@rcsaxe
Account based inUnited States
About this account
- Account based in
- United States
- Connected via
- United States App Store
Account-level information from X, not a live location or the device used for a specific post.
Director of Global Data & Analytics @abinbev | NYC Street Performer | MTG Cube Lover | Opinions are mine | He/Him
White Plains, NY
Joined March 2017
- Tweets23.9K
- Following571
- Followers3.3K
- Likes39.3K
Ryan Saxe retweeted
I often get emails from junior engineers asking me what to do.
Here's what I just wrote to one of them
All I want is to play with Rex 😭
I have a suspicion that I’m gonna end up deleting like 90% of my dotfiles because of it and I cannot wait.
Ryan Saxe retweeted
interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had
with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works
already-clean.goodfire.com/
Ryan Saxe retweeted
I just called Chungus (my OpenAI Dot) on the phone while picking up mexican food and asked it to transfer some Rex benchmarks from last night on my Mac Studio to my MacBook Pro via Tailscale and I just got to my desk and there they are. What a fucking time to be alive.
Ryan Saxe retweeted
I've published a new, general-purpose terminal specification that gives any program a way to tell the terminal what it's doing: idle, working, waiting, finished, or failed, and why. It is easy to implement on both sides. I've written more about why: mitchellh.com/writing/progra…
The problem this solves is generic and applies in compelling ways to all sorts of use cases (e.g. package managers, build tools), but it really rears a hideous head when it comes to AI and the "agentic inbox" problem.
I found over 250 different agent orchestrators (with various goals) that all individually implement a heuristic based approach to detecting whether things like Claude Code are working, blocked, done, etc.
If you look at @herdrdev's commit history, you can see around 10 compatibility changes to detect Claude Code within just the past 3 months. (This isn't a dig at Herdr! They're doing great work with what they have, and everyone else is doing the same thing).
Heuristics are not the way. And over-indexing on agentic use cases isn't the way, because this problem is really interesting for other tools too: Homebrew, Terraform, Cargo, etc. etc.
Proprietary protocols and out of band APIs are not the way. They have an O(N) integration problem and require extra work to work over SSH or within local VMs/containers.
The program status specification (OSC 7501) solves this in a reliable, deterministic way that is general-purpose and can be easily adopted by the entire industry.
Read more: mitchellh.com/writing/progra…
Ryan Saxe retweeted
Reject non-understanding.
Slop grenades are not just code and PRs. We run a real risk that *reading* gets entirely discounted, because of the exhaustion of getting low-quality and unverified AI prose being thrown at you all the time.
There's a thread going viral about a 'perf improvement' attributed to a compiler change. The PR *description* itself has the AI telling this person it's *not* due to the compiler change, it's how the algorithms and data structures have changed. And the implication is that a similar win could have been attained otherwise.
I want AI in the service of understanding the universe and enhancing human cognition and creativity.
Ryan Saxe retweeted
This is Knap. It's a new language I created that turns data into Markdown. The syntax should feel familiar and comes with wonderfully pleasant features to modify and format plain text.
Knap is open source. Over a million people already use Knap directly or indirectly because it started as the templating language I made for Obsidian Web Clipper. Now any tool can use it.
You can add Knap to your app or use it via CLI. Create a batch of Markdown files from JSON or CSV, or pipe Defuddle directly into Knap to generate Markdown from HTML/URLs.
The Knap site is color-coded to help beginners understand the relationship between variables (blue), filters (orange), and logic (green). I've tried to document it as comprehensively as I can.
Try the Knap Playground so you can see how fun the syntax is to use!
Knap is pronounced /knæp/ (with a hard k). It's named after knapping (with a silent k): the process of shaping stones to form arrowheads, scrapers, and other tools.
Enjoy!
This, and the comment thread, got me way more excited than I thought I’d be.
It’s gonna be hard to get me out of tmux (no herdr didn’t do it, I tried), but maybe this will do it.
Quick Superlogical demo to end the week. This demo is just of the basic functionality, but I want to highlight just how FAST it is. There's a WHOLE lot more functionality already present and coming but I wanted to keep the demo short. But seriously, check out that speed.
I'm showing the macOS app here and I get asked a lot so let me just say up front: its not macOS only. We're still figuring out exactly what platforms will be stable enough for the initial public release, but we're working on others. In particular, the web interface is very functional but I'll demo that another time (or someone else will).
This is also all showing off the architecture I previously talked about for terminal multiplexing which is significantly different from other mainstream multiplexers: this is all a custom binary protocol where the server is maintaining N replica distributed terminal state machines. It works!
Tons of hard work going into this by the team @almonk @dizzyup @pearkes. Hope we can bring you something you can use soon. ❤️ Happy to answer questions in the replies.
Honestly, I’ve found all of this true of the later models in the Opus 4 series as well.
Especially where giving examples constrains the system. I would argue not just constrains, but homogenizes. You end up with a similar recipe that’s force fit to your examples.
Ryan Saxe retweeted
Introducing the world's fastest tokenizer implementation, Gigatoken!
Gigatoken is ~500-1000x faster than HuggingFace, and ~100x faster than OpenAI's tiktoken for most tokenizer definitions on most machines.
These baselines are already multithreaded Rust implementations! 🧵
probably still going to keep my ssh from my phone into a tmux session, but maybe the apps are finally getting good enough I should take a look at the officially supported ones.
This looks like a pretty good dive into it for codex.
Ryan Saxe retweeted
i have seen enough proof now that using a coding agent is a deep skill
it's confusing because the people you see heavily using them produce horrible results
but that's because it's a skill! you can get better and the ceiling seems pretty high - this is very exciting to me
Ryan Saxe retweeted
i need someone at @OpenAI and @AnthropicAI to teach the models that while prototyping, backwards compatibility is just a bad idea
Ryan Saxe retweeted
"Algebrica" is a free and open mathematical knowledge base. All entries are progressively being released in Markdown format on GitHub for anyone who wants to study mathematics freely and openly.
Alongside the texts, the individual SVG illustrations are also made freely available. They are minimal, mathematically accurate, and designed to be easily reusable in notes, lecture material, or educational resources. Since they are vector-based and code-driven, they can also be modified or improved simply by editing the source.
Another step toward making the knowledge base more open, transparent, and genuinely useful over time.
I’ve been doing this since September and would strongly recommend.
In fact, this has become so much a part of my workflow that I explicitly have a design system as part of my “skill-builder” and often build skills towards web-app artifacts in which they are “zippable to share”.
Ryan Saxe retweeted
Also live today: ARC Prize 2026 - 3 tracks, $2,000,000 in prizes available!
Get involved:
• Play a Game: arcprize.org/tasks/ls20
• Build Agents: docs.arcprize.org
• Win Prizes: arcprize.org/competitions/20…
Ryan Saxe retweeted
I'm dying for pi-mono-style minimal library that handles the hard parts of email (auth, syncing with local state, etc.) and gives me an opinionated way to add agentic loops on top of that. I want to build my own agents and logic and guardrails, I don't trust vendors right now.
