This week - how to teach a robot to dance; another misbehaving AI chatbot removed from the internet; robots made from ice; AI accelerated by light; and more!
More Than A Human
I was pleasantly surprised to see this short video on BBC explaining that ageing can be reversed and there are people actively working on that problem. And you don't have to wait for anti-ageing therapies to become available - you can change your lifestyle right now or, if you like to live dangerously to live longer, try out some experimental therapies. Overall, it is a nice, well-rounded starting point for everyone to get interested in anti-ageing research and therapies.
Running a leading AI lab is not cheap and not profitable (yet). According to its annual report filed with the UK’s Companies House register, DeepMind has more than doubled its revenue, raking in £266 million in 2019, up from £103 million in 2018. But the company’s expenses continue to grow as well, increasing from £568 million in 2018 to £717 in 2019. The overall losses of the company grew from £470 million in 2018 to £477 million in 2019.
Remember Tay, Microsoft's AI chatbot that shortly after relating into the internet started to write homophobic, racist and sexist comments? The same thing happened recently in South Korea to Lee Luda, another AI chatbot imitating a 20-year old student, and had to be pulled from Facebook after engaging in homophobic slurs.
One of the interesting problem in AI is the control problem - if we create super-intelligent AI, can we make sure it will act in our interests? Well, according to this study, we can't.
Researchers working on processors using light waves instead of electrons believe they can address hardware bottlenecks we are approaching and bring more computing power required to run more powerful AIs. The technology is still in initial stages but it promises to reduce latency and power consumption while increasing processing power.
How does a robot dog learn how to dance? Adam Savage and the Tested team examine and dive into Boston Dynamics' Choreographer software that was behind Spot's recent viral dancing video.
And here is an interview with Aaron Saunders, Boston Dynamics’ VP of Engineering, which gives a bit of behind the scenes sense how Boston Dynamics engineers taught their robots to dance. This might seem like a silly idea - make robots dance - but in fact, it was a big research project on how to translate human movement to a robotic body and how to quickly teach a robot new skills.
Researchers from Harvard developed fish-inspired robots that can synchronize their movements like a real school of fish, without any external control. The video inside the article shows how these robots self-organised themselves and it looks very much like what a real school of fish would do.
I don't think anyone expected to see robots made from ice but here we are.
Bionic Hive's Squid robot not only roams on the floor but it also can attach itself to the shelves and reach any place in the warehouse.