"Introducing" without actually just linking the search page. That's not an introduction, that's a teaser trailer. If they want this to be a viable search it needs to be available quickly, and anonymously from something quick to type in. Google would have been annoying as shit if you had to go to google.com/search , let alone then log in.
ChatGPT Search
31–40 of 1001 posts
Re: ChatGPT Search
#32"Introducing" without actually just linking the search page. That's not an introduction, that's a teaser trailer. If they want this to be a viable search it needs to be available quickly, and anonymously from something quick to type in. Google would have been annoying as shit if you had to go to google.com/search , let alone then log in.
Re: ChatGPT Search
#33AKA Bing Search in ChatGPT. So it is not it's own search engine and is still using Bing for its results just like the rest of them.
If it's better than Bing or Google in presenting the relevant result in a condensed way, it's still a win for the users.
This doesn't matter if the results are user-hostile, as both search engines are.
Re: ChatGPT Search
#34How long until advertisements are subtly introduced? I didn’t notice any specific brand of limoncello recommended in their demo.
Re: ChatGPT Search
#35Re: ChatGPT Search
#36Hah, OpenAI is becoming an ads business too. So much for something new, same old funding model for every centralized platform on the web.
OpenAI has ads? I thought it was mostly a freemium business model.
There aren't any ads in their demo, we haven't seen the real deal yet, but I'll be watching HN for that day.
Re: ChatGPT Search
#37Asked it to generate code for a library that was released in the past year - GPT-4 couldn't do it and this one just did it flawlessly. I am super impressed.
Re: ChatGPT Search
#38Re: ChatGPT Search
#39I have turned bearish on Perplexity recently, this confirms it
Perplexity is a really terrible name for a product and that alone will hold it back from being a real competitor.
Re: ChatGPT Search
#40The only other way to kill the web without killing LLMs in the process would be to create a way for people to upload structured public content directly into an LLM’s training. That would delay public content into release batches unless training can be sped up significantly.