1 Liquid & Solid Brains
Mapping the Cognition Space
@ricard_sole
+
The debate over "understanding" in #AI #LLMs @melaniemitchell
https://mastodon.social/@teixi/113245780190962868
2 #DwarkeshPatel
w #DylanPatel & #JonY
https://mastodon.social/@teixi/113246369169408095
3 #CCN2024 keynote by @NicoleCRust
https://mastodon.social/@teixi/113247415592235319
4 #MLStreetTalk w @tdietterich
https://mastodon.social/@teixi/113331859727155714
5 #MichaelLAnderson author of #AfterPhrenology on #BeyondPhrenology
https://mastodon.social/@teixi/113704895461800063
Most silliest, funniest, unexpected podcast by the usually tempered #DwarkeshPatel
w
#DylanPatel https://www.semianalysis.com/
#JonY https://www.asianometry.com/
On the many #ChipFoundries insights
#hyperscalers financials
#datacenters #ESG compliances
#TSMC #ASML competition
#US #CN export wars
#Bitcoin #GW records vs #AI forecasts
+
how complex, complicated, specialized are the design, validation, and production of chip fabric knowledge crafts
+
#windowsXP old #techdev software
When AI’s Large Language Models Shrink Smaller models trained on more data challenge the dominance of ChatGPT, GPT-4, and company
https://spectrum.ieee.org/large-language-models-size
"We won’t be able to make models bigger forever. There comes a point where even with hardware improvements, given the pace that we’re increasing the model size, we just can’t." -- #DylanPatel #SemiAnalysis
Trillion-parameter LLMs make it more or less impossible for others to reproduce. That means external researchers aren’t able to probe these models for potential safety concerns and that companies looking to deploy LLMs are “tied to the hip” of OpenAI’s dataset and model design choices.
Teixi
Mike Amundsen