysbio·Sep 26

I strongly suggest reading their report, not just the heavily hyped press release (maybe written with the near-term IPO in mind). After reading, ask yourself a simple question: How would you respond if a PhD student, given the same broad research brief, had made this discovery?

Amir Mitchell avatar
Amir Mitchell@mitchell_sysbio·Sep 26

Scientific discussion around Anthropic’s ART “discovery” has been mostly negative. Scientists repeatedly questioned the novelty, belittled the experimental work, or argued that the finding itself may be unimportant. I'm not arguing if they are correct, but IMO most discussions missed real breakthrough here. (2/9)

Ophir Shalem avatar
⟳ Amir Mitchell retweeted
Ophir Shalem@ophirshalem·Sep 24

Happy to share the revised and now published in Cell Systems version of our paper, where we explore why dCas9-KRAB(KOX1) protein levels drop so drastically after neuronal differentiation in WTC11 and KOL2.1J iPSC lines. Paper: doi.org/10.1016/j.cels.2026.…

Amir Mitchell avatar
Amir Mitchell@mitchell_sysbio·Sep 23

We can debate the hype around this discovery, but the *methodology* is the real transformative part. The most important skill Universities are NOT teaching is how to use AI to drive innovative research. Not surprising since most faculty are completely oblivious to this paradigm shift (for most, #AI is very just capable chatbots 🤦‍♂️)

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: anthropic.com/news/claude-di…

Amir Mitchell avatar
Amir Mitchell@mitchell_sysbio·Sep 17

This will go down in history as one of the biggest covert IP heists in history. Sad that universities are NOT joining forces to set up their own cloud based ai service 😞

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: anthropic.com/news/life-scie…

Amir Mitchell avatar
Amir Mitchell@mitchell_sysbio·Sep 9

Sad, bit this was a long time coming, Academia remains oblivious to the real threat of openAI/Anthropic accounts. Universities need to run open LLMs on their clusters and invest in open-source harnesses. Alex Krap spelled it out, yet Universities still fail to realize they also have valuable "alpha" to protect ☹️

wtf is this way to handle mathematicians work and scientific communication TLDR: Leven and Tristan worked over several months on one of the Millenium Prize Problems with various AIs to reach final interesting results. OpenAI apparently heard about it in the last days and prompted their latest models to work on the direction Leven and Tristan found fruitful. They then tried to push for controlling communication of the result and dropping Leven from authorship with some very bad taste social pressure. Hope this is not a glimpse of the future we’ll get in science research with these dominating players playing marketing games hurtful for the real scientific community.

Post attachment