Atlas Shrugged and the irreplaceable founder, pumping stations and civic pride, and thoughts on the eve of AGI
by Jason Crawford
·
January 26, 2025
·
3 min read
Much of this content originated on social media.To follow news and announcements in a more timely fashion, follow me on Twitter, Threads, Bluesky, or Farcaster.
Marian Tupy at HumanProgress.org is hiring analysts to explore the economics and psychology of human progress
Alan Tomusiak is hiring scientists to work on the problem of genome instability (@alantomusiak)
Ashlee Vance is hiring for his new publication, Core Memory: “Are you an ambitious type based in DC who can write a weekly newsletter that dives into tech-related legislature and discern what’s real and has real money involved versus political garbage? … Can you do this with some flair but not let your politics color the facts of what’s going on? Can you spot interesting military and infrastructure bids and break them down? Can you make this a must read for people in the tech industry? Can you go deeper on the juicy stuff and really add context? If so, let’s talk. I’ll help give you a big audience and develop your following” (@ashleevance). Email him: ashlee@corememory.com
Edge City Austin, March 2–7: “explore how frontier tech can be built for human flourishing. Live, cowork, and collaborate in this fun week before SXSW” (@JoinEdgeCity)
“ANWAR will have its US premiere at Sedona Film Festival! Screenings with Q&A Feb 25 & 27. ANWAR is a sci-fi short about a mother who chooses to live forever, and a son who longs for heaven” (@FawazAM). ANWAR was screened at Progress Conference 2024, with discussion from the writer-director and producer.
“The Stargate Project is a new company which intends to invest $500 billion over the next four years building new AI infrastructure for OpenAI in the United States. We will begin deploying $100 billion immediately” (@OpenAI). Lots of skepticism about how real this is: it’s unclear how secured the funding is, and “intends” may be doing a lot of work here. But I wouldn’t bet against Sam
DeepSeek releases R-1, a model on par with OpenAI’s o1. “Fully open-source model,” MIT licensed. Lots of chatter about this because (1) DeepSeek is a Chinese lab, (2) they have distilled some of the models down pretty small, and at least some of them are open, to the point where you can run them on your laptop, (3) there are some claims about the model costing very little to develop, followed by counterclaims that China is hiding the fact that they’re in violation of export controls (@kimmonismus, @avichal). RPI fellow Dean Ball provides some context
Anthropic introduces Citations: “Our new API feature lets Claude ground its answers in sources you provide. Claude can then cite the specific sentences and passages that inform each response” (@AnthropicAI)
Humanity’s Last Exam: “a dataset with 3,000 questions developed with hundreds of subject matter experts to capture the human frontier of knowledge and reasoning. State-of-the-art AIs get <10% accuracy and are highly overconfident” (@DanHendrycks)
Other news
Zipline piloting drone delivery in Pea Ridge, Arkansas (@zipline). “They’re amazed by how quiet it is. They’re delighted by how charming it feels. They’re surprised that it doesn’t require any special packaging. And so much more. … Everyone who has witnessed this new delivery system in action at their doorstep is convinced they’ve just experienced the future of delivery.” (@keenanwyrobek)