If you're curious about what it feels like to use @roc_lang or @zeddotdev, have I got a video for you! This is me live-coding a Roc CLI and web server in Zed, explaining as I go: piped.video/7R204VUlzGc?si=vT3s… Thanks to Func Prog Sweden for hosting the event and editing the video!

Direct link to course if you're interested: frontendmasters.com/courses/…

@rtfeldman's course is so good!

This is indeed a masterpiece of screen capture timing. 🤣

Got to come full circle with @nathansobo with helping launch Atom 1.0 way back in the day, and now bringing Auggie to @zeddotdev today. 10 years in the making.

Augment Code (@augmentcode) and Opencode (@opencode) are now available in Zed. 🎉 Both agents run natively through ACP. Install them in seconds from the extensions store and use them alongside Claude Code, Gemini CLI, or Codex.

1 click in @zeddotdev

In the typical (underspecified) explanation, you don't know whether the host revealed the non-prize by chance or by design. If it was by design, then switching improves your odds. If it was by chance, then switching neither harms nor helps your odds...so might as well switch!

As an aside, the classic Monty Hall Problem is supposed to be about probabilities, but if you allow game theory to be involved, you can arrive at the "always switch" conclusion anyway:

As (under-) specified, the correct answer is either "about ½" or "51.2% based on current global averages." "I have a kid, what are the odds that it's a boy?" "About 50-50." "FYI that kid has a brother born on a Tuesday. Now what's your answer?" "…same?" "MATH SAYS NO AHAHA" 🙄

I agree on both counts. "The Monty Hall Problem Problem" is that the typical explanation of "The Monty Hall Problem" omits critical information (specifically, that the host had no chance of opening the door with the prize). The OP this is responding to has the same problem.

AI tools get you 70% of the way there fast. Then you hit a wall. @addyosmani has been embedded in agentic engineering for years, watching what works and what falls apart in real teams. Thursday at 12pm ET, he's joining @rtfeldman to talk about what's actually working.

GPUI application spotlight. 🔦 Fast Forward: A window switcher for macOS!

Anyone know of a place that sells a shirt with the Ghostscript tiger on it? (I'm aware that "upload an image and they'll put it on a shirt" sites exist; I'm not looking for one of those!)

I have lost count of how many agents have adopted ACP since we announced it (checks watch) 2 months ago 😳

When I started working in python, I got lazy with “single assignment”, and I need to nudge myself about it. You should strive to never reassign or update a variable outside of true iterative calculations in loops. Having all the intermediate calculations still available is helpful in the debugger, and it avoids problems where you move a block of code and it silently uses a version of the variable that wasn’t what it originally had. In C/C++, making almost every variable const at initialization is good practice. I wish it was the default, and mutable was a keyword.

Do I know someone who's built ad servers, ad networks, or ad tech before and is in SF?

👀 Want to see how Zed is built? Join our first Office Hours - PR reviews with @conradirwin. Learn how we make decisions, what gets merged, and get your PRs reviewed live.

Claude Code has this feature called /compact which is a shortcut for printing "Error: Error during compaction: Error: Conversation too long."

>create Google Voice number >get no calls for 12+ months >give the number to Quickbooks when they ask me to verify my identity >immediately start getting multiple spam calls every day

We tried something a bit meta this time. Steve Klabnik wrote about how hard it is to have good conversations about AI, so we sat down with him to figure out how to do it better: smaller groups, clearer goals, and actually learning instead of just arguing. Link in 🧵