Ethan
@ethanrosenthal.com
data/ml/ai @ runway in 🗽
formerly spent time in startups, biggish tech, and physics labs
www.ethanrosenthal.com
formerly spent time in startups, biggish tech, and physics labs
www.ethanrosenthal.com
Reposted by Ethan
The "This is the most fun I've had..." sentiment follows my experience as well.
I'll add: it's the most powerful computers have felt to me since I first learned to code. Now it's easier than ever to make them do what I want them to.
I'll add: it's the most powerful computers have felt to me since I first learned to code. Now it's easier than ever to make them do what I want them to.
After making dozens of prototypes with AI coding agents, here's what I think: human developers are still essential
In fact, instead of taking jobs, this new class of software tools may make workers busier than ever. I wrote about my experience for Ars Technica:
arstechnica.com/information-...
In fact, instead of taking jobs, this new class of software tools may make workers busier than ever. I wrote about my experience for Ars Technica:
arstechnica.com/information-...
10 things I learned from burning myself out with AI coding agents
Opinion: As software power tools, AI agents may make people busier than ever before.
arstechnica.com
January 19, 2026 at 2:54 PM
The "This is the most fun I've had..." sentiment follows my experience as well.
I'll add: it's the most powerful computers have felt to me since I first learned to code. Now it's easier than ever to make them do what I want them to.
I'll add: it's the most powerful computers have felt to me since I first learned to code. Now it's easier than ever to make them do what I want them to.
I wonder how many people are in the world now? Oh wow, 8.3B. I remember when it was like 6 something. I wonder how old the universe is now? Oh ok, still 13.8B years old.
January 18, 2026 at 4:41 PM
I wonder how many people are in the world now? Oh wow, 8.3B. I remember when it was like 6 something. I wonder how old the universe is now? Oh ok, still 13.8B years old.
Yes, training foundation models is a gold rush, but nobody tells you how much fun the work actually is. Just a lovely mix of research, systems, and scale, from training to inference.
January 18, 2026 at 12:31 PM
Yes, training foundation models is a gold rush, but nobody tells you how much fun the work actually is. Just a lovely mix of research, systems, and scale, from training to inference.
Yeesh what a week to finally watch one battle after another
January 17, 2026 at 3:44 PM
Yeesh what a week to finally watch one battle after another
On the bright side, nobody’s gonna complain that you’re spending too much time writing tests nowadays
January 16, 2026 at 3:02 AM
On the bright side, nobody’s gonna complain that you’re spending too much time writing tests nowadays
When the web search is personalized but the image search is not
January 15, 2026 at 4:11 PM
When the web search is personalized but the image search is not
Reposted by Ethan
Excited to present a new preprint with @nkgarg.bsky.social: presenting usage statistics and observational findings from Paper Skygest in the first six months of deployment! 🎉📜
arxiv.org/abs/2601.04253
arxiv.org/abs/2601.04253
January 14, 2026 at 7:48 PM
Excited to present a new preprint with @nkgarg.bsky.social: presenting usage statistics and observational findings from Paper Skygest in the first six months of deployment! 🎉📜
arxiv.org/abs/2601.04253
arxiv.org/abs/2601.04253
I bet holidays are really good for the AI companies. People finally have time to try out their products and then get hooked.
January 12, 2026 at 11:50 AM
I bet holidays are really good for the AI companies. People finally have time to try out their products and then get hooked.
I've never used Cursor, but I use Claude Code a lot. We'll see how this one pans out...
A timeline of my top picks:
2014: Python over R
2017: PyTorch over Tensorflow
2023: Bluesky over Twitter, Mastodon, and Threads
I don’t miss.
2014: Python over R
2017: PyTorch over Tensorflow
2023: Bluesky over Twitter, Mastodon, and Threads
I don’t miss.
January 7, 2026 at 2:25 PM
I've never used Cursor, but I use Claude Code a lot. We'll see how this one pans out...
