


Software, film, and everything in between.


@harrysolovay built traits.sh which I’ve been using for a few days now and at first I was like “so just many md skill files littered around”? But it’s way better than that. It’s gotten me to a point this week where the AI just does what I expect it to do.

Stoked about Traits. traits.sh

Get ready for some delectable design patterns.
🔥 𝗝𝘂𝘀𝘁 𝟯 𝗱𝗮𝘆𝘀 𝘂𝗻𝘁𝗶𝗹 𝗘𝗳𝗳𝗲𝗰𝘁 𝗠𝗶𝗮𝗺𝗶 #𝟯! Really excited about this one! We’ve got three amazing speakers joining us on Monday, each bringing a different perspective on TypeScript, Effect, AI, and how we build software. 🎤 𝗝𝗼𝗵𝗻 𝗗𝗲 𝗚𝗼𝗲𝘀 (@jdegoes · @ZivergeTech) “Harnessing AI Models with Effect TS” 🎤 𝗛𝗮𝗿𝗿𝘆 𝗦𝗼𝗹𝗼𝘃𝗮𝘆 (@harrysolovay · @crosshatchdev) “Advanced Design Patterns with Effect” 🎤 𝗗𝗲𝘃𝗶𝗻 𝗝𝗮𝗺𝗲𝘀𝗼𝗻 (@devinjameson · @foldkit) “Foldkit: Frontend for Agents and Humans” If you’re into software engineering, curious about @EffectTS_, or just enjoy learning from people doing really interesting work, come hang out! We’ll have some great talks, food, drinks, and plenty of time to meet other developers and exchange ideas. And if you’re a developer, CTO, or engineering leader in Miami, I’d especially love to see you there! 📅 Monday, November 9 · 5:30 PM 📍 @thelabmiami · Wynwood 🤝 Organized by @SemantivAI 🎟️ luma.com/0yaas5h2 See you Monday! 🌴 #MiamiTech #TypeScript #EffectTS #AIEngineering #EffectMiami

We’re using Traits to develop an app that takes content creators' long-form videos and generates clips in their style, also using Traits. 300+ Traits guide AI codegen, like this bug fix to keep editors open across refreshes. We're starting to see unique/creative AI outputs!

I sometimes accidentally create traits for things that could be deterministic lints. We need a system by which traits can identify such cases and forgo decision model usage / fall back to static analysis. Also, shoutout to @MattiaManzati for his excellent work on effect/tsgo.

We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…

A formal Traits spec is underway. It will enable pluggable segmentation and enrichment. Segmentation: how to chunk source code before evaluation. Enrichment: how to gather dynamic context for evaluation (relevant symbol+file references, remote data, etc.). Stay tuned.

Traits is how we reclaim consistency in the age of AI codegen. I turned OpenClaw's AGENTS.md, docs, and skills into 545 Traits and ran them on 1.9% of the code, revealing 1,692 flagged violations, such as logging transcripts of users’ audio, a privacy violation.

An internal model at OpenAI has produced a claimed refutation of ETH! Russell Impagliazzo, Mohan Paturi, and I have been asked to help organize an open and independent examination of this claim. 1/4

It feels possible to me to check deterministically if a PR can be automerged or not. Given very, very strict constraints: - Only reorganises code - Does not change the module graph - Does not make test coverage worse - X other rules Feels like there is a tool here waiting to be created. Obviously doesn't work for new features. Might not even work for bugfixes. But it might enable a class of refactorings that help us all make changes more confidently.

Add a trait... ... and never make that category of correction again.
"Trait-driven development" I'm no longer really operating directly on code. Instead, I create traits which operate on the code + ensure agents never need the same feedback twice.

If we are to accept AI codegen as the future of software, then traits becomes the technique that separates the good engineers from the vibe coders
"Trait-driven development" I'm no longer really operating directly on code. Instead, I create traits which operate on the code + ensure agents never need the same feedback twice.

Step 1 in enabling infinite traits: Traits are now indexed into a local SQLite database. ~100× faster listings, ~4× faster lookups. Up next: embeddings-based dynamic trait activation. @CrosshatchDev

Hey, it is Wednesday OpenCode hacked eMudra had picked a private key value of 4, thus everything is now deemed unsafe. PHP, running 70% of the internet, has an insane hack... Happy getting internet getting humped day

If you're in Miami, make sure to join us at the Effect meetup!
🚀 Excited to welcome 𝗛𝗮𝗿𝗿𝘆 𝗦𝗼𝗹𝗼𝘃𝗮𝘆 (@harrysolovay) to 𝗘𝗳𝗳𝗲𝗰𝘁 𝗠𝗶𝗮𝗺𝗶 #𝟯! I first discovered Harry through his incredible hardcore TypeScript tutorials on YouTube. His depth of knowledge, creativity, and ability to push the boundaries of the type system immediately stood out. 🎥 youtube.com/@harrysolovay Harry is now building @crosshatchdev, an Effect-native toolkit for x402 payments, enabling applications and AI agents to seamlessly pay for and monetize APIs through programmable micropayments. Alongside projects like Liminal, his work continues to push the boundaries of what’s possible with TypeScript and Effect. It’s a real privilege to have him share his ideas with the Miami developer community! 📅 November 9, 2026 📍 The LAB Miami, Wynwood 🎟️ RSVP: luma.com/0yaas5h2 @EffectTS_ #EffectMiami #EffectTS #TypeScript #AI #FunctionalProgramming #Crosshatch