AI Software Just Learned to Make AI Software

Due to recent advances in AI and a process called "deep learning," machines have begun to learn like humans do.

Learning how to learn is the method that these machines are taking on.

“Currently the way you solve problems is you have expertise and data and computation,” said Jeff Dean, who leads the Google Brain research group, at the AI Frontiers conference in Santa Clara, California. “Can we eliminate the need for a lot of machine-learning expertise?”

Meaning, can we bypass the need for programmers.  It looks like it.  It looks like machines will be doing much of the coding in future machines.  At this point, we're at the behest of computing power.

The process is being researched at a conglomeration of places, including  the nonprofit research institute OpenAI (which was cofounded by Elon Musk), MIT, the University of California, Berkeley, and Google’s other artificial intelligence research group, DeepMind.

The more realistic this becomes, the easier it will be to move forward faster in self-accelerating ball of technological breakthrough.

“Easing the burden on the data scientist is a big payoff,” says Otkrist Gupta, a researcher at the MIT Media Lab.  “It could make you more productive, make you better models, and make you free to explore higher-level ideas.”