Notes From the Desk are periodic posts that summarize recent topics of interest or other brief notable commentary that might otherwise be a tweet or note.
2024 - Rise of the robots
Last year many of the concerns of rapidly advancing AI were often countered with “but it will be decades before robots take over manual jobs”.
Foundational groundwork achieved in 2023 will likely facilitate the kind of explosion in robotics capabilities this year that we saw in the large language models (LLMs) last year.
Mobile ALOHA is demonstrating some incredible humanlike dexterity with delicate tasks. Video demonstrations can be seen in this X post.
Google’s RT-X and AutoRT are providing the path to building the LLMs of robotics. Robotics are not going to lag far behind as many people anticipated last year.
Picture a future in which a simple request to your personal helper robot - “tidy the house” or “cook us a delicious, healthy meal” - is all it takes to get those jobs done. These tasks, straightforward for humans, require a high-level understanding of the world for robots.
Today we’re announcing a suite of advances in robotics research that bring us a step closer to this future. AutoRT, SARA-RT, and RT-Trajectory build on our historic Robotics Transformers work to help robots make decisions faster, and better understand and navigate their environments.
2024 - Post-truth society advances rapidly
Hacker News largely dismissed my concerns last year over the approaching post-truth society. But dismissals are not going to stop progress and progress continues unabated. AI accelerates post-truth civilization essay was written in response to the dismissals.
“It was only a few months ago that distinguishing between AI and human art could be done at a glance. Now it takes a bit more than just a glance and some of the best images can no longer be deciphered by those without art or photography backgrounds. Likely we are on pace that by the end of the year nobody will know the difference from the best generated images.”
Mind Prison - June 30, 2023
It seems we have arrived essentially where I had predicted June of last year. A reddit post just recently went viral on creating a fake verification photo.
And if that seems impressive you must watch this video. The following is completely fake as well. This X post by Ethan Mollick describes the details and has more video references.
Me manipulating the microphone was in the real reference video, it incorporated it into the fake avatar video, along with fake blinks, mouth movements, and the voice, which was all AI generated.
— Ethan Mollick
Consider the potential AI safety implications for future powerful AI. This year expect some nightmares of cybersecurity failures as well as disturbing political manipulations. AI safety, alignment of AI, is premised on the idea we will be able to solve problems before they occur. We are already failing in that endeavor with primitive weak AI.
No compass through the dark exists without hope of reaching the other side and the belief that it matters …
Read the in-depth AI and societal introspection I wrote at the beginning of 2023 AI and the end to all things ...
Its pretty concerning, that is for certain. Keeping optimism in today's world, let alone optimism for our children's future can be hard.
The antichrist will be an AI.