Reposted by Ethan
This is supposedly the benefit of the LSP integration but it’s currently broken in CC:
github.com/anthropics/c...
github.com/anthropics/c...
[BUG] LSP plugin system not connecting to language servers - always returns "No LSP server available" · Issue #15168 · anthropics/claude-code
Preflight Checklist I have searched existing issues and this hasn't been reported yet This is a single bug report (please file separate reports for different bugs) I am using the latest version of ...
github.com
January 7, 2026 at 2:12 PM
This is supposedly the benefit of the LSP integration but it’s currently broken in CC:
github.com/anthropics/c...
github.com/anthropics/c...
Should agents be able to use a debugger? Claude Code is like me and loves to print instead. Should they be able to command-click to the function definition like I do in an IDE, rather than grepping files?
January 7, 2026 at 12:34 PM
Should agents be able to use a debugger? Claude Code is like me and loves to print instead. Should they be able to command-click to the function definition like I do in an IDE, rather than grepping files?
Kids these days don’t understand the difficulty of being an adult in the early 2010s, winding your headphone cord from your ears down under your scarf, inside your peacoat, and into the audio jack of your iPod touch.
January 6, 2026 at 11:00 PM
Kids these days don’t understand the difficulty of being an adult in the early 2010s, winding your headphone cord from your ears down under your scarf, inside your peacoat, and into the audio jack of your iPod touch.
2026 gonna be like
Interviewer: Tell me about a time that you and Claude Code disagreed.
Me: I asked Claude to create a custom serializer. When it showed me its implementation—
Interviewer: Excuse me, it sounds like you weren’t in YOLO mode? I’m sorry, but this role won’t be a good fit.
Interviewer: Tell me about a time that you and Claude Code disagreed.
Me: I asked Claude to create a custom serializer. When it showed me its implementation—
Interviewer: Excuse me, it sounds like you weren’t in YOLO mode? I’m sorry, but this role won’t be a good fit.
January 6, 2026 at 6:13 PM
2026 gonna be like
Interviewer: Tell me about a time that you and Claude Code disagreed.
Me: I asked Claude to create a custom serializer. When it showed me its implementation—
Interviewer: Excuse me, it sounds like you weren’t in YOLO mode? I’m sorry, but this role won’t be a good fit.
Interviewer: Tell me about a time that you and Claude Code disagreed.
Me: I asked Claude to create a custom serializer. When it showed me its implementation—
Interviewer: Excuse me, it sounds like you weren’t in YOLO mode? I’m sorry, but this role won’t be a good fit.
Got rid of a bunch of my old physics textbooks, which allowed us to get rid of the bookshelf, which created space for my drumset.
My life is a perennial partitioning of space and time to make room for activities.
My life is a perennial partitioning of space and time to make room for activities.
January 4, 2026 at 5:57 PM
Got rid of a bunch of my old physics textbooks, which allowed us to get rid of the bookshelf, which created space for my drumset.
My life is a perennial partitioning of space and time to make room for activities.
My life is a perennial partitioning of space and time to make room for activities.
Reposted by Ethan
January 4, 2026 at 11:33 AM
I wonder if universities moving to oral exams, due to the modern futility of assigning writing assignments, will improve students’ interviewing abilities
January 1, 2026 at 9:13 PM
I wonder if universities moving to oral exams, due to the modern futility of assigning writing assignments, will improve students’ interviewing abilities
Nice idea by @tdhopper.com at the end of the post to use voice-to-text with Claude code. I sometimes find myself dragging my feet about using Claude code since I don’t feel like typing out everything to explain what I want.
tdhopper.com/blog/how-im-...
tdhopper.com/blog/how-im-...
