Links for 2023-01-26
What a compute-centric framework says about AI takeoff speeds https://www.lesswrong.com/posts/Gc9FGtdXhK9sCSEYu/what-a-compute-centric-framework-says-about-ai-takeoff
“Legal standards facilitate robust communication of humans' underspecified goals…we demonstrate that large language models (LLMs) are beginning to exhibit an "understanding" of one of the most relevant legal standards for AI agents: fiduciary obligations.” https://arxiv.org/abs/2301.10095
Making Cognitive Enhancement Palatable https://parrhesia.substack.com/p/making-cognitive-enhancement-palatable
Finally, a Fast Algorithm for Shortest Paths on Negative Graphs https://www.quantamagazine.org/finally-a-fast-algorithm-for-shortest-paths-on-negative-graphs-20230118/
How AI Found the Words to Kill Cancer Cells: Predictive Model Allows Researchers to Encode Commands for Cells to Carry Out https://www.ucsf.edu/news/2022/12/424406/how-ai-found-words-kill-cancer-cells
Hearing AI-ds https://keyhe.com/portfolio/hearing-ai-ds/
Critique of some recent philosophy of LLMs’ minds https://www.lesswrong.com/posts/ejEgaYSaefCevapPa/critique-of-some-recent-philosophy-of-llms-minds
“Here I take the raw data from the National Crime Victimization Surveys and construct summary results missing from recent reports.” https://inquisitivebird.substack.com/p/constructing-missing-crime-tables
“The White House Office of Science and Technology Policy estimates it would cost just over $24 billion to have prototype vaccines ready for each of the 26 known viral families that cause human disease.” https://progress.institute/why-barda-deserves-more-funding/
A simple, clear explanation of how a linear electric motor works https://youtu.be/uf_Z57gAJTc
Thanks to gwern, I was able to fix a bunch of links on this old post: Genetic confounding: When you hear "privilege" or "discrimination" always think "genetic confounding" or risk wasting billions on futile efforts https://axisofordinary.substack.com/p/genetic-confounding
Your frequent reminder that the world could have defeated climate change in the 1980s if all industrialized nations had followed France's lead:
This would have been vastly cheaper than what nations are now pledging to spend in order to combat climate change. Most importantly, it would have been a solution compatible with further growth. But environmentalists sabotaged this technological solution.
And the same people who now claim that we face an extinction event and that we need to dismantle capitalism are still against nuclear. They claim it is too expensive (more expensive than dismantling capitalism?). They claim it is too dangerous (more dangerous than global extinction?). And they claim that it takes too long (France showed that it can be done quickly.).
Some additional notes:
Nuclear waste isn't a problem: https://threadreaderapp.com/thread/1550148385931513856.html
Nuclear fuel isn't limited: https://threadreaderapp.com/thread/1515325071124045829.html
Nuclear power is safe: “440 reactors producing electricity in their fourth to sixth decade of service...100s more power ships & submarines...just 3 accidents...” https://democracyjournal.org/arguments/a-1908-lesson-in-risk-and-reward/
Nuclear power is expensive because of overregulation: https://rootsofprogress.org/devanney-on-the-nuclear-flop