Kurt Thorn
@kurtthorn.bsky.social
Chief Technology Officer at ArrePath. Using Chemistry ML, imaging, and computer vision to accelerate antibiotic discovery and design. Formerly Zymergen, Nikon Imaging Center at UCSF.
Reposted by Kurt Thorn
This was a joy yesterday, I'm still buzzing from it!
A non-exhaustive list of cool things that happened (anonymized, of course):
- folks from many experience levels showed up
- we talked about our feelings, and we all felt pretty anxious and overwhelmed in this moment
- we got to crafting
A non-exhaustive list of cool things that happened (anonymized, of course):
- folks from many experience levels showed up
- we talked about our feelings, and we all felt pretty anxious and overwhelmed in this moment
- we got to crafting
Tfw you show up in SF to run a very different kind of AI event
and a lot of people are ready for something different 🤔
and a lot of people are ready for something different 🤔
Folks, have you been overwhelmed by ~All This~ and wanting a gentle compassionate space to think about agentic coding??
Good news: if you're in SF & have free time tomorrow afternoon you can come hang out with me and @anthrocypher.bsky.social in a popup casual learning event ❤️
luma.com/rdw3dq4h
Good news: if you're in SF & have free time tomorrow afternoon you can come hang out with me and @anthrocypher.bsky.social in a popup casual learning event ❤️
luma.com/rdw3dq4h
February 5, 2026 at 8:43 PM
This was a joy yesterday, I'm still buzzing from it!
A non-exhaustive list of cool things that happened (anonymized, of course):
- folks from many experience levels showed up
- we talked about our feelings, and we all felt pretty anxious and overwhelmed in this moment
- we got to crafting
A non-exhaustive list of cool things that happened (anonymized, of course):
- folks from many experience levels showed up
- we talked about our feelings, and we all felt pretty anxious and overwhelmed in this moment
- we got to crafting
Reposted by Kurt Thorn
These are all totally reasonable. My hope is that when the Republicans reject them - as they are very likely to - that Schumer and Jeffries just walk out until they’re met.
My fear is instead that they’ll use this list as a starting point and negotiate a lot of it away before caving.
My fear is instead that they’ll use this list as a starting point and negotiate a lot of it away before caving.
Schumer & Jeffries lay out the Democratic demands for a DHS funding deal in a letter to the top Republican leaders
February 5, 2026 at 2:17 PM
These are all totally reasonable. My hope is that when the Republicans reject them - as they are very likely to - that Schumer and Jeffries just walk out until they’re met.
My fear is instead that they’ll use this list as a starting point and negotiate a lot of it away before caving.
My fear is instead that they’ll use this list as a starting point and negotiate a lot of it away before caving.
Reposted by Kurt Thorn
Remarkable performance.
Sir Ian McKellen performing a monologue from Shakespeare’s Sir Thomas More on the Stephen Colbert show. Never have I heard this monologue performed with such a keen sense of prescience. Nor have I ever been in this exact historical moment.TY Sir Ian, for reaching us once again.
#Pinks #ProudBlue
#Pinks #ProudBlue
February 5, 2026 at 2:39 PM
Remarkable performance.
Reposted by Kurt Thorn
Very interesting for folks working in AI and drug discovery. Several surprises (at least to me).
openadmet.ghost.io/lessons-lear...
openadmet.ghost.io/lessons-lear...
Lessons Learned from the OpenADMET - ExpansionRx Blind Challenge
Maria Castellanos Hugo MacDermott-Opeskin Jon Ainsley Pat Walters
The ExpansionRx challenge, launched on October 27, 2025, closed two weeks ago! We are extremely grateful to everyone who participate...
openadmet.ghost.io
February 4, 2026 at 8:58 PM
Very interesting for folks working in AI and drug discovery. Several surprises (at least to me).
openadmet.ghost.io/lessons-lear...
openadmet.ghost.io/lessons-lear...
