[FoRK] Your prediction, please

Ken Meltsner meltsner at alum.mit.edu
Mon Jun 24 10:53:18 PDT 2013


Relatively accurate predictions might be possible if we can break down the
target into more manageable chunks:

* First commercial "Google is my co-pilot" capability -- I think it's a
toss-up whether this goes into high-usage environment (long-distance
trucking) or high-value (police?).  I think this is the big entry point --
all of those UPS trucks crossing the country with highly-regulated drivers,
mandatory breaks, etc.  If you still have to have a driver in the seat,
that really limits you to applications where a second driver or a single
driver (with breaks), or a single driver on amphetamines would be more
expensive.

* First consumer co-pilot version ("Mercedes and Microsoft announce...")

And likewise for first unassisted (driver in car, not in seat), first
autonomous (no driver), etc.

I would guess that we might see first commercial co-pilot applications --
limited to specific long-haul, low-density highways -- within 15 years.  We
might get really limited applications, e.g.road train like slaved trucks
for military transport beforethan that, but I'd bet that the first civilian
applications will be for trucking in the Midwest and Mountain states.
Outside of the US and Canada, it might happen faster in countries like
Australia -- they already have multi-tractor road trains, so computer
piloting is probably less scary than the status quo.



Ken Meltsner


More information about the FoRK mailing list