A-a-ron ⬣

Full-stack engineer | TypeScript, React, WordPress, AI workflows

[A-a-ron ⬣ @aaronolin]

@AItechcx @thsottiaux OMG! Please share this. I would adopt this for sure!

[A-a-ron ⬣ @aaronolin]

@Vivek4real_ Try the $understand-anything skill. It's even more amazing

[A-a-ron ⬣ @aaronolin]

@rezoundous Codex

[A-a-ron ⬣ @aaronolin]

@theo Internet in the United States 😞

[A-a-ron ⬣ @aaronolin]

@thsottiaux >bump< 😅

https://t.co/bcQ7sKBm5Y

[A-a-ron ⬣ @aaronolin]

@thsottiaux Please fix the Project ordering 😔 It's still working as it should in the iOS app, but not the desktop app. I am getting lost with the projects I was working on since they're not automatically sorted by last updated. I am loving the new app though! Great work!

[A-a-ron ⬣ @aaronolin]

@jquave @ReyaneRemache @gamejawnsinc @hrkrshnn I think the main issue is, Grok has the entire codebase, vs. Codex / Anthropic, they have bits and pieces of your project and it will only possibly send only the files you're working on.

[A-a-ron ⬣ @aaronolin]

@thsottiaux Also, I used Projects heavily in the old ChatGPT application, but now in the ChatGPT Work application, I can no longer see any of my ChatGPT Projects anymore, and all I see are just chats. There's a ChatGPT Classic app that I can use, but didn't know if Projects were going away?

[A-a-ron ⬣ @aaronolin]

@thsottiaux Please fix the Project ordering 😔 It's still working as it should in the iOS app, but not the desktop app. I am getting lost with the projects I was working on since they're not automatically sorted by last updated. I am loving the new app though! Great work!

[A-a-ron ⬣ @aaronolin]

@coinbureau Frontier providers also have to watch their pricing so that no one is forced to go to Open Source models 😉

[A-a-ron ⬣ @aaronolin]

@yozm_it Codex 5.5 Extra High, I am getting 768, ChatGPT Pro getting 960

[A-a-ron ⬣ @aaronolin]

@hxutixnnn @thsottiaux Go to a new chat session, and it's at the bottom. I was pulling my hair out trying to find where they stuck it

[A-a-ron ⬣ @aaronolin]

@thsottiaux @Dimillian This would obviously be an opt-in thing, and you can choose the min and max models Codex/GPT would infer for the actual prompt. It would be nice to have some kind of control over it in the settings.

[A-a-ron ⬣ @aaronolin]

I think we're at a point where #Codex needs a feature to infer the users prompt, and then automatically choose a reasoning level within the given model base. It could use a mini model to infer the prompt, and then change the model. I believe Warp has this same feature, and they categorize inference based on three groups. Cost Efficiency, Responsiveness, and Genius.

Sometimes when I am in the zone, I notice I am stuck on Extra High for a prompt that could have been Low reasoning 🤔

[A-a-ron ⬣ @aaronolin]

@thsottiaux @Dimillian 👀

[A-a-ron ⬣ @aaronolin]

@Anubhavhing @DonvitoAI Nah, "why sleep"

[A-a-ron ⬣ @aaronolin]

@wshxnv reeellleeeaaassseee.... 😅 This is awesome!

[A-a-ron ⬣ @aaronolin]

@danielrmay @0xnova @ClaudeDevs And you chose MOAA?! Amazing!

[A-a-ron ⬣ @aaronolin]

@ramneeksidhu01 @OpenAI My man!

[A-a-ron ⬣ @aaronolin]

@aidenybai This is awesome. Going to also set up nestjs doctor as well with my agent. Great stuff

[A-a-ron ⬣ @aaronolin]

@thsottiaux Official mobile app please!