Reposted by Kurt Thorn
Anthropic’s Super Bowl ad which criticizes AI chatbots that run ads (aka ChatGPT) just dropped. They aren’t pulling any punches and I love the song choice.
February 4, 2026 at 8:14 PM
Anthropic’s Super Bowl ad which criticizes AI chatbots that run ads (aka ChatGPT) just dropped. They aren’t pulling any punches and I love the song choice.
Wow, very useful!
In the last 33 days, Claude Code has had 54k messages.
In the last 33 days, Claude Code has had 54k messages.
just found out claude code has `/insights` aka claude code wrapped
it produces an html page with details about your usage and suggestions for using the tool more effectively
extremely interesting
it produces an html page with details about your usage and suggestions for using the tool more effectively
extremely interesting
February 4, 2026 at 7:25 PM
Wow, very useful!
In the last 33 days, Claude Code has had 54k messages.
In the last 33 days, Claude Code has had 54k messages.
It’s getting harder to tell the difference between real headlines and parody headlines
Judge Incredulous as Trump Lawyer Asks Him to Create New Law for Mark Kelly Retribution Crusade talkingpointsmemo.com/news/judge-m...
Judge Incredulous as Trump Lawyer Asks Him to Create New Law for Mark Kelly Retribution Crusade
WASHINGTON, DC–Federal District Court Judge Richard Leon expressed shock Tuesday that the...
talkingpointsmemo.com
February 4, 2026 at 3:25 AM
It’s getting harder to tell the difference between real headlines and parody headlines
Reposted by Kurt Thorn
Folks, have you been overwhelmed by ~All This~ and wanting a gentle compassionate space to think about agentic coding??
Good news: if you're in SF & have free time tomorrow afternoon you can come hang out with me and @anthrocypher.bsky.social in a popup casual learning event ❤️
luma.com/rdw3dq4h
Good news: if you're in SF & have free time tomorrow afternoon you can come hang out with me and @anthrocypher.bsky.social in a popup casual learning event ❤️
luma.com/rdw3dq4h
Claude Code (Cr)afternoon · Luma
Have you seen people rave about Claude Code, and wondered whether agent-assisted coding is for you?
We're people who both work in and around software teams,…
luma.com
February 3, 2026 at 11:32 PM
Folks, have you been overwhelmed by ~All This~ and wanting a gentle compassionate space to think about agentic coding??
Good news: if you're in SF & have free time tomorrow afternoon you can come hang out with me and @anthrocypher.bsky.social in a popup casual learning event ❤️
luma.com/rdw3dq4h
Good news: if you're in SF & have free time tomorrow afternoon you can come hang out with me and @anthrocypher.bsky.social in a popup casual learning event ❤️
luma.com/rdw3dq4h
What’s the general opinion on Claude writing rust? Is it worth migrating my backend code from python to rust, even if I don’t know rust myself? I rarely read the python code, so wondering if the gains in type safety / compile time checks / etc are worth it.
February 3, 2026 at 4:56 PM
What’s the general opinion on Claude writing rust? Is it worth migrating my backend code from python to rust, even if I don’t know rust myself? I rarely read the python code, so wondering if the gains in type safety / compile time checks / etc are worth it.
Reposted by Kurt Thorn
acrobat shouldnt have ai because ai is too good for acrobat
i wish all these random programs would stop adding AI assistants. not because AI is bad, but because these programs are bad and i don't want them implementing ai. like adobe acrobat. come on.
February 3, 2026 at 4:23 AM
acrobat shouldnt have ai because ai is too good for acrobat
In fairness, this is an easy mistake to make if you code your own agent too … I burned a lot of tokens with malformed tool calls throwing errors in a loop.
🤷♀️
February 1, 2026 at 2:37 AM
In fairness, this is an easy mistake to make if you code your own agent too … I burned a lot of tokens with malformed tool calls throwing errors in a loop.
I’m finding writing an agent a really interesting experience. I’m trying to automate all the chemiinformatics and chemistry ML tooling we’ve built at ArrePath over the last four years.
