co-founder, @coplane - the intelligence layer for the enterprise back office. previously: @Stripe, Google (BigQuery and Kaggle)
Austin and New York
- Tweets 10,189
- Following 1,794
- Followers 3,354
- Likes 35,518
Pinned Tweet
If you're excited about helping define what the UX and pattern language for a future of agentic and self-assembling enterprise software, please reach out!
We're looking for NYC-based design engineers to join our small, talent-dense team.
This is an opportunity to define the frontier of HCI. We're building the platform for self-assembling, self-optimizing software already in production with consequential enterprises.
Our eng team is made up of folks who have built and maintained large open-source projects like Neovim and Analytics.js and scaled massive data systems at companies like Google, Stripe, Wix, and Segment.
DM if you are interested or know someone who might be!
3
20
5,095
“because what does a beautiful thing tell you? Well it tells you the person who made it really cared… And so if you care about the infrastructure being holistically good, indexing on the superficial characteristics that you can actually observe is not an irrational thing to do.”
@patrickc
1
392
Style as Anti-slop
“because what does a beautiful thing tell you? Well it tells you the person who made it really cared… And so if you care about the infrastructure being holistically good, indexing on the superficial characteristics that you can actually observe is not an irrational thing to do.”
@patrickc
3
278
“The cost of ditching a library or framework and using a bespoke solution has gone down, but not disappeared entirely. They will continue to play a substantial, but slightly different role. Well-defined conventions and abstractions will help the LLM work faster, with fewer tokens and fewer errors.”
And reduce the decision burden on humans guiding the agents. Abstractions are valuable because they hide decisions from you, leaving only the ones you actually want to make.
Topcoat is pushing the boundary of server applications with Rust. Given the discussion falling out from @dhh's keynote, I thought I would write a bit about why I'm working on Topcoat and the cool features that were recently shipped: tokio.rs/blog/2026-09-24-top…
1
3
368
Victor Mota retweeted
Mark Carney is for me, right now, the most interesting man in the world. Gift link for his recent NYT interview: nytimes.com/2026/09/23/world…
17
33
369
27,103
My tinfoil hat theory is that they only rolled these fixes out now that they have text watermarks so they can continue to tell what was produced/distilled from Claude 😅
Opus 5.5 communicates more naturally, addressing some of the most common feedback we heard on Opus 5.
It puts the most important information up front and follows the writing rules you give it, which makes long sessions easier to follow.
1
193
Victor Mota retweeted
this is a very good premise
recommended reading. i belive this to be true at this point in time.
alexn.org/blog/2026/09/22/ai…
17
29
697
54,398
apparently @Cloudflare is building a durable stream API called "K2". the primitive matters and it's plain weird that it was missing so long. we'll keep making S2 more powerful for users, some exciting stuff in the works!
2
6
57
4,927
I’d fully agree with this if while sampling generated code I found the quality to be high - and it was purely a matter of not understanding what’s going on across the code base.
But without a lot of hand holding and deep understanding of the codebase I just don’t find models to produce good code. I’m hoping that will change!
It’s not just about not understanding, it’s that you can’t expect incrementally bloated and confused code to produce something globally coherent over a long horizon.
I think what people really mean when they say this is “I don’t understand as much of the codebase as a I used to.”
Well we already celebrated everyone becoming “managers of agents” earlier in the year, so time to learn how to be a manager.
You can’t oversee every little detail but still need to maintain quality control. There’s a whole art to this game. Welcome.
1
4
484
I'm just thankful Claude hasn't hyperfocused on "steel thread" as a technical term yet - it'd be really over for me
2
163
Jev (and fast/cheap generalized classifiers) can make AI qualitative linters actually useable!
Imagine defining free form rules to be checked at specific granularities
9
23
317
20,803
Jev-shaped:
a) classification
b) scoring
c) state machine or decision tree traversal (computer/browser use)
1
6
1,015
Finding the right constraints to set on a system is extremely powerful.
You think you're limiting the optionality, but you're really enabling optimization of the system in ways that enable many more things.
Jev limits the output to ~classification, and by doing that unlocks an architecture that can handle totally new use cases.
1
716
Finding the right constraints to set on a system is extremely powerful.
You think you're limiting the optionality, but you're really enabling optimization of the system in ways that enable many more things.
Jev limits the output to ~classification, and by doing that unlocks an architecture that can handle totally new use cases.
Jev is such an obvious and useful “type” of model. It’s genuinely shocking we were this far into the AI revolution before someone shipped a general purpose classifier.
I bet @CompleteSkeptic woke up every morning and scrolled X half expecting to be ape’d. Well done @typesafeai.
1
9
1,381
Instructor walked so Jev could run
Jev is such an obvious and useful “type” of model. It’s genuinely shocking we were this far into the AI revolution before someone shipped a general purpose classifier.
I bet @CompleteSkeptic woke up every morning and scrolled X half expecting to be ape’d. Well done @typesafeai.
2
312