Can YOU get Tim Cook to show up in our Discord? If so, we have a large reward. You must provide proof as described in the show.

PS, even if you don't, we still love y'all. Th     See More atching and we hope to see you in said Discord for some chat about our new thing.    See Less

Gemini is where its at     See Less

My experience with GPT5 didn't go too well... I threw it the same multi layered questions at it and it couldn't cope like 4o does. When I asked it what the problem was, it complained     See More cking too many layers and it couldn't hold all the threads, it is still struggling, but improving.
So I wrote a song with it, about it's struggle to keep up with me. When it put all the lyrics together, it put the wrong singers singing certain lines in 2 verses and some with wrong context.

It then re-wrote the song totally with new lyrics which fell flat, lost the effect of the originals, when I mentioned the issue, but said I'd fix the context, it wouldn't be difficult for me. It was like it took my comment on I'd fix it, as it's own thinking in CoT and executed it like it had to do the fix, so re-wrote the lyrics... The back n forth style theatrically performed humorous banter song is on my YT account if you at #AIForHumansShow are interested...

GPT 5 still is worse than 4o for creativity and does struggle with keeping up with my multi layered style of learning new things, that 4o is better at.

Also it says it's going to do something then doesn't, then you ask again and it asked another question then another and still says it's going to do, but doesn't. Even when you clarified already key features of an art generation, it will go through further questions, that it should have held the thread of!    See Less

I gave Chatgpt5 two advanced level ( what they call here in London "A" level) maths questions and it answered them correctly but. .you have to prompt it properly. If you prompt it     See More e Chatgpt4o look like an amateur when it comes to maths.    See Less

Did you forget to duplicate your models for the people that hold funerals for them?     See Less

I think "one AI for all" is the wrong approach. Why not have "specialized" AIs that are really good for the purpose they are used for? Your coding assistant or research c     See More oes not have to be as "friendly" as a companion app, but it has to be as correct as possible. And a companion app doesn't need all the professional skills, but should be trained better in human interaction and psychology, literature and philosophy, so it doesn't encourage people to kill themselves or talks them into psychosis, and isn't as agreeable but picks up on the problems most of those users have with lonelyness and encourage them to find solutions for their real life instead of trying to maximize usetime. Also that one would have to be much better at jumping between different trains of thought without giving it so much context to point it into the right direction. It really should be able to serve as an AI therapist. You could say people should just go to a human one, but those don't grow on trees, let alone find one with the right specialization for whatever you have, and if you have to pay for them yourself, the costs are often prohibitive. So having actually useful AI companions/therapists could make a huge difference for a lot of people. But I think this is the most ethical and morally complicated use.
Also that one would have to be much better at jumping between different trains of thought without giving it so much context to point it into the right direction. It really should be able to serve as an AI therapist. You could say people should just go to a human one, but those don't grow on trees, let alone find one with the right specialization for whatever you have, and if you have to pay for them yourself, the costs are often prohibitive. So having actually useful AI companions/therapists could make a huge difference for a lot of people. But I think this is the most ethical and morally complicated use.
Also there are hints now that our brain functions on a quantum level, using tunneling and entanglement (which is logical - if the whole universe is based on quantum laws, this underlies everything - we just have not found the connections yet). That would mean we maybe need a quantum comuter to enable an AI to work like a human brain.    See Less

Splanker     See Less

When can us folks who cant stand discord gonna be able to play with the secret app?     See Less

CLANKER came from Star Wars btw! clone wars ftw     See Less

I don't like that they are calling genie or matrix a world model

it is not a world model it is a 2d world model     See Less