I am not willing to make this prediction, but I’d be interested in hearing if anyone thinks we will NOT see fully autonomous vehicles on the road by 2030. Even in 2010, it seemed just around the corner — but that turned out to be wrong.
Technically, we already have fully autonomous vehicles on the road, they're just not generally allowed to be used as such and have obvious failure modes not shared by human drivers.
However, if by this you mean "autonomous vehicles that are clearly superior to reasonable human drivers under all reasonably frequent scenarios (including inclement weather) in any developed country" then yes, I feel very confident we won't see that in the next decade, probably two. I think the hyperoptimistic HN bubble is extremely far off the mark with respect to autonomous vehicles and always has been.
I thought of this after posting. Technically, there have been vehicles allow to operate fully autonomously on the road, yes. But I mean for non-research purposes, full-featured autonomous driving for consumer-owned vehicles.
I don't think it'll make much economic sense for consumers to actually buy them, but I do expect to see services accessible to the public that let you use an autonomous vehicle in places with effectively ideal conditions before 2030. I also don't regard that as a particularly significant advance in the broader sense of the objective because I think it's easier to get from zero to driving in Phoenix than it is to get from there to handling, say, traffic in Mumbai, both in technological and social terms.
2030 will at least have certain areas that are autonomous vehicles only, like maybe a forward-thinking region implements a fully autonomous truck lane on a highway.
If anything were to get in the way of fully autonomous vehicles by 2030, I would think it will be a government's lack of political will to create the infrastructure or regulation needed for driverless vehicles.
It doesn't matter if there are SOME self-driving cars on the road in 2030, or if they're level 4 or level 5. [1]
What matters is if they're economically feasible by 2030, and say have 100M or 1B customers. Is Waymo bigger than what Google search is now, or another self-driving company bigger than Facebook?
My answer is no for consumer applications, e.g. rideshare. I wrote in this thread:
Self-driving cars won't impact the average consumer in the next 10 years. It will continue to be cheaper to operate rideshares with human drivers in most parts of the world and most terrains/climates. It will make sense for commercial applications though.
[1] There are people in this thread arguing that this has already been achieved. I don't agree, but if the threshold is one car demonstrating one thing, then meh I might not argue.
Yep, there's no way they'll be widespread, and for the limited places they are used, they'll have a human ready to grab the wheel. Largest users might be trucking companies. I think the hype will fade this decade without any fanfare as the reality becomes clearer.
Sure. I predict that there will not be a fully autonomous vehicle widely available for sale to the public and allowed to function in FA mode on all roads, as of Dec 31, 2029.
("widely available for sale to the public": at least as available as a Tesla is right now.)
On the other hand, I predict that lots of cars will have emergency auto-assist features: collision avoidance warnings, lane-keeping assistance, seatbelt tighteners, and perhaps a last-action feature that evaluates whether it's best to slam on the brakes, veer sharply (and in what direction?) or try to accelerate out of danger -- and then applies that action without driver input.
I believe there will be L3 available in the next decade. A consumer-owned car which can take over driver-responsibility in limited settings. What Teslas autopilot promises although it currently is L2.
The interesting question is which company is bold enough to do it first and if it will survive the law suits afterwards. Tesla is the most likely.
L4/5 will realize in limited scenarios (senior communities, hub-to-hub trucks) but not as taxis in dense cities. In other words Waymo will still use safety drivers in ten years or close down.
> if anyone thinks we will NOT see fully autonomous vehicles on the road by 2030.
I don't think we will in any widespread way. Perhaps we'll see some in very limited deployments, both in terms of the number of vehicles and what roads they'll operate on.
my prediction is we will see them in geo-fenced localities that have a much more controlled driving experience, and will not be available in consumer cars beyond Tesla-style "drive assistance", but instead will be fleets managed by companies like Uber.
I.e. self driving Ubers will be geo-fenced to suburban areas away from unpredictable city conditions, and used for local deliveries (think pizza, uber eats) and perhaps long-haul highway trucks (which pretty much drive straight for fifteen hours anyway).
Self driving in every consumer vehicle will be in the 2030s.
