@3LandObserver@thsottiaux Thats true. Anthropic is really strict 🥲. I prefer OpenAI as a company, but right now, Claude models are just so much faster and better at design work. They also surprisingly have higher usage limits now compared to before.
@bemis_zac@jmaloneVS@thsottiaux True, but I didn't have plans switching to Claude until I got tired of the slow Codex speeds since they never brought ultrafast to lower tiers. Pretty much got tired of waiting (like most people) 🤧.
@jmaloneVS@thsottiaux I just tried out the $20/mo Codex plan out of curiosity and now it's all I use, even though I have 50k Codex tokens 💀. I quit anthropic because of usage limits but it looks like they're a lot better now.
@TokenGremlin Yeah I think at least if they told us their estimated release dates it'd be better. I don't think waiting for a surprise is the best approach for them anymore lol.
@thsottiaux Ultrafast for all plans would be nice. Also, being able to use credits without a subscription with any model (currently credits only let you use lower intelligence models).
@thsottiaux Also, I think that it's really nice being able to know exactly what's the feature/improvement focus of new models. For example, design and animation are sooooo amazing on Claude and I really think Codex users would benefit so much from it.
@thsottiaux I think in the past, holding off on public model release dates made sense for suspense and surprise, but I think being transparent with them now makes more sense since that'll help people decide if they should switch to Claude or wait a few days for 6.2
@thsottiaux Yes! I've been wanting this BUT with Ultrafast because of that flow state. For people who only get a chance to build not too frequently, I think it'd be great because the $20/mo plan isn't enough. I did notice recently the Claude $20/mo plan actually has more usage now.
@thsottiaux So I did end up getting a Claude subscription just to try out the crazy artsy video and design stuff I've seen. I tried it with Codex too and the differences were huge. Will the new model be better at design? I still love Codex for coding though – it does more automatically.
@OpenAI Can you guys bring Ultrafast to the lower plans 🙏🏼? To me, software development is an art, when the AI is as fast as my imagination, it feels truly immersive.
@OpenAI Also, are you guys planning to add motion/graphics into the next model at the same level that the current Claude Motion/Design can? Claude does make really pretty stuff, but their usage limits suck lol.
@thsottiaux I tried it out, and it is pretty fast. Nice work on that! I think speed is super important with this kind of work now. @thsottiaux Do you guys plan to bring Ultrafast to lower tiers 😗? I know it would use limits fast, but for me speed is super valuable.
@thsottiaux Hey @thsottiaux can you allow Ultrafast when using Codex credits 🙏🏼? I know a lot of people complain about not having enough usage but for me usage is basically infinite lol.
@mokiwara2y@thsottiaux I have a $200/mo plan and I literally never get below 80% usage. I don't have a software engineering job though, I just love to do side projects, so usage limits last me pretty much forever. I only upgraded to the $200 to get more Astra usage but 6.1 Sol is 🔥
@thsottiaux Thanks again for everything you've been doing Tibo. I still can't get over how negative people in the comments are 💀. These are amazing tools, and people just want to take everything they can for free.
@thsottiaux It's crazy how negative the comments are. Dude, I'm really happy how you guys are making all of these improvements. If people really need more than the $200 plan, then they most likely have enough income to support the $500/mo plan or use the API 💀.