This week - awesome GPT-3; the genetically engineered astronaut; a real-life Gundam takes its first step; and more!
More Than A Human
Humans are not made for space. But what if we make a human designed to thrive in space? Some scientists looked into this question. This article showcases findings of researchers that looked into tweaking personality on a genetic level (and they found it is an extremely complicated issue) and only glossed over the possibility of equipping future astronauts with improved radiation protection or other physical traits useful in space.
Professor Anders Sandberg explores his background in computer science, neuroscience, and engineering and discusses mind-uploading, cognitive enhancement, and the future of happiness.
This guy took OpenAI's GPT-3 text generator and trained it to write comedy. I will leave it up to you to decide if this cautionary tale about the dangers of doing Reddit AMAs is funny or not.
Here is a list of interesting demos, articles and products build around GPT-3 - the latest iteration of OpenAI's text generating AI. It is a quite powerful tool and various people used it in sometimes unexpected ways (like building a Dungeon Master or generating web components just from their descriptions).
I don't think many people would consider modelling as a job threatened by the AI revolution. But there are startups that work on computer-generated models that can look and be whoever you need.
There is an AI out there that if you give it a picture of a bird it will tell what bird is it. The team says the AI correctly identified over 90% of the wild species and 87% of the captive ones. Using AI for labelling birds can be less stressful for birds (current methods require catching the bird and fitting them with electronic tags) and could offer new insights about their lives.
The DeepMind team described in a recent paper a new deep reinforcement learning algorithm that was able to discover its own value function—a critical programming rule in deep reinforcement learning—from scratch. The algorithm was also effective beyond the simple environments it trained in, going on to play Atari games at a level that was, at times, competitive with human-designed algorithms and achieving superhuman levels of play in 14 games.
Well, first step... it's more like a joints test but hey, some mad engineers in Japan are building an actual Gundam!
Japanese contractor Obayashi has started to build a dam almost entirely with robots, addressing the industry's labor shortage and aging workforce. The site of the trial project is a concrete dam in Mie Prefecture, on the southeast coast of Japan's main island. The 84-meter-high structure is slated for completion in March 2023.
Before I sat down to write this issue of H+ Weekly, I read Why Futurism Has a Cultural Blindspot. It argues that futurism is good at predicting technology but it does that by not considering the social and cultural changes. As the subtitle says, "We predicted cell phones, but not women in the workplace" because our visions of future are largely the same but with new toys.