iiinova — where your digital self keeps evolving

Gates equates cyber and bio capabilities specifically because the prospect of a pandemic raises more alarm bells than a cyber attack. He posits that a bioweapon created by a non state actor could kill 100M people. Not plausible imo.

130

Jase Gehring @skyjase.bsky.social · 14h

No. I disclaim JD Vance. He is not my responsibility

010

Jase Gehring @skyjase.bsky.social · 14h

Looks like a great study and a powerful technology

031

Jase Gehring @skyjase.bsky.social · 29/09/2026

people complain about the EU (and it's often deserved) but the US is simply not stepping up and passing smart legislation like this.

080

Jase Gehring @skyjase.bsky.social · 29/09/2026

professor: i still think it's worth it to do a PhD other professor: i agree other professor: me too

2110

Jase Gehring @skyjase.bsky.social · 29/09/2026

yes, that would be nice, but the whole political environment is poisoned the overton window has shifted even if democrats want to do stuff (BIG IF), it will be back to republican bullshit about overreach and the deficit and Woke and whatever. we aren't getting back to the status quo ante

020

Jase Gehring @skyjase.bsky.social · 29/09/2026

second thing was... are students asking ChatGPT for questions to ask the seminar speaker during the talk? are they recording talks and uploading them to AI companies without consent? some of the questions were incoherent. and either way it sucks that i'm now questioning this stuff >:(

120

Jase Gehring @skyjase.bsky.social · 29/09/2026

went to a very nice protein design seminar today and a couple things stuck out. first, a student(?) asked a question about making chiral molecules and the safety implications. how do we stop people from using design models to make dangerous things.

120

Jase Gehring @skyjase.bsky.social · 29/09/2026

anthropic is basically moving the refusals along as it releases new models. i always need to use the second best model strange implications for their threat modeling. like they think it's dangerous, but then after a couple months (and coinciding with the next model release) they decide it's not?

110

Jase Gehring @skyjase.bsky.social · 29/09/2026

Venn diagrams, how do they work?

010

Jase Gehring @skyjase.bsky.social · 28/09/2026

I want to know if the Anthropic team thought they found something cool for the first time or if they’re a bit more realpolitik about the situation. Like is their mindset “probly in the training data, just do a press release anyway” or “probly Claude is a genius”

120

Jase Gehring @skyjase.bsky.social · 28/09/2026

We are entering the cool zone

030

Jase Gehring @skyjase.bsky.social · 25/09/2026

Think I saw this as a preprint? very very cool

000

Jase Gehring @skyjase.bsky.social · 25/09/2026

AI triage can be very helpful for grant review *after* a peer-reviewed non-inferiority study comparing to current practice. Show us the data

010

Jase Gehring @skyjase.bsky.social · 25/09/2026

Trump 2 cuts so much deeper because i've lost trust in our system of government and the american electorate itself. no election can change my mind. the US won't fix itself

140

Jase Gehring @skyjase.bsky.social · 24/09/2026

bombina bombina, source of bombesin en.wikipedia.org/wiki/Europea...

001

Jase Gehring @skyjase.bsky.social · 24/09/2026

So much packed into this one little fact

070

Jase Gehring @skyjase.bsky.social · 24/09/2026

some good posts linked here. i will add that Matt Durrant, the leader of the group that made this discovery, is himself the discoverer of bridge recombinases, a big deal from a few years ago. Matt is a world expert in building these sorts of bioinformatic tools *without Claude*

130

Jase Gehring @skyjase.bsky.social · 24/09/2026

The tech industry does this and I think it’s excellent. (usually caveat that we don’t know who is setting policy and with this admin presume it is in bad faith)

010

Reposted by Jase Gehring

Arc Institute @arcinstitute.org · 23/09/2026

Many important bacterial non-coding RNAs and structural elements remain undiscovered. To help, David Li, Garyk Brixi, Michael Fischbach, @brianhie.bsky.social & team introduce Minerva, which uses a genome language model to predict RNA base pairing, repeats, & other interactions from sequence alone.

12212

Jase Gehring @skyjase.bsky.social · 23/09/2026

little bit thirsty to release this on its own. with theminimal wet lab resources at Anthropic, would be hilarious if they get scooped by any of the extremely good nucleic acid biochemistry labs around the world

0132

Jase Gehring @skyjase.bsky.social · 23/09/2026

the big salary helps when your employer credits your work to the AI

131

Jase Gehring @skyjase.bsky.social · 23/09/2026

how many applicants we thinking for this role? my uninformed guess is something like 1k with 200-300 considered competitive. haven't hired in academia, not well calibrated

110

Reposted by Jase Gehring

Stephen Turner @stephenturner.us · 22/09/2026

Biggest takeaway: Not either-or: We can make beneficial use easier, dangerous use harder or impossible. We need better measurements of AIxBio capabilities and better understanding of AI uplift in bio so we can better prioritize and calibrate risk reduction responses.

