John Ferguson

@thdxr I finally got my app created and working for the most part (had to refactor to swift from Tauri for some reason, but it’s way easier this way) and I’ve got my server running on a launch agent, and this would just make that much cleaner

i'm thinking about adding an `opencode service install` command that sets up opencode server on your machine then running opencode or going to localhost:4096 would be connecting to a single instance and share resources would you use this?

@bcherny Have you found a way to handle the fact that verification can cause some huge memory usage in apps, like when running ‘npm test …’ or ‘mvn test…’ my computer gets thrashed on memory because it just builds up in the memory buffer from what I can tell

6/ Give Claude a way to verify its work Finally, make sure Claude has a way to verify its work. This has always been a way to 2-3x what you get out of Claude, and with 4.7 it's more important than ever. Verification looks different depending on the task. For backend work, make sure Claude knows how to start up your server/service to test it end to end; for frontend work, use the Claude Chromium extension to give Claude a way to control your browser; for desktop apps, use computer use. Personally, many of my prompts these days look like "Claude do blah blah /go". /go is a skill that has Claude 1. Test itself end to end using bash, browser, or computer use 2. Run the /simplify skill 3. Put up a PR For long running work, verification is important because that way when you come back to a task, you know the code works.

Dogfooding Opus 4.7 the last few weeks, I've been feeling incredibly productive. Sharing a few tips to get more out of 4.7 🧵

@bcherny Is adaptive thinking in 4.7 smarter than 4.6? I know there have been a lot of problems with it feeling “dumber” because of the way the thinking was “nerfed” recently

5/ Configure your effort level Opus 4.7 uses adaptive thinking instead of thinking budgets. To tune the model to think more/less, we recommend tuning effort. Use lower effort for faster responses and lower token usage. Use higher effort for the most intelligence and capability. Personally, I use xhigh effort for most tasks, and max effort for the hardest tasks. Max applies to just your current session; other effort levels are sticky and persist for your next session also. /effort to set your effort level.

@bcherny Is this ever coming for bitbucket or as a command to be able to run in other ci/cd pipelines?

New in Claude Code: Code Review. A team of agents runs a deep review on every PR. We built it for ourselves first. Code output per Anthropic engineer is up 200% this year and reviews were the bottleneck Personally, I’ve been using it for a few weeks and have found it catches many real bugs that I would not have noticed otherwise

@bcherny Is this available to turn on with a env var?

🎶 I've been using voice mode to write much of my CLI code this last week Can't wait to hear what you think.

@thdxr Are they still the weakest link with this much AI spend, or has it improved their performance?

if you manage a team how would you feel in a scenario where the weakest member is spending the most on LLMs like they're spending $2k a month or something

@bcherny But the JetBrains extension is so… unfinished and depressingly bad. Are there plans to update this at all? My team at work will be a lot more likely to adopt more thoroughly when that’s more polished (we’re unfortunately tied heavily to IntelliJ)

Did you know that Claude Code is more than a terminal CLI? Claude Code is also available in: - IDE extensions for every VSCode and Jetbrains-based IDE - Claude Desktop, web, iOS, and Android apps - Slack and Github Learn more: code.claude.com/docs

@hunterhammonds I would love to see what’s going on here!

I’m starting a community for cracked AI builders. Claude Code. Codex. Cursor. Conductor. It doesn’t matter. All I care about is that you’re building or you’re hungry to learn. We’re sharing workflows, skills, repos, plugins, etc. Want to join? Comment below and I’ll reach out.

@joelhooks Curious what kind of prompt you used to make the is happen?

forcing claude to review all of its fuck ups and create rules/skills to prevent them in the future

@zeeg Seriously, I sometimes jump back into planning mode to fix things because they have gone to crap

remember when the tweeters told you planning mode was useless lol that was like 2 weeks ago (i run planning mode on nearly every change now)

So @AnthropicAI (for work) and @opencode (for personal stuff) and anyone else i need the absolute Best best practices for CLAUDE.md and AGENTS.md files. I’m trying to consolidate everything into a single place and want to know where to start

@thdxr Nah fam, just look at the code in the pr 😅

we called it opencode code because you still gotta open the code and look at it

Seen discuss agent swarms, especially with @opencode, so I was wondering if you had any good resources for it that I could peruse?

@joelhooks I’m curious about how you are using swarms, and getting your agents to communicate properly. I’m trying to get to this point and have my spec driven workflow work in swarms more intentionally but I’m having trouble with it. You’re the first (and the most) that I’ve

@adamdotdev It’s finally mobile responsive! Yay! I can just use the frontend the way it is for the app I’m building! 😂

OpenCode desktop app is mobile responsive now can you guess why I finally did this?

@thdxr I doubt it because so many people gonna doing dumb shit, the only way it’ll slow down is to turn your computer off 😅

we're still a very small team - smaller than all of our competitors including the ones with 10x fewer users we're pretty overstimulated right now - so much flying at us 24/7 it's very fun but it also is a lot - kinda looking forward to the holidays to see if things slow down

@mitchellh As someone who just created a self hosted runner to try to publish an app to use opencode from mobile… this is very sad to me

I think in a moment when GitHub’s reputation amongst developers is at an all time low, making the policy change to charge for self hosted runners in any scenario instead of making other technical changes to make GH runners more attractive will go down as a very bad move.

@zeeg All I want to do is build cool shit, that does some good and helps people 😅 and some days it feels like I’m just under a a weight of not doing any of that. But some days I feel like I do decently. But I can never get away from wanting to be closer to my “own man”

need some of yall to stop doing startups and come help us build some things too many exciting things that we can bring to the masses, and not enough folks who want to be in the drivers seat

@thdxr Let’s build a new one

github is a very old app it has crazy wide scope, huge traffic, has to do some novel things its frontend architecture is all wrong lot of smart people trying to fix it even with a 5x smarter claude this isn't getting fixed much faster

@zeeg While I’m there I might “borrow” some whiskey though…

@f3rg_codes 🚨🚨🚨

need someone to go to my desktop and push some code for me so i can hack on it😅💀