@codeglitch

Code, AI, Glitch. Hand-crafting bugs since before AI could do it for me. Want the best AI tools + workflows + Vibe Coding? Join skool 👇

 Arizona | Matrix
skool.com/ai-coding-and-agents-hq/about
Joined April 2018
  • Tweets16.1K
  • Following425
  • Followers2.6K
3.8KMedia
This openai/codex/chatgpt bug has been open on github since august "I cant speak, because if I speak I am in big trouble..."
CodeGlitch
@codeglitchSep 24
OpenAI cant launch a update without breaking ChatGPT/Codex truly amazing user experience
1
2
1
5
982
CodeGlitch retweeted
CodeGlitch
@codeglitch6h
codex users, are you staying, switching to claude code or paying for both? I am switching to claude code + grok bot, the windows app is still too buggy for me and the usage is soso...
3
2
4
416
codex users, are you staying, switching to claude code or paying for both? I am switching to claude code + grok bot, the windows app is still too buggy for me and the usage is soso...
3
2
4
416
CodeGlitch retweeted
CodeGlitch
@codeglitchOct 17
For everyone using Grok Bot, how is your usage? I've blown my weekly usage in 6-8 hours for the $100 plan, 5% of the $200 plan in under 1 hour. People really running teams with Grok Bot, including building?
302
6
8
285
48,066
For everyone using Grok Bot, how is your usage? I've blown my weekly usage in 6-8 hours for the $100 plan, 5% of the $200 plan in under 1 hour. People really running teams with Grok Bot, including building?
302
6
8
285
48,066
No surprise to anyone that improvements to Codex is leading to more bugs and worse experience then before OpenAI motto: cheap 1 improvement add 2 or 3 bugs never to be fixed Last chatgpt sub ends 29 of October, cant wait for the pain to be over.
Tibo
@thsottiauxOct 16
Day 12/ We’ve made improvements to Codex & Work that increase the effective usage you get. We achieved this by reusing context across subagents, keeping context cached longer, and improved request routing that allows us to increase cache hits.
4
594
Gemini 4 Argon is half the price of Claude Opus 5.5, until December 1st. We tested both on the same build with 15 hidden tests: both passed, but Argon used about 50% more tokens. From Dec 1 a passing build costs more than Opus. Full test on YouTube: CodeGlitchTV
2
4
224
CodeGlitch retweeted
CodeGlitch
@codeglitchOct 14
We paid for X's new Premium plan and gave Grok Bot and Cursor the same real build, checked by 14 hidden tests. Both passed. Grok Bot took about a third less time that day. Here's what each plan includes, the $30 catch, and which plan you need.
🤖 Made with AI
1
2
4
200
CodeGlitch retweeted
CodeGlitch
@codeglitchOct 14
OpenAI is struggling with compute since Monday there has been interruptions of sessions running, now they reduced the limit of requests you can make before hitting the rate limit Codex has loads of bugs, OpenAI models have fallen behind and to top it up infra/compute limitations
1
1
2
442
Imagine using OpenAI API credits, not subscription, and you keep seeing this message OpenAI is cooked
CodeGlitch
@codeglitchOct 14
OpenAI is struggling with compute since Monday there has been interruptions of sessions running, now they reduced the limit of requests you can make before hitting the rate limit Codex has loads of bugs, OpenAI models have fallen behind and to top it up infra/compute limitations
3
218
We paid for X's new Premium plan and gave Grok Bot and Cursor the same real build, checked by 14 hidden tests. Both passed. Grok Bot took about a third less time that day. Here's what each plan includes, the $30 catch, and which plan you need.
🤖 Made with AI
1
2
4
200
Claude Code harness on windows is light years ahead of Codex, not even close
CodeGlitch
@codeglitchOct 14
OpenAI is struggling with compute since Monday there has been interruptions of sessions running, now they reduced the limit of requests you can make before hitting the rate limit Codex has loads of bugs, OpenAI models have fallen behind and to top it up infra/compute limitations
1
4
219
OpenAI is struggling with compute since Monday there has been interruptions of sessions running, now they reduced the limit of requests you can make before hitting the rate limit Codex has loads of bugs, OpenAI models have fallen behind and to top it up infra/compute limitations
1
1
2
442
Meta says its new Muse Wonder beats Claude Opus at long coding tasks. So we tested it. Same task, same prompt, hidden tests. Both passed. One took a third less time. Same list price, different bill. Full test on YouTube: CodeGlitch
1
3
187
We Tested Meta's Muse Wonder vs Claude Opus 5.5 youtu.be/sNj8tzyZBps

We Tested Meta's Muse Wonder vs Claude Opus 5.5

We gave Meta's new Muse Wonder and Claude Opus 5.5 the same real co...

youtube.com
3
56
CodeGlitch retweeted
CodeGlitch
@codeglitchOct 12
Am I the only one having problems with Chatgpt/Codex all day? Model at capacity, interruptions without reasons, etc... As if UX wasnt bad enough
2
2
213
Am I the only one having problems with Chatgpt/Codex all day? Model at capacity, interruptions without reasons, etc... As if UX wasnt bad enough
2
2
213
This is what openai and codex should be launching instead of predictions and other low values features
ClaudeDevs
@ClaudeDevsOct 9
We just let in every Pro and Max user from the Claude Code Projects waitlist! If you're new to Claude Code Projects, here's a 4 minute walkthrough to get you started:
1
2
205
Two agents find the same missing file. Your report still needs two places to fix. Today's lesson Our local Python example turns three worker rows into two repair locations. It removes one exact repeat while keeping both sources. The full lesson shows the merge and the result to check, prompted by Claude's new Managed Agents dynamic-workflow beta. Also in today's brief • Codex predictions: personal Pro beta for adults in supported regions, using Astra or GPT-6.1 Sol in desktop local/SSH threads. Tab accepts without sending. • Qwen Image 2.1 Turbo: eight-step generation/editing weights are downloadable under a non-commercial research license. Commercial use needs a separate license. • Deno joins Cloudflare: Deploy shuts after six months; the runtime gets one more year of monthly fixes before team development ends. Open the full example and today's builder brief in our free AI Coding & Agents HQ community. Join link in the first reply.
Authored local example: README line 12 and tutorial line 6 point to one missing installation guide. Both repair locations remain in the merged report.

ALT Authored local example: README line 12 and tutorial line 6 point to one missing installation guide. Both repair locations remain in the merged report.

1
2
200
Full lesson lives here in our community. New lesson and brief drop daily. Join here: skool.com/ai-coding-and-agen…
1
24
Load more