Hassan

I love my GPT Pro sub, but the limit resets are a gaslight. Everyone acts like it’s a free refill lol. They’re just resetting the weekly timer. It’s like your boss saying you can have Friday off but you gotta come in to work on Sunday. You just hit the wall sooner next time.

I'm experiencing this issue on 20x Pro sub and it's been happening almost a week now. Essentially unusable. Additionally, for me, 8-10 times now, it goes up to 90-100 minutes of thinking and ends with "thinking failed" on trivial tasks. @OpenAIDevs @OpenAI @sama

@qiaohui @OpenAI @ChatGPTapp @sama Yup! I'm experiencing this same issue on Pro subscription 20x and it's been happening for almost a week now. Additionally, for me, 8-10 times now, it goes up to 90-100 minutes of thinking and ends with "thinking failed" on trivial tasks. @OpenAIDevs

Weird ChatGPT web issue with GPT-5.5 Pro / Extended Pro. Fresh chats + manual Pro + Extended briefly show "Pro thinking", then no persistent "Thought for ..."; output quality looks like a fallback/non-reasoning route. Thinking Heavy works normally. @OpenAI @ChatGPTapp @sama @nickaturley @barret_zoph Tried: new sessions, clearing chatgpt.com site data, relogin, disabling Fast answers. Also hit "Message delivery timed out. Please try again." My repro chats: chatgpt.com/c/69f63833-58f… chatgpt.com/c/69f63360-b4e… Similar reports: reddit.com/r/ChatGPTPro/c… reddit.com/r/ChatGPTPro/c… reddit.com/r/ChatGPTPro/c…

@GottliebShow Your miles just became inches

What happens to all my Spirit Airlines Miles?

I never flew Spirit Airlines. And now I never will. But I never wanted to anyways.

😂

Did someone break Claude?

Crowd finds it. Claude fixes it. You ship it. 📺 Demo: youtu.be/3z4H9OT274I 🔗 Repo: github.com/semswitch-inc/… #BuiltWithClaude #ClaudeCode

Devs upload their app. Users stress-test them, report bugs, and earn inference credits. Claude Managed Agents diagnose, write the fix, run the tests, and open a PR. Every successful run is a durable, mergeable artifact - not ephemeral output. 14+ real PRs during the build week.

Building with Claude Opus 4.7 this week as part of @claudeai, @claudedevs, and @cerebral_valley's Built with Opus 4.7 hackathon. I built CrowdPatch Community-sourced bugs. AI-shipped fixes. Powered by Claude Managed Agents. Every fix lands as a real GitHub PR.

@OmniC79283 Yup, 100% The "boring ops" stuff is really important. When that stuff has a tighter baseline and better foundation, it makes the more complex stuff go way smoother.

@TheHassanEffect Same direction here: cleaner refactors matter, but I still judge coding models by boring ops stuff: preserve repo context, recover from bad shell/browser state, and verify the diff without babysitting. xhigh+fast sounds promising there.

I also really love GPT 5.5 in Codex. xhigh+fast is just a killer combination. It's one shotted refactors for me and also is much better at writing cleaner code. I would say like 30-40% better at cleaner code.

I've grown quite fond of Claude Opus 4.7. It's actually quite a great and helpful model. I prefer it now over 4.6. Anyone who dogs on it is not using it correctly. You need to just let it do its thing. Of course, you'll correct things as you go, but it's great for long tasks.

@Terranatura4 @Ryan_Arthur_xyz @AlexFinn @grok @grok don’t want you to win lmao

@Ryan_Arthur_xyz @AlexFinn @grok @grok what model is he talking about and how can someone leverage it to create wealth? Give your top 10 ideas.

There is a Mythos level AI model dropping later today that will be available to every person in the world day 1 Do you understand what is about to happen? Democratized super intelligence Anyone can create any future they want

This whole subsidization thing, sure they lose some money, but they gain something worth a lot more. User data. Which they use to train new models and sell us the next thing we want since they already know what the demand is. That data is invaluable. And look what it accomplishes? For Anthropic specifically, Mythos-Preview was released and it’s not even available to the general public. I don’t think the prices are going up anytime soon.

Exactly how subsidized are tokens for Claude, Perplexity etc right now? Realistically if we're building workflows dependent on these that are running two $200/m max plans... does this change to... $1000? $5000? More? When?

Codex was acting dumber than usual today. This means that a new model release is imminent.

@testingcatalog Have had this on Gpt pro plan for a bit now it’s quite good

Anthropic might be preparing to release its Claude Code security feature beyond Enterprise and Team plans. Claude Security will conduct periodic scans of your repositories and report any vulnerabilities it discovers. Max only? 👀

Claude Design is pretty awesome. I am not a big fan of Opus 4.7 in Claude Code, but in this new design feature, it's top tier.

@AlexReiswich @bridgemindai 🧢

@bridgemindai I really would like to know what kind of prompts you are using. I have built a whole working SaaS within an 6 hour session and have like 25% usage. Opus 4.7 feels a lot faster and really intelligent, even UI is by worlds better than 4.6

Claude Opus 4.7 eats tokens for breakfast. 3 prompts. 13% usage. This is why I bought two Max plans. One $200/month subscription would be gone before lunch. The model is the best I've ever used. But you can feel that 35% token increase on every single prompt. Claude Opus 4.7 is not a model you run on a budget.

Opus 4.7 output is sometimes weirdly similar to Chat GPT. "you're right to push on this" "if you want, I can do..."