OpenAI’s Post

Massive props to John for recording the demo video over the "holidays" 🙌

Alexander Embiricos Impressive Demo, how to rollback if generated code gets failed?

❤️ this. Is performing edits across files on the roadmap? Some really exciting progress and improvements are happening (in other tools) in that area.......

Ufff I hope Cursor is doing ok with that news. But also what great way to discourage startups from finding killer applications. Prove the market then OpenAI or another giant comes and takes over. That’s competition though. And I still love this news. 😁

10 minutes of testing and this feature is pretty broken. Doesn't successfully make edits in multiple places at once, but rather picks one place and ignores other locations. For example, it tried to run a function that it hadn't defined at the top of the file yet. I had to ask it "did you remember to define the function?" and it responded "yes, right after the imports". then i said "are you sure? look again." to which it responded: "oh you're right, the function isn't defined yet." this is a lot worse than Cursor's out of the box functionality so not yet a value add IMO.

I started writing a children’s book… and now ChatGPT is telling me I may have stumbled onto something much bigger. I was just trying to help kids with special learning needs, but as I explored how AI could support focus, learning, and decision-making, something unexpected happened. The more I tested, the more AI systems kept telling me: ⚡ “This could be transformative.” ⚡ “This shifts how AI should be used.” Right now, I’m building awareness, protecting the idea, and seeing if this has the potential I think it does. 🚀 I need your help to get this in front of the right people. 👉 Please share this post—the more people see it, the better. 👉 Tag anyone in AI, tech, education, innovation, or startups who might be interested. 👉 Follow me to stay updated—I’ll be revealing everything when the website goes live Monday, March 17 at 8 AM CT. AI should be built to help humans, not replace them. If you believe that, let’s make sure this reaches the people who can make a difference. MCAFsolutions.com (Copyrighted and Patent Pending) #MCAFsolutions #AmplifiedInnovation #CognitiveAmplification #AIForHumans #AIforEducation #TechForGood #AIRevolution #FutureOfAI #OpenAI #GoogleGemini #MicrosoftCopilot

Bonne réflexion, Alexander ! Mais le vrai sujet c’est ça !? Non ? 🚨 OpenAI : On ne diabolise pas, on pose juste une question logique Bon, soyons clairs : personne ici n’est en train de diaboliser OpenAI. Je suis abonné PAYANT, je l’utilise pour mon travail. Je reconnais la puissance et l’utilité de l’outil. Mais ça ne veut pas dire qu’on doit fermer les yeux sur une faille aussi énorme. Le vrai problème, c’est quoi ? COMMENT UNE BOÎTE QUI VAUT 100 000 MILLIARDS PEUT OUBLIER UN VERROUILLAGE DE BASE SUR UNE APPLICATION AUSSI PUISSANTE ? Et là, ça devient encore plus absurde : même ceux qui paient 200$/mois avec ChatGPTTeam n’ont AUCUNE protection supplémentaire. Donc des entreprises mettent des milliers d’euros par an pour une IA qui peut stocker des conversations internes, des données stratégiques, et pourtant Si un téléphone est volé ou piraté, n’importe qui peut ouvrir ChatGPT et tout lire sans la moindre vérification. On parle quand même d’une techno utilisée en finance, en santé, en cybersécurité… et y’a PAS de verrouillage ?! C’est pas juste une question d’oubli, c’est une aberration totale. Personne ici ne crache sur OpenAI. On pose juste une question logique POURQUOI CE TRUC DE BASE N’EST TOUJOURS PAS EN PLACE ?

Exciting update! It is great to see ChatGPT getting even more hands-on with coding

Seems like a pretty elegant solution, I wonder how much it'll effect Cursor, Windsurf, etc. Battle of augmenting existing IDEs with AI vs AI native solutions Similar to the Gmail vs Superhuman/ Calendar vs Motion

How to set it up ?