Low power AI chips getting better results than deep learning with a fraction of the data and processing power needs

I came across this startup: Another Brain claiming to design low power chips with embedded AI that give better results than common deep learning implementations. All with limited amount of data and very little processing power. What do you think? Is it real or too good to be true? They did raise 10M but given the hype around anything AI related, I'm wondering.

Submitted May 03, 2018 at 03:26PM by Luzututhun https://ift.tt/2jnavNY via TikTokTikk