OpenAI says it plans to end its partnership with Cursor following SpaceX’s acquisition of the company. Under the current proposal, Cursor’s direct access to OpenAI models would end on November 12. Cursor says OpenAI models represent about 5% of its traffic and that both companies are discussing a resolution, so the outcome is not necessarily settled. Even if an agreement is reached, the situation reveals a larger problem. Cursor became valuable partly because it functioned as a model-neutral workspace. Developers could choose the model behind their coding workflow without changing the interface where their projects lived. Once that workspace is owned by a company with competing models, neutrality becomes harder to maintain. AI companies increasingly want control over more of the stack: models, compute, interfaces, distribution, and the data generated through use. A product acquisition can change which models remain available, even when the end user has done nothing differently. The practical response is portability. Keep code in repositories you control, files in standard formats, and important instructions and memory organized so they can move between tools. Subscription value should also be judged by workflow, not just price. A $20 plan is valuable when its limits support a complete, repeatable process. If those limits interrupt the work, the price alone tells you very little. My earlier Skool post on AI subscription economics: https://www.skool.com/cliefnotes/codex-pluss-5-hour-cap-broke-the-workflow-that-made-it-worth-20 OpenAI and Cursor may still repair the relationship before November 12. That would reduce the immediate disruption, but it would not change the underlying lesson. The real question is not whether the partnership survives. It is whether your workflow can survive the partnership ending. A model can be replaced. A workflow you cannot move is a liability.

For a lot of builders, Codex Plus was never about unlimited use. It was a real sweet spot: map a build, implement it incrementally, test it, fix what breaks, and repeat. The weekly allowance worked because you could use it during the blocks of time when you actually had time to build. Bringing back the 5-hour window breaks that workflow. One intermediate task—with very few tools or agent calls—can drain a session and leave the work half-finished. The issue is not that usage has a cap; it’s that usage we already paid for is no longer available on our own schedule. The best description I’ve seen is that the problem is continuity. Real development is not a neat five-hour block: understand the repo, make related changes, run tests, debug failures, review the result. Being stopped before the final verification is where the value drops hardest, especially for people who build at night or in focused weekend sessions. The useful takeaway from the recent usage analysis is that model choice is now a budget decision. Use cheaper models for repo exploration, logs, and repetitive implementation; reserve the flagship model for architecture and final review. That can stretch usage, but it does not solve a workflow being cut off mid-build. I also don’t see the $100 tier as the obvious next step for every serious builder. There is a big gap between “I need one or two productive project sessions a week” and “I need a premium subscription every month.” The uncertainty around how limits may evolve on higher plans only makes that jump harder to justify. My answer, for now, is Cursor Pro+ at around $60 for broader model access and a more workable monthly coding pool, while keeping Codex Plus for focused basics and the workflows where Codex is genuinely the better environment. If a project truly needs more, I can move up for that period—but I do not want a five-hour meter deciding when I’m allowed to finish a build. OpenAI does not need to offer unlimited usage. But a weekly-only option, a larger burst allowance, or a way for an already-running task to reach a safe stopping point would better serve the builders who made Codex Plus compelling in the first place.

Grok Bot just became much more relevant for people building at a small scale. The official update says Grok Bot is now included with Cursor Pro, and Cursor Pro is $20/month. That is a very different adoption point from where this started. The original price was a real stumbling block for a lot of people. Not because the tool was uninteresting, but because $200/month made it feel like something only larger teams, technical operators, or people already getting clear ROI could justify. At $20/month, the question changes. Now a non-technical founder, solo operator, creator, freelancer, or small business owner can realistically test what it feels like to have a bot handle repeatable work, keep context, run routines, and operate more like a small AI teammate than a normal chatbot. My read: this makes Grok Bot much less of a "watch from the sidelines" tool and much more of a practical entry point for people who were curious but priced out. Official release: https://x.ai/news/grok-bot-more-plans Initial Grok Bot post: https://www.skool.com/cliefnotes/grok-bot-ai-staff-layer

Hello, can anyone let the administrator that I cant test the eduba alpha? @Jake Van Clief ?

This was the first community I joined, and my first real go at AI beyond just talking to a chat window. I found it pretty much by chance, lurked for a few months, and High Tea 6 was my first call. I didn't realise at the time how much that would shape things. The picture I got here was a healthy one. Build the thing, watch how it actually behaves, be honest about the bits that don't work. I've spoken to enough people since who came in through paid courses to know that isn't the norm. Nobody here sold me a shortcut, and I got pretty lucky landing here first rather than somewhere else. Something happened recently that made me take a step back. I met someone at an open networking event who runs corporate training and is writing a textbook alongside it. What she needs is her course turned into per-module walkthrough videos - her own slides on screen, voiceover over the top, one video per module. I built it as ICM, because that's just how I build now. Numbered stage folders, one agent reading the right files at the right moment, markdown carrying the context, and local scripts doing all the mechanical work that never needed a model in the first place. Her voice cloned from a module she'd already recorded, the deck walked and exported as frames, one script and one audio file per slide. The durations of those audio files become the cue sheet that stitches the whole thing together, so there's no separate timing step to get wrong. One module's through it end to end. 21 slides, 21 scripts, 21 audio files, one finished MP4. And there's around ten hours of modules to put through the same pipeline. As an 18-year-old that's a really good opportunity, and it's going to generate me some serious income. I wouldn't have built any of it without what I picked up in here first, so thank you to everyone who's answered a question of mine along the way. Nothing's finished yet. I've still got a lot to come and plenty I haven't cracked. It just felt worth stopping on rather than quietly getting on with the next thing.