Hacker News Reader

Has anyone seen anything to suggest this will happen for sure in 10 years? I’m sure there will be improvements in safety with these features but there are basically zero AI demos out there that are what I’d call good. For example Siri and even Google Assistant quickly show how limited they are after years and years of real world use. If these rely on the same Deep Learning technology won’t it just be the same limits apply to this much harder problem?

View on HN

Siri/Alexa/GoogleAssist etc. work well in constrained situations (e.g. changing music, turning stuff on/off, pop culture quiz questions), they have trouble parsing more general/complex questions.

Many researchers I've talked to believe the initial self-driving car deployments to the public will be have to be very constrained: fixed routes, possibly isolated self-driving car lanes, only allowed in clear/daytime conditions.

And its probably going to be here earlier than we think.

Essentially train carriages on tarmac :)

Why should driving be a harder problem then being a universal assistant (Siri's promise)?

Siri's potential problem field is huge, driving is a well defined, limited problem field.

Reply on news.ycombinator.com