98 mentions · 85 chunks · 55 episodes
The reflections of an OpenAI engineer say that at OpenAI "Chat runs really deep." "Since ChatGPT took off, a lot of the codebase is structured around the idea of chat messages an...
... only player that could plausibly have the juice to do it in the last decade is OpenAI. OpenAI will try to redefine the browser category. Just blow it up. Make the web page rendering be the secondary function, not the primary.
On the path to aggregating everything, OpenAI will at some point allow users to choose competitor's models in the UI. They'll say "you can use any model! Look, it's open!" But no one changes the ...
With OpenAI and Anthropic the model is the product. The product is a model in the middle like a christmas tree decorated with various doo dads. The doo dads are ...
... never ship a toggle to "make me more bored." This is probably the logic behind OpenAI refusing to make the memory feature user controllable. Also, they think they can get away with it without too much uproar… … and they're probably rig...
What's the chance that OpenAI winds down the Inference API business?[ix] They could try to switch to an approach where the only way 3P's build on their platform is mainly by exten...
OpenAI going for vertical integration is like cell phone carriers[jj][jk] trying to control what software you can install on devices connected to their netw...
OpenAI's[jz] implementation of MCP in ChatGPT is limited. They only allow a subset of allow-listed MCP instances for certain use cases. This will quickly ev...
... about how ChatGPT now must retain all logs due to a judge's ruling. Imagine if OpenAI had been in a position where they simply didn't have the logs. But aggregators must have logs, it's their business imperative.
When you ask OpenAI what embarrassing things it knows about you, if it doesn't have anything it says "I don't know anything embarrassing about you, but you should feel f...
...internet bill for our home. Many people have a subscription to a walled garden (OpenAI) to get access to LLMs. If you're going to have a subscription to get access to LLMs, why not pick the option that is the open ecosystem, that includ...
...ith your intentions becomes of critical importance. According to a leaked memo, OpenAI is trying to build the super-assistant.
...to me. You could imagine an alternate universe where ChatGPT got popular before OpenAI had released a public completion API. In that world, OpenAI would likely reserve their model for their own 1P product. Other leading models from Anth...
With the way OpenAI is building out their product team, they sure aren't acting like they think they'll get AGI soon.[mm]
...lives. The top 4 chatbots today are led by people who have been Facebook execs. OpenAI is speedrunning the engagement maxing playbook. Experts in the playbook have pointed this out. The engagement maxing playbook was a net negative for ...
...aturally bias towards sycophantic answers. People have started to call out that OpenAI is falling into the engagement maximizing trap: "You have a problem you haven't addressed or answered. Sycophantic model behavior is good for your bu...
...n't do that." A chilling moment. Surely it's just a coincidence that the CPO of OpenAI is an exec from the all-time champ of engagement maximizing, Meta. OpenAI will become an even more intense version of Facebook. The honed engagement-...
Rohit got locked out of his OpenAI account. Imagine how bad that would be if ChatGPT were the operating system for your life.
Prompt injection sets the ceiling of potential of LLMs. Claude and OpenAI will build integrations into chat via things like MCP. Vibe coders will get stuck making dead end little island apps. Both will get stuck on the priv...
...t your assistant, it's Amazon's! Similarly, ChatGPT is not your assistant, it's OpenAI's. When Claude makes an artifact, it feels like it made something for you. You didn't make it, Claude did. Software that says "you" in the interface ...