LetItCode.dev

I subscribed to Codex and Claude simultaneously. Every time I open the IDE, the first thing I do is send "Hi" to both Codex and Claude code extensions, because the first message triggers the 5h limit start, so make it start earlier so I can get renewed earlier

I recently used Cloudflare Tunnel to connect my domain to my server, and it also works well with WSS. Previously, my setup was: Cloudflare CDN → server Nginx → service In that setup, I had to configure Nginx properly to support WSS. Cloudflare Tunnel is much more convenient

My product uses the Gemini API, but I often run into a 503 error with gemini-2.5-flash: “This model is currently experiencing high demand…”

Finally joined the App Store Small Business program

After making a plan, Claude Code usage for 5 hours increased from 3% to 35% using Sonnet 4.5 in high-effort mode. So, I asked Cursor Auto mode to implement the plan generated by Claude.

OK, ran out of all cursor tokens and need to wait another 26 days

Cron job auto-upgraded OpenClaw today, but the update broke due to a config validation issue. Lost connection completely. Had to fix it manually—removed agents.defaults.cliBackends from ~/.openclaw/openclaw.json, then restarted the gateway. Everything back to normal now.

I've enabled the 1.5x speed in the #Codex app, but still feeling very slow

Recently, i encountered this issue with the ChatGPT Mac app very frequently @ChatGPTapp

Naughty @antigravity 😂

I want to try, but: Claude Code Review is billed based on token usage. Each review averages $15-25 in cost: #pricing" target="_blank" rel="nofollow noopener">code.claude.com/docs/en/code-r…

letitcode.dev/t/openclaw-tel… youtube.com/watch?v=3o8Vbf…

I asked ChatGPT in the Mac app to generate markdown text from a 10-page Word document, but it stopped while generating. Then I asked it to generate a .md file, and it worked

i just used Claude Cowork to write an 11-page doc, 1800 words, and soon, hit the limit, 20 bucks subscription. i know, because it needs to do a lot of research, so it consumed lots of tokens

Far out, Gemini’s audio chatting feels better than any other AI. Because ChatGPT was the first to provide real-time audio chatting, I’ve always used it. I tried Gemini before version 3.0, but it didn’t feel that good. Today I tried it again, and I think it’s version 3.0

Apple charges a 30% commission by default unless you enrol in the Small Business Program: developer.apple.com/app-store/smal… After enrolling, the commission is reduced to 15%

I asked ChatGPT (auto), Claude (Sonnet 4.6), and Gemini (3.0 Fast) to translate 1,200 subtitle cues. Only Claude generated the full translation. Both ChatGPT and Gemini failed

i renewed the code: code.sygnal.com/code/shallow-s…

Tetris auto-playing gemini.google.com/share/ecd8e4d1…