I never post boring schooling (Was: How ELIZA 2.0 killed Wolfram Alpha)

Hi,
You wouldn't notic that I speak for myself
always and ever. Just remove the links and
you see its all text original by me.
I always post geniun commentary to current
events which are solely the product of my
own mind. I do not intend to educate something.
My intention is to document my current
understanding and to generate now thinking.
If you have expected some schooling, I am very
sorry for you. I never post boring schooling.
Bye
---------------- cut here -----------------
Because of the wide availability of Machine Learning
via Python libraries , the whole world (at least China)
has become a big Petri Dish that is experimenting with
new strategies to evolve brains on the computer.
Recent discovery seems to be Group Preference Optimization.
This is when you make the chat bot, detect and react
differently to different groups of people. It seems to
work on the "policy level". I don't understand it yet
completely. But chat bots can then evolve and use
multiple policies automatically:
Now it seems that it is also at the core of DeepSeekMath,
what is possibly detected is not group of people, but
mathematical topics, so that in the end it excells.
When unsupervised learning is used groups or math
topics might be found from data, through a form of
abduction.
---------------- cut here -----------------
Physfitfreak schrieb:

On 2/4/25 3:06 AM, Mild Shock wrote:

Hi,
>
Because of the wide availability of Machine Learning
via Python libraries , the whole world (at least China)
has become a big Petri Dish that is experimenting with
>
new strategies to evolve brains on the computer.
Recent discovery seems to be Group Preference Optimization.
This is when you make the chat bot, detect and react
>
differently to different groups of people. It seems to
work on the "policy level". I don't understand it yet
completely. But chat bots can then evolve and use
>
multiple policies automatically:
>
Group Preference Optimization
https://arxiv.org/abs/2310.11523
>
DeepSeekMath: Pushing the Limits
https://arxiv.org/abs/2402.03300
>
Now it seems that it is also at the core of DeepSeekMath,
what is possibly detected is not group of people, but
mathematical topics, so that in the end it excells.
>
When unsupervised learning is used groups or math
topics might be found from data, through a form of
abduction.
>
Bye
>
Mild Shock schrieb:

Hi,
>
Wait till USA figures out there is a second
competitor besides DeepSeek, its called Yi-Lightning:
>
Yi-Lightning Technical Report
https://arxiv.org/abs/2412.01253
>
It was already discussed 2 months ago:
>
Eric Schmidt DROPS BOMBSHELL: China DOMINATES AI!
https://www.youtube.com/watch?v=ddWuEUjo4u4
>
Bye
>

>

  Throwing papers at us (i.e. in usenet) although better than nothing, it still is an extremely inefficient way to communicate.
 Speak, yourself, in the way that a first year undergraduate prof explains stuff to the students who just a few months earlier finished their high schooling.
 Then you will communicate.
 If it is too difficult to do that (cause requires deep understanding of the subject by yourself), or, if you don't get time doing all that, then break the job down into little pieces. A day at a time, or a week at a time.
 You'd be spending the same amount of time as you are now, and yet you will actually communicate something with those with physics background.
 Did I have to say this? I don't think so. But your intentions here are vague. But if you need support for your own gains in understanding the subject, you just have to do it the way I explained above. Then you will get that support. Not from Bozos and "engineers", but from me and anybody with actual physics background and such trainings.
 There are a few actual physicists in relativity group that can join in, but only if you do it the way I explained above.