The next war already happened btw. The Azerbaijani government used a large number of Turkish-supplied attack drones (self-destruction drones too iirc) against the Armenian forces.
Self-driving: Can you elaborate on this a little more? You mean having self driving trucks on supply routes makes it much harder to break supply lines?
EVs: I think this one is already mainstream. Not in the sense that most people already have or are buying EVs but everyone from car manufacturers to bureaucrats are on the train already and it is only a matter of time.
Self-driving: Right now, the biggest problem with self-driving car adoption is that when anyone dies at the hands of a self-driving car, there's a huge press outcry and the whole program gets shut down. That calculus changes dramatically if the other side is actively trying to kill you. Then, it becomes a huge advantage to not even send drivers into the war zone, and resupply with automated logistics lines. Instead of becoming an invention that potentially kills people, it becomes an invention that saves lots of people, because the baseline shifts from "nobody ever dies young" to "young people are actively trying to kill each other". It also gives a huge advantage to the side that adopts it, because they can resupply as quickly as they can build materials instead of losing costly human lives in the effort, and the vehicles themselves can often follow evasive routes that wouldn't be practical for a human crew.
Drones (like nuclear weapons before them) make war a lot more expensive. I guess only the top 5-7 armies in the world will have the technology and expertise to maintain drones. So a conflict like the one in Iraq will have a swift conclusion, but the calculus among the big nations will remain mostly unchanged IMHO.
Self-driving: I think the biggest problem with self-driving in this scenario is the very moniker "self-driving". But it is very likely there are vehicles being tested today that are programmed to avoid minefields, use sophisticated path planning algorithms to avoid detection, minimize fuel consumption etc. Sort of a jeep-sized roomba that traverses a region instead of your living room.
Drones: My 3-year-old will ask me "Daddy can you fly the drone now?" and absolutely loves chasing it. Great toy now, starting to find some industrial uses with insurance/forestry/firefighting/etc. Will absolutely revolutionize warfare when the next war happens.
Self-driving: Another wartime technology. I think this will limp along in niche markets until then, and then explode (no pun intended) when we need to maintain logistics networks under armed threat. Likely to be a cornerstone of post-war transportation networks, but not in the form they are now. Rather than struggle with sensor algorithms, we'll likely build the sensors & coordination into the road infrastructure and ban manual driving.
Robotics: This is like AI in that when it succeeds, it's no longer called "robotics". It's already used very extensively in manufacturing and logistics (have you ever looked up airport baggage handling systems on YouTube?). We aren't going to get general-purpose robots, but we might see a wave of micro-manufacturing appliances a la 3D printing (put in raw materials, get back a product made to your specifications, distribute products digitally) as well as descendants of today's household appliances (we already have this with Roombas, robotic lawn mowers, robotic tractors, etc).
EVs: Will probably go mainstream around 2023-2024. If you're already a top-notch electrical engineer you can likely get very rich on this right now by founding a company and letting a SPAC take you public next year. For the rest of it the excitement will be in buying one and upgrading our solar panels, power wall, power grid to support it.
Climate tech: Will likely ultimately fail in stopping global warming but will invent lots of useful innovations that might be applied to other industries in the future (sort of like how microwave ovens came out of radar and computers came out of artillery guidance). If you have good skills and a long time horizon, the play is to go work for one of these companies, learn everything you can, spend your time researching, and then ride it until it goes bankrupt and then apply the technologies toward something more mundane.
I'd agree that we're not in 2005 now - it feels more like 2002-03, where you can tell that there's interesting stuff going on but nothing's really ready for primetime yet. We had stuff like Friendster, Kazaa, LiveJournal, BaseCamp, then - all precursors to really important stuff, but the specific successes hadn't yet happened.