Meta's decision not to release their AI speech tool due to deepfake concerns, China's President welcoming AI tech with potential global implications, the "Waluigi Effect" article revealing the challenges of large language models, and research papers exploring fine-tuning, inverse scaling, and unifying large language models and knowledge graphs.
Contact: sergi@earkind.com
Timestamps:
00:34 Introduction
02:12 Meta's New AI Speech Tool Is Ready-Made for Deepfakes—So It's Not Being Released
03:52 China’s President Welcomes AI Tech
05:23 The Waluigi Effect (mega-post)
07:25 Fake sponsor
09:46 Full Parameter Fine-tuning for Large Language Models with Limited Resources
11:22 Inverse Scaling: When Bigger Isn't Better
13:05 Unifying Large Language Models and Knowledge Graphs: A Roadmap
15:36 Outro
