Coding on vibes since day 1
Swayzee
Joined January 2013
- Tweets1.1K
- Following1K
- Followers318
- Likes3.1K
Pretty cool Mitchell is handing out free perf reviews for your macOS apps 😂
Wonder if you can hit this half bounce metric with electron or tauri
Big fan of Sketch but you have an unsettled frame at a full bounce (not ready, 1st image). You have unsettled loading states for ~1,600ms after a full bounce (2nd image). After that, you're fully settled (3rd image). One full bounce + ~2s of loading. Benchmark is based on fully settled interactive ready final frame.
Try pointing your agent at your AGENTS.md and skills to see if there is any rules or patterns for your project that could be lint rules instead. I've tended to put them there by default without considering if they could be a good fit for static analysis
Shout out to the legend @poteto for hooking it up with some Cursor credits! Time to see about using my Mac Mini as a self-hosted cloud agent runner for mobile development 🚀
Long running/autonomous agent sessions give me an RPG farming/training bot vibe. It hums along while you're away and you come back with excitement that maybe you'll have a nice little surprise.
Also you might just be banned
Having a single agent session manage processes is buns
Does anyone have a good system for multi-session process/server discovery and management for agents?
Claude Code, Claude Cowork, Codex CLI, Codex App, Cursor, Cursor CLI, Gemini CLI, Github Copilot, OpenCode, OpenClaw, Kimi Code, Replit, v0, Lovable, ChatGPT
If you're only using 4 of these your ngmi
I can't search Google in incognito mode anymore without having to solve a captcha. Not sure if it's my ISP or what. What's the best replacement for finding web pages?
I'd like a Claude Code tab like feature in @ChatGPTapp to cycle through a response length option.
Some explanations just need a paragraph but it insists on giving you a short story.
Does Claude Code really need to show off so hard when using Git?
"The --force-with-lease option was used for safety - it ensures the push only succeeds if no one else has pushed changes to the remote branch since your last fetch, preventing accidental overwrites of others work."
Dane Grant retweeted
If you’re a dev and not using @cursor_ai yet… this post by
@danecando might convince you.
Autocomplete, Ask, Agents, & tips for getting the most out of #AI in your editor.
Link in the comments!👇
ALT A promotional graphic from This Dot Labs featuring the title "Increasing Development Velocity with Cursor" in bold white text on a dark blue and pink diagonal-striped background. On the right side, there's a photo of a smiling man holding a "Dog Man: Mothering Heights" book. A purple label at the bottom reads "By Dane Grant."
