Feeling like you're behind on AI? JJ Englert says we're all still early. Businesses are learning, workflows are changing, and there's still plenty of room to experiment, build, and figure out what works.
Delta is our daily driver, and it has dramatically improved how we work together.
We’re building Delta and DeltaDB so teams can move at the speed of thought while maintaining exceptional codebases, and we have many more features coming soon: delta.dev/roadmap. 
Git only records commits. DeltaDB, the version control system backing Delta, records every edit in between, along with the comments and conversation that shaped them. You and your team can view, inspect, and build on the same history — not only what changed, but how and why.
Last week, we turned off pull requests in the Delta repository, and moved all review into Delta threads.
Delta generates guided reviews of changes, and allows us to comment in context alongside the teammates and agents who produced the code.
Now, your team can do the same.
zed.dev/blog/delta-p... 
Delta is now in public beta. It’s a multiplayer environment for coding with agents and reviewing what they build, with every code change linked to the conversation that produced it.
Anyone can download it today on macOS, Linux, and Windows: delta.dev/download 
woooo! thank you so much for the work!
Make Zed feel like home.
Join New User Office Hours with @movraxrax.bsky.social to find your way around Zed and learn the basics of LSP configuration, settings, and keymaps.
September 22, 2026, 8:00-8:30 AM EDT
luma.com/zed-24ba 
It was a packed house last week for the SF Database Meetup.
Our CTO @nvw.io took the stage to break down sharding, why we built Neki, and why it's much more than just sharded Postgres.
Full talk on YouTube: One popular write-up described a recent AI lab incident as the work of "three consecutive secret AI civilizations."
@biilmann.blog thinks that framing lets the humans who ran the training off the hook. His new post makes the case for holding them accountable instead. biilmann.blog/articles/nat... 
In tldraw.com, stacking shapes more consistently preserves the relative positioning of shapes connected with arrows. Just because AI makes something easy to build doesn't mean it belongs on your roadmap. JJ Englert from Tenex shares how his team decides what to build by focusing on the work that strengthens their competitive advantage, not recreating tools that already exist.
Read docs to learn more: commandcode.ai/docs/mods Mods are more powerful than plugins.
Command Code can build mods and modify any behavior from programmatic hooks to a completely different inference provider.
We bet you haven’t experienced something as powerful before.
Build a mod now and make Command Code do anything you want. 
Last week these models went live in Command Code.
DeepSeek V4.1 Flash
Muse Spark 1.3 (with max reasoning)
Ling 3.0 Flash Sante (free model)
🐐 
fwiw charlie did dabble on a lightweight model trained locally and it's amazing: github.com/celoyd/potato mostly riffing on this underappreciated masterwork from @vruba.bsky.social 
few people are talking about how good llms are
Join us for the @changelog.com IRL at PlanetScale HQ.
We'll have three special guests joining us to discuss background agents, developer tools, and more.
Happening in San Francisco this Thursday, September 17th. RSVP:
Our own Tyler Benfield is speaking at Builder’s Day 🌈
After 15+ years across startups, consulting, and larger orgs, plus talks at Epic Web Conf, KCDC, and All Things Open.
Talk details coming soon.
Oct 26, SF ↓
builders.day/speakers/ty... S3 doesn't work the way you think it does. In this video, we uncover how S3 handles millions of requests per second on basic hard drives.
Sunday ship: built-in herdr support in Command Code
Zero setup. Shipped v1.54.0.
· working / idle / blocked in the sidebar
· works in cmd -p, not just the TUI
· releases the pane on exit, even on kill
· /herdr to check status
🐐 Extending $60 usage on DeepSeek V4.1 flash to 10 days. 🐐