egamirorrim

I don't really open the IDE any more

AI;DR

God I hope so, I'm rinsing my cc limits across two max accounts with ease atm

when it all gets too much add --cwd to the claude agents invocation to make it only show agents in the current directory.

I run about ten tabs, one for each project, with claude agents as a session manager.

Zero trust proxies are the enemy of easy automation, change my mind.

I'm constantly frustrated when trying to do something simple like use an API Key with services wrapped in things like Google's IAP.

Much easier with a VPN

What on earth is the use case for a docker container in a browser?

Isn't part of this the inevitable cache invalidation that comes from switching between providers (Opus to Gemini)?

So AI it was unreadable

Truly terrifying. But also shocking that a 'hacker' is using windows

Love this, but doesn't having to mouse with a touchpad do your head in?

I'm so disappointed that Anthropic are jerking us around so much with Fable. Really makes me feel like cattle they're unsubtly milking.

Arrrgh I bet I can't use it on Vertex

I'm the same. Every time I see people talking about explaining the project every session I'm puzzled. I've just never had to do that.

Russians are responsible for their leaders

How can I have any confidence in the security of your product?

It's extremely hard to convince myself to use a product for the huge variety of often sensitive agent tasks when it's not open source. I understand the business reasons for that, but it's unusual in this space at the moment.

Instead: Can you post any independent security assessments perhaps? Fundamental things like SOC2?

The UK gov needs to sod off with all this 1984 BS

Asking for a friend - what are these devices called?

I wish they'd release more stuff that didn't rely on me routing all my data through their cloud to work. Obviously the LLM is cloud based but I don't want any more lock-in than that. Plus not everyone has their repositories in GitHub.

You can get Claude Code to write tests too...

I mean if we're building an unrelated wishlist... Can 20x max users get auto mode already? Or can the enterprise plans get something equivalent to 20x max?

Given I'm running two max accounts to get the usage I want, can we get a 25x and 40x tier? :-)

This is bloody great Boris. Thank you.

I don't suppose you could share a little on that patching process?

So impressed by your eye, I didn't see this

Can I inject params to prompts?

...as long as you're in the US. The website doesn't do anything for rest of world and they can't get Alexa+.

This is really great, thanks so much for your work.

That's proper old

It's more like you want Claude to interact with X, and you go to see if there's an MCP server for it.

Claude could use the API directly but most MCP now comes with OAuth so you can let it act as you, in case API keys are hard to come by or chargeable. Sometimes with a good skill or a preconfigured CLI tool skills can be just as good if not far more powerful than an MCP server.

But the trigger you'd look for to decide to use an MCP is 'i wish Claude could access X'. My top examples:

- pulling designs from figma to implement them - fetching ticket context for a job from JIRA - getting a stack trace to investigate from Sentry

I've often thought about doing this, and it's probably easier when ever just with a tailscale exit node or something. What's always given me pause though is that this basically just an ec2 instance or other cloud instance that you route all your traffic through? Doesn't that mean it's just an non anonymous as your home IP if so?

I get so annoyed when Google launches things like this without Vertex support. Instant no-way from compliance and I'm left sitting on the sidelines.