27:23
1h 27mA mesmerizing shot from within the water and sometimes from above, being dragged across behind the ship with innumerable birds covering the sky. Whenever the camera is above the water, and on the other times, it's just a wallpaper of moving ripples and water. It's very cinematic.
Also the shot of one of the fishermen just sitting alone and watching tv, barely reacting to what. What's on tv, and then me realizing, like, this is entertainment for them and this is the rest of their lives, that they're not doing the gruesome fishing work that they're barely getting paid for.
Also these shots of birds flying at night, where, like this darkness of the sea and the sky dissolve. And sometimes the camera keeps moving angles and becoming upside down, so you can't even tell which direction the birds are flying and which way they're oriented. It's mesmerizing. Cinematography.
This capability of developing training data for RLVR is what makes the RSI process of automated development of models using LLMs an AGI. It's the process of model development that becomes an AGI, while the individual LLMs don't count as AGIs, since they can't learn novel deep skills. In contrast, in-context learning (that an individual LLM possesses) is too weak, and continual learning with anything resembling the current methods is likely to either follow the weakness of in-context learning, or to inherit the sample inefficiency of pretraining.
Degradation of natural boundaries. We rely on friends and loved ones for emotional support, but they aren't ever-present, so we have to also learn to cope on our own, which is important for developing a stable identity.[2]Always-available AI companions degrade that, which is likely one contributor to existing cases of AI psychosis.