Vibe coding through the GPT-5 mess

Last week, we ran into an issue: no one on the show had ever vibe coded. So for today’s episode, we set out to fix that. Everyone came prepared with a vibe coded project — and some misadventures to go along with it.

This week on The Vergecast, we’re discussing the fallout of the messy GPT-5 launch and why the new model isn’t living up to the hype. We’re also exploring one of GPT-5’s big promises: vibe coding. OpenAI has put a lot of emphasis on the model’s improved coding abilities and even suggested that the average user can prompt their way to a helpful interactive experience. We put that to the test and quickly learned the tool isn’t quite ready for users as unprepared for coding as us.

After that, we dive into a busy week of corporate shenanigans, sizing up whether some of the biggest moves are serious business strategies, time-wasting stunts, or a little bit of both. On the docket: Perplexity trying to buy Chrome, Apple suing Apple Cinemas, Apple refusing to settle with Masimo, and Elon Musk threatening to sue Apple.

Finally, the Thunder Round returns for a lot of talk about smartwatches and the pitfalls of AI. Can you replace your phone with your watch? Can you trust what a chatbot says about itself? We have some strong feelings about it.

If you want to read more about the stories we discussed, check out the links below.