I think that's the only way forward. Ring fencing robot axis in certain areas at least minimizes the danger of either a hack or a fatal accident since you can insulate the area.
There is also the issue of latency since even a 0.1s delay in geolocation can have serious consequences in an evasive maneuver since the map inside the AI brain has to be accurately cross-referenced with where the car is in relation to that. IMHO true driverless cars are many years in the future except in small geofenced zones, as u point out.
Good article. The main questions I have wrt FSD are: a) the legal ramifications of a fatal accident, where the system is being used (open or closed), was responsible, b) how easily can FSD systems be hacked by a shady actor to cause mayhem? As we know no system is ever foolproof and outside of limited scope use (such as certain cities wrt taxis, long distance highway driving) where the known unknowns are programmable (eg differentiating b/w a motorcycle and cyclist when taking evasive action) and the unknown unknowns are limited, my view is that FSD has a long way to go before becoming an everyday reality.
Good questions. Several years ago Uber AV was involved in a fatal accident that was a precursor to Uber shutting down its AV project. Legal liability and insurance are two important outstanding questions. The questions also lead me to expect that in near term both Waymo and Tesla will roll out services in predefined, ie geofenced, areas. Cybersecurity is an issue. No reports I am aware of yet involving Waymo, however there have been hackers that have shown off their ability to hack a Tesla.
I think that's the only way forward. Ring fencing robot axis in certain areas at least minimizes the danger of either a hack or a fatal accident since you can insulate the area.
There is also the issue of latency since even a 0.1s delay in geolocation can have serious consequences in an evasive maneuver since the map inside the AI brain has to be accurately cross-referenced with where the car is in relation to that. IMHO true driverless cars are many years in the future except in small geofenced zones, as u point out.
Good article. The main questions I have wrt FSD are: a) the legal ramifications of a fatal accident, where the system is being used (open or closed), was responsible, b) how easily can FSD systems be hacked by a shady actor to cause mayhem? As we know no system is ever foolproof and outside of limited scope use (such as certain cities wrt taxis, long distance highway driving) where the known unknowns are programmable (eg differentiating b/w a motorcycle and cyclist when taking evasive action) and the unknown unknowns are limited, my view is that FSD has a long way to go before becoming an everyday reality.
Good questions. Several years ago Uber AV was involved in a fatal accident that was a precursor to Uber shutting down its AV project. Legal liability and insurance are two important outstanding questions. The questions also lead me to expect that in near term both Waymo and Tesla will roll out services in predefined, ie geofenced, areas. Cybersecurity is an issue. No reports I am aware of yet involving Waymo, however there have been hackers that have shown off their ability to hack a Tesla.