🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Zed Wow, my mind was blown when Richard Feldman demonstrated this feature!!!! 🤯🙌🏾👏🏾 It was like watching your pairing partner code.

Nathan (Nathan Sobo) was recently on the Changelog podcast with hosts Adam Stacoviak and Jerod Santo to talk all about agentic editing in Zed. Check it out! youtube.com/watch?v=bcdVT6…

🔥

I'm not a scientific researcher, but this is why I avoid AI summaries. I like using AI tools to quickly get me draft outputs that I can verify/revise via other techniques (that don't hallucinate). But I can't verify a summary without reading the entire unsummarized version. 😛

Modular's second internal tech talk explores it's next generation graph compiler architecture, the layers of novel MLIR dialects, fancy fusion and dynamic shape support, focus on control and performance, plus how it all ties together with AI kernels written in Mojo🔥🤓

With Claude 4 out today, we had to add support for it quickly so we could give it a try. On Stable, we now support Claude 4 Opus and Claude 4 Sonnet via Anthropic API keys, and we also have Claude 4 Sonnet support through the Zed Pro plan. zed.dev/releases/stable

I haven't done web dev in long enough that I'm just a spectator here, but it's definitely funny timing to me that just yesterday I saw a comment about someone being perplexed at why people still like Elm despite its last compiler release being 2019.

Starting in 1 hour!

Starting now!

If you've tried AI agents for coding in the past and had a bad experience, e.g. because they keep getting stuck or going off the rails, I think trying Claude 4 Opus is worth your time. It's costly, but the failure modes are in a totally different (lower) category of frequency.

macOS's tty impl (in the kernel) has a 1024 byte internal buffer limit for cooked mode. After that, writes succeed but the data is silently dropped. iTerm2 works around this by sleeping ~16ms between writes. My day is ruined and my disappointment is immeasurable.

Episode 107: Language Design Deep Dive with José Valim! 🎉 I tried something different in this episode: we were both using Zed to write example code in realtime in shared buffers, Google Docs-style, as we talked about language design. It was really fun!

come over to talk to me and the team tomorrow! it's been awesome talking to folks from our community and learning from your experience using Zed! :)

Billion-dollar startup idea: it just calls a customer service number for you and keeps saying the word "representative" until either it gets you to a human or the call tree gets stuck in an infinite loop that bankrupts the company.

Huge milestone for the Zig team: the self-hosted (no LLVM) x86_64 backend is the default for debug builds! I'm an aarch64 user so I have to wait a bit more. :) But the amount of effort Zig puts into DX can't be overstated. ziglang.org/devlog/2025/?u…

Nathan Sobo is here to make the case for software craftsmanship in the era of vibes! "As software engineers, we should measure our contribution not in lines of code generated, but in reliable, well-designed systems that are easy to change and a pleasure to use."

This was a great conversation, and Jerod Santo asked fantastic questions! 😃

Going live with Zed tomorrow to talk about how I integrate AI into my daily workflow. We’re going to go through a real commit and will have live Q&A too!