@Adidotdev i see the complete opposite.
literally every problem we've spent past 2-3 decades solving have re-emerged, but in different shape. e.g. security, identity, permissions, scaling, privacy, virtualization, code distribution, orchestration, etc.
it's the perfect time to build!
reminder: if your user can run docker, your user is root.
microsandbox has no daemon and runs rootless. nothing on your host is sitting there doing root favors.
you get a hardware-isolated microVM, full stop.
native SSH support now in [email protected]. 🍻
previously, you'd run sshd inside the sandbox and expose a port. well, no more. point your existing ssh/sftp clients straight at it.
available in all SDKs and CLI today. 🪅
I understand the concern of skills atrophying when using agents. But so far I am not seeing it. Instead I have learned all sorts of dark secrets of linux networking I somehow didn’t learn before agents building a networking product.
early usage of opus 4.8, and it tells me: "hey we've been at this for 12 hours. let me add the task to the plan, and call it a night. if i continue, i'll likely introduce bugs due to fatigue".
lol, not sure what to think of this. 😂
your sandboxed agent calls OpenAI. its transcript, secret placeholders included, ships to S3 for archival. a prompt injection tries to exfil keys, or at minimum fingerprint which credentials exist. a release pipeline tries to send $NPM_TOKEN somewhere weird.
four scenarios, four different right answers. most sandbox providers stop at secret injection. that's 1 of 4. what about the other three? @microsandbox today gives you all 4.
i dig into all the details in this article.
fwiw, microsandbox supports bring your own kernel. definitely niche, but folks have asked for kernel features that didn't make sense to ship by default.
https://t.co/4463binbyf
we recommend basing your patches off our libkrunfw fork for full compatibility.