News for hackers. Stuff that matters.
Re:Gemini Flash by numbers on Tuesday May 14, 2024 @03:34PM (#40359198)
It's ironic that when you ask these AI chatbots what their own context size is, they don't know. ChatGPT doesn't even know about 4o existing in 4o.
Re:Gemini Flash by advisedwang on Tuesday May 14, 2024 @04:08PM (#40359584)
Ask a human how many neurons they have. Hell, humans haven't even consistently understood that the brain is where cognition happens.
Re:Gemini Flash by SoftTalker on Tuesday May 14, 2024 @03:38PM (#40359245)
Does a monkey know that it is a monkey?
Re:Gemini Flash by verdverm on Tuesday May 14, 2024 @03:42PM (#40359297)
I think "yes" is the most likely answer here

animals have a lot more intelligence than they typically get attributed

Tool use, names, language, social structure and behavior, even drug use has been shown across many species

Re:Gemini Flash by chaorace on Tuesday May 14, 2024 @04:22PM (#40359731)
Okay, but the monkey doesn't know that it knows that it's a monkey.
Re:Gemini Flash by verdverm on Tuesday May 14, 2024 @06:57PM (#40361095)
are you sure?

Many animals recognize themselves and their species as separate concepts

Re:Gemini Flash by keefle on Wednesday May 15, 2024 @03:06AM (#40363939)
He meant something more meta I believe. Knowing you are a monkey is one thing, and knowing that you know you are a monkey is a another thing. It's about being cognisant of the fact that there is something called knowledge and you have it
Re:Gemini Flash by chaorace on Wednesday May 15, 2024 @10:53AM (#40367758)
Precisely. To put it more concretely: it is no small feat to grasp the abstract distinction between known-knowns, known-unknowns, unknown-knowns, and unknown-unknowns. They do not know what they do not know.
Re:Gemini Flash by fourthark on Tuesday May 14, 2024 @10:00PM (#40362312)
How do you know?
Re:Gemini Flash by simonw on Tuesday May 14, 2024 @05:37PM (#40360436)
The models didn't exist when their training data was collected.

But... that's not really an excuse any more. Model vendors should understand now that the most natural thing in the world is for people to ask models directly about their own abilities and architecture.

I think models should have a final layer of fine-tuning or even system prompting to help them answer these kinds of questions in a useful way.