Hi Ed, Thanks for the feedback. My main point is not that these programs are always wrong. It's that they should never be used. Why predict the right answer to a question in personal finance, when you can calculate the right answer it directly? All best, Larry
I have been using openai's chatGPT since Dec 2022. What I have found, continuously, is that if you ask it a question you know the answer to (like writing a program that does something moderately complex), it always fails to deliver a correct answer (a program, in this case). I do this with openai, copilot (Microsoft), Gemini, Meta.AI, Claude.ai, Mistral.ai, poe.com, agent.ai, and perplexity.ai. They are getting better.
Most some up with useful information, but all details have to be verifited outside the world of ai.
thanks, John! Pls spread the word. Yours, Larry
Hi Ed, Thanks for the feedback. My main point is not that these programs are always wrong. It's that they should never be used. Why predict the right answer to a question in personal finance, when you can calculate the right answer it directly? All best, Larry
You've done a real service here.
Excellent report of Google AI.
I have been using openai's chatGPT since Dec 2022. What I have found, continuously, is that if you ask it a question you know the answer to (like writing a program that does something moderately complex), it always fails to deliver a correct answer (a program, in this case). I do this with openai, copilot (Microsoft), Gemini, Meta.AI, Claude.ai, Mistral.ai, poe.com, agent.ai, and perplexity.ai. They are getting better.
Most some up with useful information, but all details have to be verifited outside the world of ai.
Thans again for the excellent post.