I assume there's no reason to use GPT-4-turbo for API calls, as this one is supposedly better and 2x cheaper.
GPT-4o
481–490 of 1001 posts
Re: GPT-4o
#482Ohhhhhhhh, boy... Listening to all that emotional vocal inflection and feedback... There are going to be at least 10 million lonely guys with new AI girlfriends. "She's not real. But, she interested in everything I say and excited about everything I care about" is enough of a sales pitch for a lot of people.
The movie “Her” immediately kept flashing in my mind. The way the voice laughs at your jokes and such… oh boy
Re: GPT-4o
#483Impressed by the model so far. As far as independent testing goes, it is topping our leaderboard for chess puzzle solving by a wide margin now: https://github.com/kagisearch/llm-chess-puzzles?tab=readme-o...
Wow from adjusted ELO of 1144 to 1790, that's a huge leap. I wonder if they are giving it access to a 'scratch pad'
Re: GPT-4o
#484I added gpt-4o support to my LLM CLI tool: pipx install llm llm keys set openai # Paste API key here llm -m 4o "Fascinate me" Or if you already have LLM installed: llm install --upgrade llm You can install an older version from Homebrew and then upgrade it like that too: brew install llm llm install --upgrade llm Release notes for the new version here: https://llm.datasette.io/en/stable/changelog.html#v0-14
Whenever I upgrade llm with brew, I usually lose all my external plugins. Should I move it to pipx?
Re: GPT-4o
#485The most impressive part is that the voice uses the right feelings and tonal language during the presentation. I'm not sure how much of that was that they had tested this over and over, but it is really hard to get that right so if they didn't fake it in some way I'd say that is revolutionary.
Yeah, the female voice especially is really impressive in the demos. The voice always sounds natural. The male voice I heard wasn't as good. It wasn't terrible, but it had a somewhat robotic feel to it.
Re: GPT-4o
#486Sorry to nitpick, but in the language tokenisation part, the french part is incorrect. The exclamation mark are surrounded by spaces in french.
"c'est un plaisir de vous rencontrer!" should be "c'est un plaisir de vous rencontrer !"
Re: GPT-4o
#487Impressed by the model so far. As far as independent testing goes, it is topping our leaderboard for chess puzzle solving by a wide margin now: https://github.com/kagisearch/llm-chess-puzzles?tab=readme-o...
woah, that's a huge leap, any idea why it's that large of a margin?
using it in chat, it doesnt feel that different
Re: GPT-4o
#488I think it’s safe to say Siri and Alexa are officially dead. They look like dusty storefront mannequins next to Battlestar replicants at this point.
Or Apple is rarely if ever the first mover on a new tech and just waits to refine the user experience for people?
Maybe Apple is not that close and Siri will be really far behind for a while. I just wouldn't count them out yet.
Re: GPT-4o
#489I am still baffled at how I can not use a VOIP number to register, even if it accepts TXT/SMS. If I have a snappy new startup and we go all in VOIP, I guess we can not use (or pay to use) OpenAI?
That's what we get when an entire industry uses phone numbers as a "proof of humanity"...
Re: GPT-4o
#490The crazy part is GPT-4o is faster than GPT-3.5 Turbo now, so we can see a future where GPT-5 is the flagship and GPT-4o is the fast cheap alternative. If GPT-4o is this smart and expressive now with voice, imagine what GPT-5 level reasoning could do!