Today we're announcing Coder Agent Relay: a new way to give Cursor Cloud Agents self-hosted execution environments on your infrastructure.
Launching with @SpaceXAI for regulated enterprises where code can't leave the network.
๐ https://t.co/I2IqK0tbP9
G2 named Coder a Leader across every segment in the Fall 2026 IDE Grid Report. One reviewer: templates run identical and reproducible for every developer and AI agent.
https://t.co/BZvO6UfFqM
What happens when AI gets really good at finding open source vulnerabilities, but there still aren't enough people to fix them?
6,000+ found. 97 fixed.
The latest [Dev]olution minisode digs into the bottleneck AI can't solve.
Watch now ๐
https://t.co/VpLXfiMFYU
45% of Commit Your Code talks are about production AI. One is horror stories from unsandboxed agents. We're a Gold sponsor, in Plano today. Come argue with us about where agents should actually be allowed to run.
New Coder whitepaper out today: AI risk doesn't grow as adoption matures, it changes shape, and it's highest right at the start.
Here's the infrastructure that closes the gap: https://t.co/yd1EZsYSu4
Coder Agents with kubectl on a live EKS cluster. It caught a CrashLoopBackOff, traced the root cause, fixed it, and converted the app to Helm. Then it tried to read kube-system and got a hard forbidden.
Check out @techie_piyush's demo https://t.co/RCYC5aNxwv
Rick Clark's prediction: by 2028, "we tested this" stops meaning anything to an auditor. The forcing function is regulation catching up, not the technology changing.
Full segment on [Dev]olution: https://t.co/OiXIAX2GZi
Most tech waves make your current job easier. This one changes what the job is.
Rob Whiteley on 25 years of platform shifts, and why AI is the first that rewrites the role itself.
https://t.co/OCcdwVWSE3
๐จ Claude can now do what got people fired.
Six months ago, sending company code to an AI tool was a fireable offense at a lot of regulated companies. Thatโs still the default with most AI coding tools today.
Coder Agents fixes it. Itโs a self-hosted AI coding agent, your code and workspace stay inside your own infrastructure. Claude plugs in as the model, and the only thing that ever leaves your walls is the inference call.
โ no API keys on developer laptops
โ every action tied to a named person
โ full audit trail
โ model-agnostic: swapping Claude for GPT or Gemini later is a config change, not a migration
For banks, hospitals, defense contractors โ, anywhere policy requires that โthe code doesnโt leave the building,โ this is the first setup that gets past that wall without the security risk.
Check here: https://t.co/FlGJmq7LPq
Thanks to Coder for sponsoring this post.
Seniority used to be measured in years of writing code. @rwhiteley0 on [Dev]olution: the variable that matters now is whether you're in the room while software development gets rewritten. https://t.co/OCcdwVWSE3
"I don't just want to code to code." Nnenna Ndukwe on why the value behind the work is the part engineering keeps skipping. New [Dev]olution episode: https://t.co/PF24Izn2Tp
74% of companies will deploy agentic AI within 2 years. 21% have governance for it. Gari Singh calls the gap "Ninja IT," agents executing faster than approval chains can track. https://t.co/cvrTptcdWF
98% more PRs merged, 91% longer review time. AI is moving the developer bottleneck from writing code to trusting it. Gene Kim on what that shift breaks first: https://t.co/ulJd2c6Ckk
Boot a complete dev platform from a USB stick. At our London meetup, Adรจle Devantoy demos The Box: Coder + k3s + Terraform packaged as a single NixOS appliance. https://t.co/IgkWVi9M6O
AI tool adoption โ AI maturity. On June 30, we're mapping the real stages of the AI journey for engineering teams.
Last call to register before the session tomorrow: https://t.co/CeVgHQ5FWh
518.7M pull requests merged in 2025, up 29%. Commit comments down 27%. AI is accelerating code output while review shrinks. The teams winning aren't writing the most code. They're redesigning what happens after.
Learn more on the Coder blog:
https://t.co/GCV9fmZ35H
Five months to patch. One day to weaponize. CVE remediation has a math problem, and AI-generated per-repo fixes aren't the answer.
https://t.co/vKhUlBlMWb
78% of enterprise devs use AI tools daily. 12% of orgs govern them. That's not an adoption problem, it's an infrastructure problem.
June 30, we're talking about it with engineering leaders who've lived both sides.
https://t.co/CeVgHQ5FWh
Engineering leaders are at wildly different points on the AI adoption curve. June 30 we're unpacking what separates the teams ahead from the ones stuck.
Register here: https://t.co/CeVgHQ5FWh