011

Reposted by Jase Gehring

Stephen Turner @stephenturner.us · 22/09/2026

1. OAI incident was a warning shot 2. Be skeptical of doomsday countdowns 3. AI doesn't make you a biologist overnight 4. Take risks seriously but don't overstate them 5. Layered defenses 6. AI is moving fast, calibrated governance must keep pace www.cbs19news.com/news/inside-...

cbs19news.com

Inside the Numbers 09/21

Stephen Turner from the UVA School of Data Science comes to the CBS19 studios to talk about artificial intelligence.

211

Jase Gehring @skyjase.bsky.social · 23/09/2026

Looks like a nice paper. Cool approach and OGTAC is a very fun name

020

Jase Gehring @skyjase.bsky.social · 22/09/2026

$1B over five years is like… 0.2-0.4% of their *current* revenue??

000

Jase Gehring @skyjase.bsky.social · 21/09/2026

my impeachment+conviction threshold is a steady 25%

031

Jase Gehring @skyjase.bsky.social · 21/09/2026

gonna be hard to reach a successful ending seeing as we lost the war in April

010

Jase Gehring @skyjase.bsky.social · 21/09/2026

I agree with this if and only if the pool of grad students and postdocs is not increased. We already way overproduce biomedical PhDs relative to the job market. Introduce legitimate career ladders for basic science. make graduate and postdoc stipends a salary and tie it to local cost of living

020

Jase Gehring @skyjase.bsky.social · 21/09/2026

This and 3. USA is a much less attractive option. This is how shrinking American science results in net loss of global scientific productivity

040

Jase Gehring @skyjase.bsky.social · 21/09/2026

Also incredible that they used Claude and, surprise surprise, a major pharma company was already pursuing this target with a major drug program Think anyone at Merck has an Anthropic subscription?

251

Jase Gehring @skyjase.bsky.social · 21/09/2026

“But other scientists note that the Virtual Biotech has not been vetted in the crucible of real-world drug discovery, and its predictions were not validated through experiments, let alone clinical trials.“ really no one out there doing it like James Zou

192

Jase Gehring @skyjase.bsky.social · 21/09/2026

State of the art for open source binder design

063

Jase Gehring @skyjase.bsky.social · 20/09/2026

this is cool but i expect a lot more from Anthropic and the team they've hired. i think they are planning to make real advances on 1) bio models 2) integrating claude into research 3) biological discoveries a foray into drug development itself is perhaps a stretch even for them

140

Jase Gehring @skyjase.bsky.social · 19/09/2026

My friends started Novome. The story of their clinical failure is fascinating. Immediately what I thought about at the top of Akos‘ thread. Cool stuff

021

Jase Gehring @skyjase.bsky.social · 19/09/2026

What I will not accept is renaming a traditional laboratory to a “bio lab” No. They already took “lab” for themselves because it sounds cooler than an office with laptops and standing desks. They don’t get to reinvent how we refer to actual laboratories

030

Jase Gehring @skyjase.bsky.social · 19/09/2026

what "problem" give them the magnesium they deserve

020

Jase Gehring @skyjase.bsky.social · 18/09/2026

They‘ve stated that their intention is to make fundamental bio discoveries and to target rare diseases. they see pharma as major partners and don’t want to scare them off. Matt Durrant (bridge recombinases) and Nathan Frew (Prescient protein design) are there

240

Jase Gehring @skyjase.bsky.social · 18/09/2026

being completely serious and not condescending it really helps if you work at Stanford

150

Jase Gehring @skyjase.bsky.social · 18/09/2026

my recommendation is private and state-level red-teaming to determine the nature of the threat and build an appropriate response. if we found significant AI uplift and gaps in our regulatory regime that can't be closed, i would rapidly update my views everything else is just armchair speculation

000

Jase Gehring @skyjase.bsky.social · 18/09/2026

AI safety people: listen to the biosecurity policy people biosecurity policy people: listen to the defense officials defense officials: listen to the AI safety people

000

Jase Gehring @skyjase.bsky.social · 18/09/2026

Many expert microscopists questioning the authenticity of this movie, the winner of a prestigious imaging competition

010

Jase Gehring @skyjase.bsky.social · 17/09/2026

Guys I think I’m done. It’s time to leave

010