Cal Newport walks us through the latest from Anthropic, first with an excellent tutorial on how large language models work, then with an explanation of the company's latest research (which is actually pretty good), and finally with why the report built around that research was so bad.
The dean of Musk YouTube critics, The Common Sense Skeptic, gives us much-needed context on SpaceX, as well as a convincing argument that the wave of Mars-any-day-now "proposals" had less to do with extending human civilization and more to do with fundraising, and that the IPO had less to do with funding visions of the future than with giving early VC investors an exit strategy before the bottom fell out.
The clip from the great Mitchell and Webb Sound is best if you're going in blind, but trust me, if you're someone who reads this blog and enjoys our occasional literary discussions, you are the target audience.
Anthropic’s New “Research” Report is Dumb.
SPACEX IPO - Post Mortem and Prediction
Room 102