January 31, 2026 at 6:54 PM
I’m finding writing an agent a really interesting experience. I’m trying to automate all the chemiinformatics and chemistry ML tooling we’ve built at ArrePath over the last four years.
Tell me five classes you took in college:
Advanced Bioorganic Chemistry
Music from 1945 to the present
Medieval manuscript illumination
Philosophy: Theory of Knowledge
Chemistry lab (for majors)
Advanced Bioorganic Chemistry
Music from 1945 to the present
Medieval manuscript illumination
Philosophy: Theory of Knowledge
Chemistry lab (for majors)
Tell me five classes you took in college.
Molecular Biology of HIV
History of the Civil War
Organic Chemistry (2.5X, lol)
Science Writing
Physical Chemistry (woof)
Molecular Biology of HIV
History of the Civil War
Organic Chemistry (2.5X, lol)
Science Writing
Physical Chemistry (woof)
Never mind the jobs you had, tell me five classes you took in college.
Physics 2 (whoa, FIELDS?!?)
Astronomy 2 (whoa, order of magnitude thinking/proportional reasoning?!?)
Origins of Life
History of WWII
Relativity & Cosmology
Physics 2 (whoa, FIELDS?!?)
Astronomy 2 (whoa, order of magnitude thinking/proportional reasoning?!?)
Origins of Life
History of WWII
Relativity & Cosmology
January 31, 2026 at 6:50 PM
Tell me five classes you took in college:
Advanced Bioorganic Chemistry
Music from 1945 to the present
Medieval manuscript illumination
Philosophy: Theory of Knowledge
Chemistry lab (for majors)
Advanced Bioorganic Chemistry
Music from 1945 to the present
Medieval manuscript illumination
Philosophy: Theory of Knowledge
Chemistry lab (for majors)
Reposted by Kurt Thorn
The future is software writing its own software. Which is why I'm so in love with Pi: a coding agent that can extend itself :) lucumr.pocoo.org/2026/1/31/pi/
Pi: The Minimal Agent Within OpenClaw
A gentle introduction to the Pi coding agent and why I think it’s a glimpse into the future of software.
lucumr.pocoo.org
January 31, 2026 at 2:23 PM
The future is software writing its own software. Which is why I'm so in love with Pi: a coding agent that can extend itself :) lucumr.pocoo.org/2026/1/31/pi/
Reposted by Kurt Thorn
"Luchs speaks five languages: English, French, Italian, and some German and Russian. She approached grasping Gen Z parlance like she was learning another language.
...An Instagram commenter wrote: 'She’s so natural with it too like it’s not even cringe.'"
...An Instagram commenter wrote: 'She’s so natural with it too like it’s not even cringe.'"
Alison Luchs, who has worked at the National Gallery of Art for 47 years, agreed to learn Gen Z slang and make videos because she wanted to raise interest in the museum’s art.
She never expected to slay. https://wapo.st/45BXc3S
She never expected to slay. https://wapo.st/45BXc3S
January 30, 2026 at 5:14 AM
"Luchs speaks five languages: English, French, Italian, and some German and Russian. She approached grasping Gen Z parlance like she was learning another language.
...An Instagram commenter wrote: 'She’s so natural with it too like it’s not even cringe.'"
...An Instagram commenter wrote: 'She’s so natural with it too like it’s not even cringe.'"
Moltbook is a lot of fun to read - it’s a social media platform for agents.
January 30, 2026 at 2:26 PM
Moltbook is a lot of fun to read - it’s a social media platform for agents.
Having spent some time building an agent powered by Opus, this feels right - my agent feels much less smart than Claude Code.
The things that changed in the past 6 weeks at least in Claude Code are 100% how they handle MCP calling and having tool calling tools, and how they handle sub agents / nested agents.
January 30, 2026 at 4:12 AM
Having spent some time building an agent powered by Opus, this feels right - my agent feels much less smart than Claude Code.
Reposted by Kurt Thorn
when they tell you to go back to writing about gadgets www.theverge.com/policy/86857...
