Onwards, 2024.

With ChatGPT in 2024...

Indeed, it is a strange world out there. So, we should
all make plans to prepare ourselves for what lies ahead...
Perhaps, with a little help from ChatGPT...

Testing ChatGPT (March 2024).

Testing ChatGPT-o1-Preview (September 2024).
Problems with ChatGPT-o1-Preview (September 2024).

Friendly ChatGPT Plans.

ChatGPT would like to see itself as your friend...
Useful, informative and friendly:

It is also eager to help you out with your plans... (December 2023).
But, well, making plans, that is a difficult thing...

Well, well...
There is something wrong here...

-Simon

Simon Laub (Let me Google that for you).
www.simonlaub.net

Testing ChatGPT4, March 2024.

March 11th, 2024.

The chatbots improve, dramatically, all the time...
But, there are still problems...

And now, something about monkeys and bananas:

Well, well...
As Peter meets 2 talking monkeys, this was not the answer we were looking for...

-Simon

Simon Laub (Let me Google that for you).
www.simonlaub.net

Testing ChatGPT-o1-Preview,
September 2024.

September 14th, 2024.

The chatbots improve, dramatically, all the time...
Now, some of the problems, from earlier models (back in March),
appears to be fixed
(with the September release ChatGpt-o1-Preview)...

Well, well...It is getting better ...

Didn't ask for a lunchbreak, but got one on all days,
except Tuesday and Thursday?
And, well, the timeinterval between different activities Saturday and Sunday
doesn't appear to be super consistent?

-Simon

Simon Laub (Let me Google that for you).
www.simonlaub.net

Problems with ChatGPT4-o1-Preview...

September 16th, 2024.

According to OpenAI, ''o1 is a new large language model
trained with reinforcement learning to perform complex reasoning''...

So, what about this problem:

Well, ChatGPT-o1's reasoning powers still isn't on a super human level,
when it comes to reasoning from first principles
(September 2024)...

(Where: First principles thinking is the practice of questioning every assumption
you think you know about a given problem, then creating new solutions from scratch).

But, it can now solve the ''boxes with fruit'' problem: