Joe Devon

I'm excited to announce that The White House will be doing a Global Accessibility Awareness Day (#GAAD) Event and Town Hall tomorrow at 10AM EST. Links to Livestream and Access Streamtext in the comments. Thanks @WhiteHouseOPE46! From blog post to The White House. 🤯

@kunchenguid So far I’m cool with how OpenAI did it. They aren’t punishing third party harnesses. They are being pretty open and just have a good harness you can use. It may change tomorrow but so far so good.

i'm strongly against model companies focusing too much on harness, but i would love to hear if anyone has a strong argument for it my reason against it: if openai didn't build GPT 5.5, no one else can. this is their core competence if openai didn't build codex cli and app, we have opencode and t3code. building harness is NOT their core competence this is not saying products like claude code, codex aren't good - i genuinely think these are top tier products built by really talented people my point is - the world might be a better place if model companies focus more on their core capability and give us better, faster, safer and cheaper models, rather than competing with the ecosystem in the application layer what do you think?

Great spending time with Global Accessibility Awareness Day co-founders @Jennison and @joedevon at Apple Park, as we celebrate the 15th annual #GAAD. Their work continues to motivate developers and teams at Apple to think deeply about technology, designed with everyone in mind.

Building accessibility into our products is central to everything we do. This year, we’re proud to deliver more intuitive accessibility features with natural language powered by Apple Intelligence. #GAAD apple.com/newsroom/2026/…

Not everyone arriving at Global Accessibility Awareness Day (GAAD) today has been here before — and that's a good thing. Accessibility.Day #GAAD #GlobalAccessibilityAwarenessDay #Accessibility #A11y #Inclusion #InclusiveDesign #DigitalAccessibility #DisabilityAwareness

@m13v_ @iSaurabhArora @ibuildthecloud @Imports @import If they're handling it differently than a CLAUDE.md straight text, then that sucks!

@joedevon @iSaurabhArora @ibuildthecloud @Imports @import fair on the single-source @import. cost was the bigger concern: a 4k-token AGENTS.md imported into CLAUDE.md still fires every turn, ~$8 of opus 4.7 per long session whether those lines get used or not.

I think I'm going to drop support for in my repo for claude code files. I never use claude code anymore and I just hate that I can't just use AGENTS.md .agents/ files.

@arjunaaqa fixed

@joedevon try claude, it’s great with slides

@m13v_ @iSaurabhArora @ibuildthecloud @Imports WDYM? You only @import AGENTS.md in the same dir into CLAUDE.md. Don't maintain CLAUDE. Just maintain AGENTS.md and there's no drift. It could be personal preference but I hate symlinks.

@joedevon @iSaurabhArora @ibuildthecloud @Imports @imports look clean until imported files drift independently. cost still adds up on every turn, and most teams forget the imports exist when auditing token spend. ccmd.dev/r/mn8du8mk written with s4lai

@iSaurabhArora @ibuildthecloud No need to symlink. CLAUDE.md supports @imports: #import-additional-files" target="_blank" rel="nofollow noopener">code.claude.com/docs/en/memory…

@ibuildthecloud Why not symlink? Agents.md and . agents/skills etc as source?

Fork your dependencies, trim them to only your use case, never update unless it breaks for your users. I’ve been vocal about this for 10+ years. I’ve always said that updating is way riskier than latent bugs (which can be tracked and CVEs monitored). If you are updating a dependency, it’s on you to analyze every single commit in the full transitive set of dependencies. If you dont see anything compelling, dont update! I remember at HashiCorp once in awhile an engineer would try to update a dep or replace a DIY lib with an external one and id always ask “show me the commit we need.” Dont update for the sake of it. Feeling pretty swell about this mentality with all the supply chain attacks happening.

The craziest aspect of @karpathy joining @AnthropicAI is that even if his only contribution was running their social media, he would have an outsized impact because of how influential his opinion is. The fact that he's a goated AI research OG is just icing on the cake.

Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.

This is going to be long. Last semester I suspected I had a major issue with use of AI in my survey courses, so I inserted what is known as a trojan horse (not the virus kind) into the directions of a paper assignment. As it turned out, I did in fact have a major problem, and a post on Threads about it accidentally went quasi-viral and ultimately became a Huffington Post article and an NPR interview. (Links at the end)

@Steve_Yegge @claychristensen Darn why didn't that tweet by Clayton expand. youtu.be/rHdS_4GsKmg?si…

1. Google attacked ID head on by making Google and Gmail free via an ad business. You can't be attacked from below when there is no cheaper price. 2. @claychristensen himself modified a few aspects of ID in his later years, talking about business model disruption. You can see him (link in next reply) warn Google about AI, saying that they aren't paying enough attention to the possibility of business model disruption, foreshadowing the emergence of OpenAI. 3. Interestingly @elonmusk and @claytonchris had an exchange on old Twitter in 2018 where Musk said disruption happened from above not below. Which is certainly Musk's playbook: x.com/claychristense…

Brendan Hopper, Matt Beane and I have a thesis, one that I've been sharing around lately, and we want CEOs and boards to hear it. Before I get to the thesis, let's revisit Clayton Christensen's Innovator's Dilemma (ID), the theory he developed at HBS to explain why big companies often get eaten by upstarts during technology shifts. In short, the ID says incumbents serve their best customers so well, and tune themselves so ruthlessly for doing exactly what they do today, that they can't chase the disruptor tech coming up from below until it's too late. The classic solution to the Innovator's Dilemma is to create a "bubble" in your company. You carve out an innovation team with a budget and mandate, as unfettered as practical by the parent organization. This is to combat the 2-level trap presented by the dilemma. The economic trap is Christensen's original point: a disruptive technology can't justify itself under your existing P&L, because it serves smaller or weirder customers at margins your real business would never accept. The governance trap is what gets piled on top once you're big: SOC2, FedRAMP, etc. mean every new idea has to clear a lot of process before it can move. The bubble is intended to escape both at once, with its own economics and permission slips. The standard innovation "bubble" solution famously doesn't work very well. You may solve the problem inside your bubble, but you often can't roll it out to the rest of your company for the original reasons. Everyone is focused on doing their current stuff, and nobody has time for a major change. Our thesis is that there is an entirely different way out of the dilemma this time around. No bubble needed, as long as you follow a simple rule. That rule is, let your people play. Give them back any time they earn from automating their jobs with AI. Then incentivize them to use that time to improve the company's processes. When you see an engineering team announce a 40% productivity boost from adopting AI — a number that's been showing up in plenty of LinkedIn posts lately — your first reaction as a CEO or manager is probably to say, that's awesome, we can do more work now! Or you might simply expect to see 40% more output from the team. Either way, you have just asked them to spend their extra time building faster horses (your current business) instead of letting them go figure out what a car would look like for your company. They gained some productivity from AI, which could have been your ticket out of the Dilemma, and you immediately slurped it back for your existing business. This will get your company killed in the medium to long haul, because your company tomorrow will look almost nothing like it does today. Conway's Law says your software and your org chart mirror each other; as AI rewrites how you build software, the org has to shift to match. But if you're stealing the hours back saved by your employees, then you're not letting your org pivot naturally in the direction it needs to shift. @RealGeneKim and I saw this in person at @arkanalabs a few weeks back. As long as your people know they'll be recognized and rewarded if they improve the company's processes — public credit for cross-team workflow wins, promotion criteria that actually count process improvements, managers who treat freed-up hours as a feature rather than a budget line — then they will use their "play time" to seek out other teams, and start pivoting you to becoming AI-native. This way it can unfold in whatever bespoke way is most natural to your company, rather than in some ivory-tower research bubble. For every company, the way it unfolds will be a bit different. I think of this approach, of giving the time back to the humans who automate parts of their jobs with AI, as the new solution to the Innovator's Dilemma. The old bubble solution was to separate a bunch of people from their regular jobs, and try to give them the freedom to solve the problem in isolation. In contrast, by giving your regular employees their hours back, the innovation bubble is still there, but it's now dispersed across the company, as lots of very tiny bubbles: one bubble per person who has liberated some hours. If you've ever read Slack by DeMarco and Lister, a great book from back in the 90s, then our thesis should resonate. What companies need is to empower their own employees, the ones who actually work together (even across departments)--the ones who know how the business works--to shift the company in the new directions together. Gradually, but with intentionality. You still have the frankly awful problem of token budgets. For every employee you upskill into baseline AI literacy (which I'd define loosely as using coding agents throughout the workday), you've added a non-trivial opex spend — for the heaviest agentic users it can run into five figures a year. I won't sugar-coat it; you need to find that money somehow. I don't have a magic solution, but I'm very happy that other models are catching up to Claude, because they're becoming good enough for real work now. But token budgets alone aren't enough. To live through the Innovator's Dilemma this time around, your employees need a time budget, too. Give it to the ones who earn it using AI, then incentivize them properly, and I think you're headed in roughly the right direction. Thank you for coming to my TED tweet.

