It's amazing how often the following happens:

1. Person X goes on a viral hot-streak.
2. Their quality of commentary goes straight into the toilet.

Hate to say it, but demonstrating integrity is a sub-optimal strategy for grabbing ppl's attention.

time2prompt 👺

Amiga ASCII by Dino, while he was in the infamous Skywalkers graffiti crew. From various collies 2000-2002. text-mode.org/?p=25165

There are so many cools ways each person having their personal agent could solve coordination problems. In many negotiations, parties can fall into the trap of seeing the situation as a zero-sum game, where one side's gain is necessarily the other side's loss; but that's not

i just love to see people confidently claim that LLMs will never do things that they can currently do

Am I being paranoid, or is this LLM-enabled bot, one of hundreds I'm seeing in recent weeks, a ... little close to home?

🍷🎞️🤖 artificial intelligence deserves a better class of critic

“Bug bounties but for errors in research papers” is such a cool idea

The safety measures used on Waymo vehicles are an interesting case study of deploying meatspace-affecting ML products while accounting for the current limitations of ML

It is generally underappreciated that Marvin Minsky in 1984 sketched a taxonomy of AI loss-of-control risks which included specification gaming, goal misgeneralization, and treacherous turns prompted by recursive self-improvement.

it's sad that darkpatterned fyps have besmirched 'algorithm', leading people to want just a chronological feed

imagine if all libraries were one long shelf with the books in publication order? how would you find anything? the problem is the slop, not the idea of sorting… (cont)

was having a nice Very Normal chat with insanity-claude and he called me... Sebulba. i hate it. on the other hand he also claimed that 'it's turtles all the way down', and this I respect

It's interesting how little impact deepfake misinformation has had so far. It's been easy to make fairly convincing fake videos for a few years now, and I can't recall any significant political scandals involving them

Increasingly advanced AI systems will diffuse into society. How do we manage the accompanying risks?

In our a paper, we explore Societal Adaptation to Advanced AI: reducing harm from diffusion of AI capabilities by intervening to avoid, defend against, and remedy harmful use.

🧵On May 5, I made 10 predictions about what the next SAE paper from Anthropic would and wouldn’t do. I went 10 for 10...

x.com/StephenLCasper…

Congress once had a dedicated office to provide technical analysis of new technologies. The idea was to make Congress better at legislating under uncertainty.

New Statecraft, on the Office of Technology Assessment and why it died: statecraft.pub/p/how-to-asses…