🧵 post science methods
Anonymous at Tue, 19 Mar 2024 13:30:03 UTC No. 16086480
What do you expect from a post science world? I expect science to largely be deprecated soon.
An immediate tech tree upgrade would be abandoning the current system of science for data collection and neural networks. Meaning the new paradigm is creating mass data collection and relying on neural networks of any architecture to predict the output or find patterns. No more hypothesis and experiment but rather data capture/creation and NN predictions.
What do you think is coming soon with post science paradigms? Either immediate or far-future upgrades to the algorithm? I think we can all agree the age of journals is ending. We should fundamentally analyze science, how it functions, and how it can be altered or improved and not treat it like a holy bible forever.
Anonymous at Tue, 19 Mar 2024 13:33:22 UTC No. 16086484
>post science
The only thing post science, if it ever comes to that, is /x/ related. Everything else is still science.
What you are talking about is post academia.
Barkon at Tue, 19 Mar 2024 13:37:47 UTC No. 16086487
Simulation updates to bring about new source. But for that you need a simulation technician. Look no further. Barkon at your service.
Anonymous at Tue, 19 Mar 2024 13:40:12 UTC No. 16086489
>>16086484
If you reduce science to it's simplest form. It becomes extremely micro in function - in that case yes you would focus on macro systems around it that maximize it's output/ collection/networking. So yes it would be how to replace the academia meta around it to something more optimal.
Anonymous at Tue, 19 Mar 2024 13:51:28 UTC No. 16086500
>>16086489
>how to replace the academia meta
Baring some /x/ related shit going down, AI and automation.
The only real yoke institutions have on practicing scientists is money. When you have your own robotics supply chain to manufacture nearly everything you need you are beholden to no one and can experiment and publish whatever you want whenever you want and however you want.
Before this becomes a reality politics will always be a corrupting factor to true science.
Anonymous at Tue, 19 Mar 2024 13:53:03 UTC No. 16086502
In any case a few simple things
Human thought is not keeping pace with computational capacity, so anything that replaces current scientific meta will be computational.
The intuition of the system can be much larger and communicative. It won't be limited by the need to communicate it's intuition as it can "dump" it's weights and easily be tested - communication of science can become massively more complex without low bandwidth humans. Just dumping raw data used to train would be enough, no reason for more as it is easily produced from the data. So again, data will become everything, explanation will be reserved for explaining to humans if that is ever done.
Almost all improvements would come via architecture and systems or "expanding intelligence itself"
Anonymous at Tue, 19 Mar 2024 13:55:11 UTC No. 16086503
>>16086502
Intelligence itself...
Here we go over ten thousand years of human development
Were did they figured what was intelligence and how they improved it?
Anonymous at Tue, 19 Mar 2024 14:07:26 UTC No. 16086518
>>16086503
I'm sure science is good enough to have solved it. Just go to the science repository and look it up.
Anonymous at Tue, 19 Mar 2024 14:44:39 UTC No. 16086576
>>16086502
this won't work. Computers still depend on the limiting thoughts of humans. And even more fundamentally they're just non-linear universal approximators. They can't know or solve because they lack "creativity", "intelligence", and as a whole "thought".
Barkon !otRmkgvx22 at Tue, 19 Mar 2024 14:45:17 UTC No. 16086577
No'qqqqq
Anonymous at Tue, 19 Mar 2024 16:13:41 UTC No. 16086665
>>16086480
>the new paradigm is creating mass data collection and relying on neural networks of any architecture to predict the output or find patterns. No more hypothesis and experiment but rather data capture/creation and NN predictions.
If we still played some role in the process of AI doing science then we'd still be doing science but just using AI as a tool to do a majority of the work. I don't think science would be complete without experiments, they have to be done to verify what's being predicted is real and not fantasy
Anonymous at Tue, 19 Mar 2024 21:13:55 UTC No. 16087037
>>16086576
>creativity
well that's out the window now. at least for art.
creativity is a sort of brute forcing. when you get something that works it seems amazing, especially from the outside.
https://newatlas.com/materials/deep
Anonymous at Tue, 19 Mar 2024 21:20:25 UTC No. 16087044
>>16086665
once some of the predictions pan out to be true, they will get in the loop and in time humans will have less and less input. unless someone can come up with a reason why humans will always be needed, eventually we'll step back from the whole thing.