Hype, Sustainability, and the Price of the Bigger-is-Better Paradigm in AI
With the growing attention and investment in recent AI approaches such as large language models, the narrative that the larger the AI system the more valuable, powerful and interesting it is is increa...
On the opposite, a race for scale in AI, leads to other power redistribution, as we discussed with @meredithmeredith.bsky.social and @sashamtl.bsky.social
arxiv.org/abs/2409.14160
Bigger models require data and infrastructure that consolidate big actors
Let us actually democratize AI โค๏ธ
3/3
09.12.2024 23:55
๐ 42
๐ 13
๐ฌ 1
๐ 1
It would be so nice if you upload material online after seminar is done.
07.12.2024 14:23
๐ 1
๐ 0
๐ฌ 0
๐ 0
I will keep the fact that you have seen many calibration plots, means that at least people do more than AUC
27.11.2024 20:43
๐ 0
๐ 0
๐ฌ 0
๐ 0
The brand new `legendry` package by @teunbrand.bsky.social takes ggplots2 guides to the next level ๐คฉ
teunbrand.github.io/teunbrand_bl...
22.11.2024 03:44
๐ 234
๐ 55
๐ฌ 16
๐ 8
It feels that the longer I do statistics, the harder it becomes. I never get an analysis just "right". There is always one more assumption to change or one more term to add to my model.
16.11.2024 12:50
๐ 111
๐ 14
๐ฌ 7
๐ 2