Ben Werner

Mostly auth You can spin up 10 VMs that are forked from a development environment that has everything setup like an authed Claude code, an authed GitHub cli, and an authed gcloud cli. Since it’s a vm you can do long running back ground agents, or experiments, or 10 different rewrites stuff like that.

@benwerner What pain point does this address?

Can't see myself moving to managed agents. It's like Replit vs a local IDE: it's marginally more convenient, but not by 10x. Not even 2x. On my local, I can give agents access to any data, code or CLI I wish to. I don't even do it, I ask them to give themselves access. That was @openclaw's main innovation IMHO: sidestep the integration/access conversation by running things locally with very few guardrails.

@ibuildthecloud I recently flipped to agent in sandbox. Curious to see what you think of rigkit.dev

What do you all think, agent in sandbox or agent out of sandbox? Seems like agents want to be out of the sandbox. I don't like it, but I get it. But I don't like it.

Git worktrees are so clearly not the solution what are we doing

@amitpr nix is great - but you often need a bit more. Like auth, editor/ssh entrypoints, db branches we're launching rigkit.dev - should be interesting

Two months ago I wrote about Nix being the only sane way to manage a Linux system (or, thousands?). Feeling both more vindicated (Mythos + Shai-hulud) and doubtful (Nix can be painful!) today.

@damian_b Yea there is a reason employees get $2000 MacBooks. We are figuring all of this at @freestyle_dev As for auth we think we need something like rig.freestyle.sh to solve that weird problem

The sandboxing debate is becoming more widespread, but I fear it is focusing on the wrong things. Given too many people are ranking startup times, I decided to write this rebuttal. From Peter’s tweet, his postulation is the correct one: "How would we build software if tokens don’t matter?". In this future, sandboxing will not be about shaving milliseconds off boot. Timings will still matter, but there will be many more layers to this cake. Most critically, sandboxing will be about giving agents a high-fidelity operating environment. I think the winners will have to get these three things right: 1. Full-capability sandboxes Most sandboxes people are hyping are not actually capable enough for serious agentic work. A useful coding agent needs to clone arbitrary repos, install dependencies, run tests, spin up services, debug failures, and iterate. For example, many production repos depend on Docker. If your sandbox cannot run Docker, it is probably not ready for the workflows agents are about to own. 2. Trusted credential handling The moment agents do real work, they need access to real systems. GitHub, Linear, Slack, GMail, cloud accounts, internal dashboards, deployment tools. You cannot just hand the model your secrets and hope for the best. The right abstraction is some version of a trusted proxy: the agent can request actions, but secrets stay controlled, scoped, audited, and revocable. 3. Agent-to-agent communication Today, most sandboxed agents are isolated workers. I do not think that will last for long. The future probably looks more like fleets of specialized agents, each with different tools, permissions, memories, and objectives. Some review code. Some reproduce bugs. Some run benchmarks. Some test security. Some interact with external apps. Those agents will need to talk to each other, delegate, verify, and coordinate. So the sandbox stops being a box, and it starts to become more of a network substrate. That is why I think this market is still wide open. Nobody really knows what the agentic software stack will look like yet. But I am absolutely confident that the winning sandbox provider will not be the one that starts fastest. It will be the one that safely enables agents to do real work.

yknow back in the day using shadcn and tailwind was only for the cool kids

@ibuildthecloud @ifeanyi_we We’re launching rig.freestyle.sh that fixes everyone’s problems here

what is "set conditions". everyone runs CI in github actions or something similar, so it's fairly easy to reproduce your execution environment. But if your happy with a worktree, why would a worktree in a container be harder? It's just now you can do more stuff. And yes, it's the only way I can manage a crap ton of things in parallel. Worktrees is just chaos.

I hate worktrees so much. What is wrong with you all.

if claude is dumb but has taste, and gpt 5.5 is smart but with no taste, what is grok?

Giving a like to the bait articles which use photoshop over ai for the craft

ok guys why does claude keeping using "load-bearing"

@RhysSullivan @dillon_mulroy @kr0der Goated stack

@kr0der effect + tanstack start + cloudflare + drizzle has been nice

what tech stack do you guys use for side projects? i usually use nextjs/tailwind/trpc/prisma/vercel but i’m trying out tanstack start/drizzle/cloudflare

wake up babe, Claude shipped more dots

@ibuildthecloud We have just the solution: freestyle.sh/products/git

We really do just need a GitHub for agents. Not for humans. Humans need not apply. That would actually fix GitHub for humans.

Shocker, people are skeptical when someone named “Mr. Wonderful” tries to move in next door

Why is this article framed like this? Is this implying the random VC who threw in cash on an early round deserve it more than the guy who made the thing?

OH in SF: I love cats, they're like plants that move

We've been cooking - come drop by and hang out with us!

codex using for (;;) instead of while (true) to escape its RL is proof we're won't escape humanity's inevitable annihilation