The AI lab gets to throw Microsoft’s supercomputing and cloud computing muscle at its bid to build artificial general intelligence (AGI)....
The news: Microsoft says OpenAI will help it jointly develop and train new AI technologies for its Azure cloud computing service. It will also work with it to develop new supercomputing hardware to try to achieve AGI—machines with the capacity to learn tasks like human beings do. That’s a holy grail of AI that still remains (and may always remain) out of reach. OpenAI’s founders, which include Elon Musk and other tech leaders, reckon AGI could help solve longstanding challenges in areas that range from climate change to healthcare.
Show me the money: Since it was set up in 2015, OpenAI has developed AI that’s sought to defeat human players at games like Dota 2 and frighteningly effective language AI, among other things. It began as a non-profit research lab with the mission of developing safe AGI, but AI models need mountains of data to crunch, and that requires huge amounts of expensive computing power. So, earlier this year, Open AI set up a new for-profit arm to help pay for its work. (Open AI calls its model “capped-profit” because investors can only achieve maximum 100x return on their investment. Still, not bad.)
Go big or go home: The investment is part of Microsoft's new strategy, and will help the firm keep pace with rivals like Google, Facebook and Amazon who have all spent big on AI in recent years.
While many AI researchers believe AGI is unachievable, OpenAI may be hoping to reach it by scaling existing deep learning approaches massively, rather than by developing new ones. There’s no guarantee that throwing more brute computing power will crack the problem, if it can even be cracked, so Microsoft’s billion dollar bet is definitely a risky one.
Sign up here for our daily newsletter The Download to get your dose of the latest must-read news from the world of emerging tech.
It was the second attempt, after a scheduled liftoff last week was postponed because of technical issues....
The news: India’s Chandrayaan-2 mission to the moon launched at 04.13 EDT today from the Sriharikota space station in the country’s southeastern coastal state of Andhra Pradesh.
Second go: The launch was due to take place exactly a week ago on July 15, but it had to be postponed just 56 minutes before launch because of a leak from a helium gas bottle in the rocket’s engine, according to the Times of India. It was simple to fix, but left unresolved, it could have resulted in total failure, a source from the Indian space agency Isro told the BBC.
Two firsts: If successful, it would be the first Indian spacecraft to land on the moon, and the first ever to land on the moon’s south pole. It would join a club of just three countries so far to land on the moon (the US, China, and the Soviet Union). The lander (named Vikram, after Isro’s founder) is scheduled to touch down on September 6 or 7 between two of the moon’s craters, Simpelius N and Manzinus C.
The mission: There are three components: the orbiter, the lander, and the rover. The lander will search for water and minerals, and measure moonquakes, while the orbiter will spend a year taking images of the moon’s surface. The rover, which has a life span of just 14 days, will roam up to half a kilometer from the lander, taking pictures and collecting data to send back to Earth. Chandrayaan-1, the country’s first moon mission in 2008, contained only an orbiter and impactor designed to slam into the lunar surface.
Plummeting costs: The budget for the mission is just $141 million, indicating just how cheap it has become to send things into space.
Want to stay up to date with space tech news? Sign up for our newsletter, The Airlock.