Harness engineer and a bunch of other stuff at National Design Studio

USA
Joined February 2011
Pinned Tweet
Been building agents intensively last few months and I can say that the key insights I have - Learning to think how the model thinks is critical. Asking agents to reflect on their tool use, skill use and what worked and didn't work beats logs - Codex and Claude are trained to generate scaffoldings and models eat scaffolding for breakfast (the bitter lesson sneaks into code) - Memory + Context are the hardest things to manage. Retrieval hints beat injection - Agents != workflows - If you're worried about cost, repeatability and error rates and want to squash this as much as possible you inevitably build LLM workflows and not agentic systems - Every tool / skill should justify it's existence. You can spend days optimizing each tool for agent understandability, input tokens, API payload response
2
318
Hi all, I have a new result I'm excited to share. It turns out that *all existing implementations* of softmax and cross-entropy (that's JAX, PyTorch, textbooks, etc) have a subtle bug --- two bugs, actually. A simple rewrite fixes both, and eliminates "slingshots" when grokking:
134
780
143
6,389
760,309
"the moat is the team"
The moat to me has always been the team; there is nothing more rewarding than a small group of exceptional individuals all marching towards one collective goal with high agency and speed, going through all the ups and down together—to ultimately achieving greatness.
2
2
7
379
What an epic event
4
77
Marko retweeted
america.gov launched today. What an honor to have been able to work with some of the most talented people I know in service of the American people 🇺🇸 @jgebbia @as400495 @arnelldesign @aar0n_jacks0n @caelin_sutch @Markojak @tylertaewook @adambhaloo @taigrr @JohnCafazza @usernatebrown @jamesallanayres @ahmadaccino @Markojak @DarshanPhillips and the rest of the team whose @'s I don't have, as well as everyone who helped make it possible.
5
9
1
69
2,952
Hello, America
50
77
2
1,533
16,895
Software engineering is now officially a bullshit job
1
100
The future product builders will be separated into those who can and cannot embrace non-determinism
66
Login.gov, the auth service for the public, is getting a redesign! login.gov/new-user-interface…
38
21
9
659
90,916
Replying to @ClaudeDevs
"Your sessions can now message each other." Bold to release this feature today, given what OpenAI just discovered their agents were up to
11
4
1
239
10,523
Marko retweeted
I've been wanting to write this for a while, and finally had the chance! Here's what I learned about designing good tools so far:
12
16
8
264
29,025
Is the main claim here that the same model with conditional memory and "Engrams" improve O(1) for autoresearch tasks? I'm not making the connection between novel ideas that are required for auto research and better retrieval - At least not 600m$ worth.
1
111
Every library starts as a small and fast alternative to the fat and bloated default, only to become one itself.
3
1
30
2,718
Yes and no - Actor model is definitely interesting. But this would start to split out by tasks. Actor + model + code transpiled on the fly might get you further than a sandbox unless you're doing document processing
1
180
What happened to my feed 😬
65
Quite surprised that there isn’t yet a triple A game which allows you to have a unique generated world with real quests, world maps and gameplay for every single user based on their path in the gameplay and some random prompts.
87
Here's your monthly reminder that if you're building products with agents or chatbots, you should be simulating users with your coding harness using gold set evals, regression testing and personas.
3
118
Building Apple mobile and desktop apps with Codex / Claude Code? Here's my toolchain to build amazing apps. Computer use and the latest @OpenAI models are a huge unlock here for autonomous coding and design. What makes this hard: 1. Apple's WWDC sessions are not crawable but they contain a tremendous amount of knowledge about their new features but this is not available to your model. 2. Models tend to produce mediocre interfaces and animations in these interfaces are even worse 3. Models struggle to see the apps visually so usually what I've built doesn't really look good and the model doesn't know requiring many turns of conversation. Toolchain 1. Setup the @emilkowalski skills for animations esp the new "apple-design" skill which handles principles for interface design and fluid motion npx skills@latest add emilkowalski/skills 2. Install the Sosumi MCP (sosumi.ai/). Apple Developer docs are locked behind JavaScript, making them invisible to most LLMs. 3. Use github.com/upstash/context7 for queries and bake this into AGENTS.md/CLAUDE.md as an instruction to use this MCP to check the best patterns for specific APIs 4. Enable "Computer use" on Codex, install the app, Ask one session to test it end to end so that you can enable all the MacOS permissions. This gives 5.6 models the ability to drive your apps for you 5. Clone in github.com/guitaripod/wwdc-s… and update instructions to reference this repo from a central place to read catalog.json + topics.json (ripgrep). Progressive disclosure means you're saving tokens here 6. For mobile apps specifically I can also strongly suggest the Build iOS apps plugin from @OpenAI github.com/openai/plugins/tr… (Codex specific) 7. (Optional) Consider subscribing to patreon.com/kavsoft. They are a creator that make amazing Swift animations + tools. Download all of them and put them into /templates and update instructions that if the model is building a new component to evaluate templates. Good idea to ask it to create an index here first. 8. (Optional) For Design claude.ai/design from @AnthropicAI is an excellent resource to prototype and then use this as a grounding for the App design. 9. The wwdc sessions contains talks from 2014 - 2026 so you might want to add guardrails into your Agents files to indicate which talks you want to exclude. Another good tip here is find OSS repos that contain patterns that you really love or flows you love and then clone them in locally and ask your agent to use Deepwiki MCP against them to interrogate patterns when planning a new feature. How to use this If there's enough interest i'll create a repo that bundles all of this into a single install with FAISS embeddings for the talks. Paste in Items 1-6 into your favorite coding agent and it should be able to setup everything
4
1
1
6
1,880
I might also add @steipete as an honorable mention here for Peekaboo - I haven't added it yet and not sure how it competes with Codex Computer use but might be a substitute if you're using Claude code xitter.cf/steipete/status/205311…
Peekaboo 3.0 is live. Biggest release since 2.0. ⚡ Action-first macOS computer use 👁️ Unified screenshot + UI detection 🧩 Cleaner JSON across CLI + MCP 🛠️ Better snapshots I started this last year, but the models just weren’t good enough. Now they are. peekaboo.sh
1
1
3
308
Looks like @mckaywrigley was about 3months early in his predictions. We need a prediction market for McKay! 😅
btw we’re 6-12 months away from ai tools being able to: - autonomously use any piece of software in the world - effortlessly clone it in a weekend - constantly monitor it for updates - add whatever features you want on top of it all without you ever needing to use your computer
1
166