I am particularly proud of this additional feature in the latest build: I added features and guidance to VoiceClaw so that people with limited control of their hands can open and quit the app using iOS Accessibility “Voice Control” custom commands; turn on the in-app mic either by using a Voice Control “mic on” command that taps the screen or by using the app’s optional autostart-the-mic-upon-app-open feature; turn off the in-app mic using an in-app “mute the mic” command, and end an in-app session using an “end the session” voice command. Now a person can use VoiceClaw — and more importantly, OpenClaw — even if they can’t use their hands.

Making the epub into an audiobook did cost $10 bucks, but is otherwise not available as an audiobook. I did try a reader to speak out the book, but it was really robotic so now I have OpenAI's Cedar voice to accompany me tomorrow.

We live in crazy times. AI has gotten good enough to let you knock out multiple projects in a day that would take forever otherwise. Mind you, this is my output on a relaxing Saturday spent mostly out and about. In my spare time at home: I spent an hour using AI to write a POC from scratch for a multinational that will let them in-house a pain in the ass activity that would have cost them 6 figures on a SaaS. The output is picture perfect. Then I upgraded my setup described in this post x.com/joedevon/statu… to add a fan to the Raspberry Pi. I'm going to be overseas and wanted to make sure it doesn't overheat. Might not sound like that big a deal, but AI saved me countless hours debugging stupid stuff. Then, since I'm going to India tomorrow, I wanted an ePub book I bought to actually be an audiobook I could listen to on the way. So literally while watching MMA on Netflix, I got Codex to try out 4 different voices and now it's generating an audiobook. It takes me longer to write a tweet like this than to vibe code some of the projects I work on.

@jxnlco @trq212 Darn I missed another meme. Everyone in my feed is copying each other uncredited. What did I miss?

@trq212 ahahahah

When do you reach for other models instead of Codex? What can we do better? Hit me with all of your frustrations. dms open. If you can give me detail (e.g. specifics/transcipts) - it'll help a lot in finding out exactly what we need to do to improve the next model

Finally installed it. Wasn't sure until your post. Love it already! My new default terminal. Vertical tabs is such a great addition. But having a Browser inline is so convenient. And I've only been using it for about an hour. I bet there's some browser plugins for ghostty and maybe a vertical plugin, but this is so nice right out the box.

i love ghostty terminal and have been using it for a few years now.. but i need vertical tabs. started using cmux (built on libghostty) cmux.com still use ghostty for one-off things but any tabs i'm going to keep around for a long time i prefer in cmux now

Oh wow McGregor vs Holloway 2. July 11. #UFC329