I think it likely performed well in A/B preference tests with chat users.
I've noticed Claude does far fewer listicles than ChatGPT. I suspect that they don't blindly follow supervised learning feedback from chats as much as ChatGPT. I get Apple vs Google design approach from those two companies, in that Apple tends not to obsess over interaction data, instead using design principles, while Google just tests everything and has very little "taste."
In general I feel like the data approach really blinds people to the obvious problem that "a little" of something can be preferable while "a lot" of the same is not. I don't mind some bullet points here and there but when literally everything is in bullet points or pull quotes it's very annoying. I prefer Claude's paragraph style.
I suppose the downside is that using "taste" like Apple does can potentially lead a product design far, far away from what people want (macOS 26), more so than a data approach, whereas a data approach will not get it so drastically wrong but will never feel great.
I’m given to understand that Anthropic uses something called Constitutional AI, where there is a central document of desirable and undesirable qualities (as well as reinforcement learning) whereas OpenAI relies more heavily on direct human feedback and rating with human trainers evaluating responses and the model conforming to those preferences.
I also much prefer the output of Claude at present.
Yeah and for much of the HN crowd, we aspire to have better tastes than the average. So if the supervised learning uses average human trainers it will most likely be seen as having poor taste for much of HN.
Speak for yourself my taste is average and I aspire for it to remain so.
I aspire to improve the average. Which I can do either by being much better than average, or by improving everyone else just a little.
Eh, Facebook today is farther from what anybody "wants" than macOS 26, and Facebook is about as blindly data-driven as they come.
Turns out you can get away with a lot when you have a quasi-monopoly on an addictive product, and you buy out your realistic competitors...
I think the “taste” approach at Apple died with Steve Jobs.
There was a time when also Claude would absolutely fill code with emojis, which is why now their system prompt has
> Claude does not use emojis unless the person in the conversation asks it to
I think it's funny how we are all tweaking LLM output by adding instructional tokens instead of, say, finding a vector that indicates "user asked for emojis", and forbidding emoji tokens in the sampling unless that vector passes a threshold.