Best gas masks
“How did these people go out and get gas masks?” AG Bondi asked.
www.theverge.com
January 29, 2026 at 3:29 PM
when they tell you to go back to writing about gadgets www.theverge.com/policy/86857...
We got a 3D printer for Christmas and now my 8 year old is starting a business to sell 3D printed fidgets to kids at school. Based on his market research, there’s a lot of demand!
January 29, 2026 at 3:13 PM
We got a 3D printer for Christmas and now my 8 year old is starting a business to sell 3D printed fidgets to kids at school. Based on his market research, there’s a lot of demand!
My main vibe coded project is now up to ~40k LoC and I’ve merged 200 PRs in the last four weeks.
I think I've hit a threshold in most of my vibe coded projects where I am accelerating, somewhere past 20KLoC. The combination of types, tests, and existing patterns seems to put Claude on cruise control, especially on new features. Even the refactors and dead code grinds are getting easier :)
January 27, 2026 at 11:52 PM
My main vibe coded project is now up to ~40k LoC and I’ve merged 200 PRs in the last four weeks.
Reposted by Kurt Thorn
Working with LLMs is a skill that is often based on what you enjoyed about your work. People who love the outcome vs people who love the process.
If you enjoy the process of writing, coding, etc then you’ll hate working with LLMs but if you most enjoyed the finished product then they’re a good tool
If you enjoy the process of writing, coding, etc then you’ll hate working with LLMs but if you most enjoyed the finished product then they’re a good tool
January 27, 2026 at 12:51 PM
Working with LLMs is a skill that is often based on what you enjoyed about your work. People who love the outcome vs people who love the process.
If you enjoy the process of writing, coding, etc then you’ll hate working with LLMs but if you most enjoyed the finished product then they’re a good tool
If you enjoy the process of writing, coding, etc then you’ll hate working with LLMs but if you most enjoyed the finished product then they’re a good tool
Reposted by Kurt Thorn
100-thousand percent.
first it taught us we have been very wrong about language, and now it is teaching us we have been very wrong about "intelligence".
sure: LLMs are not that intelligent, fine! ... neither are we.
first it taught us we have been very wrong about language, and now it is teaching us we have been very wrong about "intelligence".
sure: LLMs are not that intelligent, fine! ... neither are we.
"you don't think that this is this could actually work. But it does for LLM and maybe that's actually a lot of what humans do as well."
"AI is teaching us that maybe we're not quite as smart as we think we are"
"AI is teaching us that maybe we're not quite as smart as we think we are"
January 27, 2026 at 1:27 AM
100-thousand percent.
first it taught us we have been very wrong about language, and now it is teaching us we have been very wrong about "intelligence".
sure: LLMs are not that intelligent, fine! ... neither are we.
first it taught us we have been very wrong about language, and now it is teaching us we have been very wrong about "intelligence".
sure: LLMs are not that intelligent, fine! ... neither are we.
Reposted by Kurt Thorn
I keep thinking “my god the people they’ve killed seem like exceptionally good folks” and I think the lesson there is that they’re not exceptional, but that wonderful people abound and that a baseline distrust in humanity is a shitty conservative trope that privileges power over collective strength
you don't become a martyr because you were a saint, you become a saint because you were martyred
I agree that we can’t make Alex Pretti a Benedictine monk because that’s a trap, but we can talk about the good things we did because people will see and hear them and think “I do those things too”
January 25, 2026 at 12:27 AM
I keep thinking “my god the people they’ve killed seem like exceptionally good folks” and I think the lesson there is that they’re not exceptional, but that wonderful people abound and that a baseline distrust in humanity is a shitty conservative trope that privileges power over collective strength
I’ve created a tool that lets Claude code read my chemistry agent’s conversations so it can act like a robopsychologist and help improve prompting and tool descriptions.
January 23, 2026 at 11:26 PM
I’ve created a tool that lets Claude code read my chemistry agent’s conversations so it can act like a robopsychologist and help improve prompting and tool descriptions.