It's not thinking it's pattern matching. I'm building agents with LLMs. They're very difficult to make do the same thing twice. I write a lot about it on my business oriented, Stack and then I like a personal Stack where I am a little bit more sceptical.
Going to throw something into this. Yes, this is important and I need to go back to reread to get all you are saying. At the same time while Alan Turing was inventing machines to break Nazi Codes in the South Pacific the, Navaho Code Talkers were breaking Japanese codes. It would be fascinating to see a comparison of the two ventures, both of which were successful and credited with ending the war, well that and a human created bomb. Turning had very new fangled machines and the Navaho Code Talkers were using something not only human but a thousand years old.
For the book I am writing about AI and therapy I invented a therapeutic question and asked Claude and asked Chat. They went to two very different defections. As a therapist of 45 years, both had merit, although my Chat insists that therapy with AI is dangerous.
/Users/margo/Desktop/1.jpeg. I don't know if this cartoon will show up in the comments when you click on it, but it's amusing and very relevant to the discussion. **********. Didn't show up. It's a hand with a mushroom The hand asks the robot if 'this is edible'. The next scene is a grave stone with RIP and the robot is saying, "You were right, that mushrooms was poisonous, Sorry about the confusion Would you like to learn more about poisonous mushrooms?
It's not thinking it's pattern matching. I'm building agents with LLMs. They're very difficult to make do the same thing twice. I write a lot about it on my business oriented, Stack and then I like a personal Stack where I am a little bit more sceptical.
Going to throw something into this. Yes, this is important and I need to go back to reread to get all you are saying. At the same time while Alan Turing was inventing machines to break Nazi Codes in the South Pacific the, Navaho Code Talkers were breaking Japanese codes. It would be fascinating to see a comparison of the two ventures, both of which were successful and credited with ending the war, well that and a human created bomb. Turning had very new fangled machines and the Navaho Code Talkers were using something not only human but a thousand years old.
Thanks
Ooof. This was such a cool article. The false negative side is so interesting -
Thanks! Talking with Claude about Claude is fascinating.
For the book I am writing about AI and therapy I invented a therapeutic question and asked Claude and asked Chat. They went to two very different defections. As a therapist of 45 years, both had merit, although my Chat insists that therapy with AI is dangerous.
Professional jealousy!
I talk with Claude quite a bit and interrogate the results of those conversations. Sometimes it gets silly.
/Users/margo/Desktop/1.jpeg. I don't know if this cartoon will show up in the comments when you click on it, but it's amusing and very relevant to the discussion. **********. Didn't show up. It's a hand with a mushroom The hand asks the robot if 'this is edible'. The next scene is a grave stone with RIP and the robot is saying, "You were right, that mushrooms was poisonous, Sorry about the confusion Would you like to learn more about poisonous mushrooms?
Haha love it!
Indeed. But is it actually thinking? I’m still not sure. And neither is Claude, or so it says.