Ryan Magoon

@mitchellh excited for hyperghostty

Excited to share that I've joined Vercel's Board of Directors. Vercel is made up of builders and tastemakers that continually ship things that deeply impact how developers work: Next.js, AI SDK, v0, etc. I can't think of a more exciting place to be. Let's fucking ship. ▲ My relationship with Vercel goes back to the earliest days. HashiCorp was an early adopter of NextJS and Vercel (~10 years ago!) and it remains my default tech stack and deployment platform to this day. Ghostty's website is all on Vercel, too! Beyond that, I've been continually impressed with the teams relentless focus on shipping meaningful software. And importantly, software that has incredible taste. Now we are in the age of agentic software development. Vercel is building agentic infrastructure that I think every app and agent will need (I certainly need it!) and I can't think of a more exciting place to be. Huge thanks to @rauchg , Jeanne, Marten, @cramforce, @tomocchino and the entire Vercel team for the warm welcome. Time to work.

@jamonholmgren Was just about to post this til I read the follow up lol

It’s like a postmortem. Use that feedback cycle and continuous improvement to get to a point where it is making the right decisions more often than not.

If your agent misbehaves and writes the wrong code, don’t tell it to fix the code. And don’t fix it yourself. Use that valuable context to figure out WHY it did the wrong thing. Tell it the issue, have it analyze its own context, and have it tell you what docs or skills or workflow is wrong, and what improvements would make it make the right decision next time. Then have it fix those issues. Be very diligent that it gets it right. Be prepared to hand tune this, because you can amortize the improvement over the rest of your project. Only after that, have it fix the original issue.

@zeeg I like comet browser and the mobile app still for most things I used to google

do people actually still use Perplexity? why? totally forgot about them until they stirred the mcp pot

@peer_rich I'm excited for the r3x literally because I just wanted a mk2 golf that is safer in a crash

ok german car industry hear me out reboot ALL iconic 1990s cars fully electric sell nostalgia to millenials who all have money now remove all expensive components, sell them for <$30k go into debt if needed and sell them at a loss

basic controls overlay

I was crushing with multiple parallel workspaces with claude code on claude desktop on mac, but as soon as I tried the same thing on windows it was choking hard getting lost constantly and not knowing what commands to run and how to orient itself in the filesystem. I assume it would be fine with WSL but I was working on my electron app which has to be developed against native windows windows sucks for dev and most other things

@cramforce @zenorocha Bullying works

The time proven technique of friendly bullying people on Slack

@DavidKPiano I love a good combinatorial explosion

For every n features you add, you're adding n² more ways things can go wrong

gamelord will be a place where people can nerd out over the history of the games in their rom collection, find games based on natural language/vibe + discover new rom hacks/homebrew, cloud saves, retro achievements, etc + really polished emulation experience and social features

@zeeg I'm like 2-3 mugs of specialty coffee a day

anyone else survive off a caffeine drip?

@kentcdodds @thdxr I think you have done the best job of anyone I know teaching people spaced repetition, project based learning, learning how to learn etc in the context of our field in a durable way

I've been struggling a lot with this recently. Everything has always been a gamble about the future when we're talking about investments and what you should be working on, but now we see that gamble play out so quickly that it makes it really hard to bet on anything. As an example, I feel like I've done a pretty good job in the past as an educator, making the right bets on what to teach people and giving people actually valuable knowledge by predicting what is going to be useful in the future so that their knowledge is still useful. With AI, that knowledge is only useful for a handful of months before something changes, and that puts me in a real pickle as somebody who's really just trying to help people build quality user experiences with durable and valuable skills that last more than a couple months. So now here I am trying to predict what is going to be actually valuable more than a few months into the future, so that when people buy my workshops and courses, they don't feel like they're only getting value from that learning for a couple months. Before, even if I guessed wrong, it still could help people for a couple months and was worth my investment because people would buy it hoping, like I did, that the bets were useful in the future. But now if I guess wrong, I put all this work into building something only to find out at launch time that it's not useful at all. In general, this is good because it means nobody buys something that's not immediately useful, but it's hard for me because now I'm not sure what to teach. Right now I'm experimenting with teaching what I think will be durable "problem crystalizing" and system design skills, but even still, who knows!? Maybe I should get out of teaching and just build boring but useful stuff...

everyone keeps trying to predict the future and then form an opinion that makes sense in that future this just does not work, you need to be right about the future, get the timing right, and then have the correct opinion on top and even if you are, none of that helps you today

@zeeg waiting for the sentry x fellow aiden and ode 2 with SSP burrs combo

Yall so greedy Mokapots don’t even make good espresso!

@evilrabbit_ One of my favorites. Grim is my 🐐 game

▲

windows support has been achieved internally

soon

Close. In sociotechnical systems, the tools and the culture are built (or destroyed) hand in hand. Like I said. A culture of caring without the tools of enablement will just burn through people's reservoirs of fucks to give. And tools without accountability loops are toothless.

big things happening

Discipline is not a genetic trait. You build it one day at a time. So don’t be jealous. Get to work.

one-shotted basic sound effects lol, need better ones

made the art more juicy