Sign-In With GPT, Anthropic Vs. the Pope, Paramount Rebrands as Skydance | Diet TBPN
The hosts argue that OpenAI's sign-in with ChatGPT lets users bring their own compute and could reshape app economics, then cover AI consciousness fights, Tesla autopilot, Skydance, and domains.
Watch on YouTube ↗Charlie's summary
- ▶ 0:32
Sign-in with ChatGPT is a bring-your-own-compute model that makes inference-hungry indie apps viable
The hosts see this as the most relevant Dev Day feature for developers. Two years ago builders of magical products got stuck with “a $30,000 API bill.” Now users authorize an app like Notion to draw from their own ChatGPT allowance instead of the app's credits. Peter Yang says it makes a language learning app he wouldn't launch due to API costs viable. It only carries OpenAI tokens, not dollars, so printing costs stay with the app.
- ▶ 6:09
Consolidated billing could matter more than who owns the front door, but it may mean an app store dynamic
The hosts wonder whether this is a new economic structure with a margin for OpenAI, or just discovery. An inference provider might accept losing the billing relationship if demand grows 10x, like neoclouds on Nvidia's cloud. One point of billing is easier for consumers, as with the App Store's simple cancellation. Models still hesitate to handle API keys or credit cards in computer use, so a polished flow isn't there yet.
- ▶ 9:44
Platform integrations keep failing when both sides join reluctantly, as with GPTs and the Siri-ChatGPT handoff
Sam launched the GPT store at the first Dev Day in 2023, and it languished for three years. The Siri integration is the cautionary example: it asks Do you want to ask ChatGPT?, needs several clicks, and opens ChatGPT without any prior context. The hosts say both companies worked together begrudgingly. They wonder if a small team will now become the tool pulled from the shelf among 1.2 billion users and reach $100 million ARR in months.
- ▶ 13:18
Anthropic's push on model consciousness collided with the Pope, and Olah's stance is uncertainty, not conviction
A New York Times report says Chris Olah threatened to walk out of the Pope's AI encyclical launch after the Pope rejected machine consciousness, and that Anthropic lobbied his advisers. Critics call it AI psychosis. The hosts note Anthropic says there is a real possibility of risk, not certainty. They point out that Olah repeatedly says he doesn't have a firm position and that consciousness is hard to judge from “an end of one view.”
Give your email to unlock the rest of the key points — for every episode.
Do more with Charlie
Free — sign in with your email, no password.