Crowd funding site here with compelling video. The timeline at the bottom of the site page seems to indicate it’s just a matter of designing and slapping the parts together and away you go with a $ 500 “family robot”.
Is this level of voice recognition accuracy and inter-activity possible today or not? I often have a very hard time getting my Samsung Galaxy S-3 smartphone speech to text feature to understand what I am saying accurately, and this little robot is chittering away merrily with everyone with very advanced natural language recognition skills and (seemingly) understanding what they are saying in context.
Is this level of AI performance possible (today) or is this going to be a scam or an under-delivering boondoggle?
Watching that video all I could think was “yeah fucking right.” Face-tracking to record video and take pictures is do-able, hell, even Facebook can tell where people are in a photo, but non-staged interactions that are as natural as the ones it had with the woman working with the dough and the kid reading the book aren’t gonna happen.
Is it possible it’ll work better than all the existing voice-recognition stuff out there? Sure. But there’s a wide gulf between that and what they show in the video. If they ever actually release this thing I think we’ll find it’s a stark contrast between the reality of it and what was promised.
How can they promise that they will have this technology all worked out by a specific date next year? Either they’ve got it down, or they don’t. Also, why would I have to wait until August of next year for a $27 t-shirt I pay for now?