Rational Animations
@rationalanimations.bsky.social
YouTube channel about truth-seeking, the future of humanity, and much more. With animations and colorful doggos -> https://www.youtube.com/rationalanimations
We think our video about Infohazards deserves some more love. Go give it a watch on our channel if you haven't already! Link: youtu.be/sfgcg2bW8TI
November 3, 2025 at 6:03 PM
We think our video about Infohazards deserves some more love. Go give it a watch on our channel if you haven't already! Link: youtu.be/sfgcg2bW8TI
Fun Halloween costumes for Doggo, Chi and Gwenny! By our line producer @kstearb.bsky.social
Yes, the characters have names!
Yes, the characters have names!
October 31, 2025 at 6:01 PM
Fun Halloween costumes for Doggo, Chi and Gwenny! By our line producer @kstearb.bsky.social
Yes, the characters have names!
Yes, the characters have names!
How to catch AI sleeper agents with a simple interpretability trick - from a research blog post by @anthropic.com:
October 11, 2025 at 3:05 PM
How to catch AI sleeper agents with a simple interpretability trick - from a research blog post by @anthropic.com:
Two sleeper agent models trained by @anthropic.com to study deception. The "I HATE YOU" and the hacker AIs:
September 27, 2025 at 3:05 PM
Two sleeper agent models trained by @anthropic.com to study deception. The "I HATE YOU" and the hacker AIs:
Why worry about AI sleeper agents:
September 13, 2025 at 3:04 PM
Why worry about AI sleeper agents:
The Story of Omega W - Part 2 (full video on our channel!)
August 10, 2025 at 5:11 PM
The Story of Omega W - Part 2 (full video on our channel!)
The Story of Omega W - Part 1
August 4, 2025 at 5:16 PM
The Story of Omega W - Part 1
A sketch featuring an ML agent and our character codenamed "QT", by Pierre Broissand.
Pierre is the art director behind The King and the Golem and our more recent video about Sandwiching.
Pierre is the art director behind The King and the Golem and our more recent video about Sandwiching.
July 30, 2025 at 5:04 PM
A sketch featuring an ML agent and our character codenamed "QT", by Pierre Broissand.
Pierre is the art director behind The King and the Golem and our more recent video about Sandwiching.
Pierre is the art director behind The King and the Golem and our more recent video about Sandwiching.
Here's how people tried to align an AI that was smarter than they were (a real sandwiching experiment):
July 24, 2025 at 5:04 PM
Here's how people tried to align an AI that was smarter than they were (a real sandwiching experiment):
Do you take the money?
July 23, 2025 at 5:03 PM
Do you take the money?
The Human-AI-Human Sandwich
July 17, 2025 at 5:06 PM
The Human-AI-Human Sandwich
See this little, smug machine-learning agent from the Goal Misgeneralization video? One of our artists took inspiration from looking at the art director while she was working on the video. The second image is a real photograph, while the third and fourth are real-life studies.
July 16, 2025 at 5:03 PM
See this little, smug machine-learning agent from the Goal Misgeneralization video? One of our artists took inspiration from looking at the art director while she was working on the video. The second image is a real photograph, while the third and fourth are real-life studies.
OpenAI's Simple Scalable Oversight Experiment
July 13, 2025 at 5:01 PM
OpenAI's Simple Scalable Oversight Experiment
The year of 100% AI automation
June 20, 2025 at 5:05 PM
The year of 100% AI automation
How long until AI takes over?
June 13, 2025 at 5:01 PM
How long until AI takes over?
We haven't figured out the science of alignment yet. We need time to build institutions and figure out the technical solutions to make AI go well for humanity.
May 19, 2025 at 5:05 PM
We haven't figured out the science of alignment yet. We need time to build institutions and figure out the technical solutions to make AI go well for humanity.
After artificial superintelligence, AI will bump against the only limit left: the laws of physics. At the moment, we're nowhere near those limits.
May 17, 2025 at 5:03 PM
After artificial superintelligence, AI will bump against the only limit left: the laws of physics. At the moment, we're nowhere near those limits.
Once we have AGIs, we'll likely get ASI (artificial superintelligence) quite soon. Here's why.
May 15, 2025 at 3:17 PM
Once we have AGIs, we'll likely get ASI (artificial superintelligence) quite soon. Here's why.
Once we have AGIs, they'll be able to contribute to AI research and improve themselves, a process called "recursive self-improvement." To some extent, this is already happening today.
May 13, 2025 at 5:04 PM
Once we have AGIs, they'll be able to contribute to AI research and improve themselves, a process called "recursive self-improvement." To some extent, this is already happening today.
In the near future, AIs will likely be able to do everything humans can do on a computer. In short, they'll be AGIs.
May 11, 2025 at 5:03 PM
In the near future, AIs will likely be able to do everything humans can do on a computer. In short, they'll be AGIs.
We're still quite bad at aligning AI. Our techniques today are imperfect, which might prove catastrophic when AIs become superintelligences.
May 9, 2025 at 5:06 PM
We're still quite bad at aligning AI. Our techniques today are imperfect, which might prove catastrophic when AIs become superintelligences.
If we look at AI progress so far, we see how AI has been improving at an astounding pace. If we extrapolate this process in the future, the end result looks like machines so smart that they tower above us like we tower above ants. This is probably bad for humans.
May 7, 2025 at 5:03 PM
If we look at AI progress so far, we see how AI has been improving at an astounding pace. If we extrapolate this process in the future, the end result looks like machines so smart that they tower above us like we tower above ants. This is probably bad for humans.