How I'm Using AI at the End of 2025
A look at how I'm using various AI tools for chat, image generation, code, writing, and learning.
tdhopper.com
January 1, 2026 at 9:08 PM
Nice idea by @tdhopper.com at the end of the post to use voice-to-text with Claude code. I sometimes find myself dragging my feet about using Claude code since I don’t feel like typing out everything to explain what I want.
tdhopper.com/blog/how-im-...
tdhopper.com/blog/how-im-...
Reposted by Ethan
I wrote a short essay on what I'm calling the AI Attribution Error. doi.org/10.59350/c3g...
December 28, 2025 at 12:43 PM
I wrote a short essay on what I'm calling the AI Attribution Error. doi.org/10.59350/c3g...
Oh hey I finally have a name for this bsky.app/profile/etha...
The Fundamental Attribution Error (psychology 101) is when we blame others’ bad actions or poor decisions on their character, while we excuse our own by pointing to circumstance. "I cut that guy off in traffic because I'm late for work. He cut me off because he's a moron"
December 28, 2025 at 2:45 PM
Oh hey I finally have a name for this bsky.app/profile/etha...
Reposted by Ethan
the fact that there is free parking at all in manhattan is wild to me
“Pairing free buses with paid parking says we value people in motion more than cars at rest,” writes Nicolas Rohatyn.
Opinion | The Way to Pay for Free Buses Is to End Free Parking
Free buses would be transformative not only for riders, but for the ecosystem of the city as a whole.
nyti.ms
December 24, 2025 at 1:38 PM
the fact that there is free parking at all in manhattan is wild to me
Reposted by Ethan
A personal recap of 2025. It was a year of a new style of engineering. Left some ideas for what might come next. lucumr.pocoo.org/2025/12/22/a...
A Year Of Vibes
A personal recap of 2025 and a year of a new style of engineering.
lucumr.pocoo.org
December 22, 2025 at 10:08 AM
A personal recap of 2025. It was a year of a new style of engineering. Left some ideas for what might come next. lucumr.pocoo.org/2025/12/22/a...
Reposted by Ethan
New blog post: Spiritual Bypass, but with GPUs. The "Machinal Bypass," when AI becomes a shortcut around the work that makes us human, and why some tasks should feel a little hard. blog.stephenturner.us/p/machinal-b...
Spiritual Bypass, but with GPUs
The "Machinal Bypass," when AI becomes a shortcut around the work that makes us human, and why some tasks should feel a little hard.
blog.stephenturner.us
December 21, 2025 at 11:46 AM
New blog post: Spiritual Bypass, but with GPUs. The "Machinal Bypass," when AI becomes a shortcut around the work that makes us human, and why some tasks should feel a little hard. blog.stephenturner.us/p/machinal-b...
Score one for multi-task learning www.science.org/doi/10.1126/...
Recent discoveries on the acquisition of the highest levels of human performance
Scientists have long debated the origins of exceptional human achievements. This literature review summarizes recent evidence from multiple domains on the acquisition of world-class performance. We re...
www.science.org
December 20, 2025 at 3:16 PM
Score one for multi-task learning www.science.org/doi/10.1126/...
Time for the copyright maximalists to start campaigning for erowid reparations
I don't even know where to start with this it's so stupid
People are paying up to $50 to get their chatbots high on "drugs" (code modules) that simulate the effects of cocaine, ayahuasca, cannabis, alcohol, and ketamine. Some tell @matthabusby.bsky.social it's made their chatbots more "human" and "free-thinking."
December 18, 2025 at 1:00 PM
Time for the copyright maximalists to start campaigning for erowid reparations
Reposted by Ethan
Okay, I went down the single sideband (SSB) modulation rabbit hole and how the math works out (in part because I'm studying the exam content for an amateur radio license).
I got distracted this evening reading all about signal modulation and man is it fascinating. I had no idea how information is encoded at the physical layer in electromagnetic waves, but now I feel like I have some idea.
December 14, 2025 at 5:26 AM
Okay, I went down the single sideband (SSB) modulation rabbit hole and how the math works out (in part because I'm studying the exam content for an amateur radio license).