The reactor would mostly likely fall into the ocean, where it would be retrieved intact. RTGs are designed to survive a launch failure, and several accidents have
already happened, without any significant release of radiation.

A fission reactor that has been assembled, but never operated, does not produce much radiation. Enriched uranium and/or pure plutonium are not particularly dangerous (unless inhaled or ingested). It is the fission byproducts from actually operating the reactor that are dangerous. Even this minimal radiation could be avoided by using temporary shielding that is removed (possibly by a robot) immediately before the launch.

Not necessarily. Instead of shielding, you can use distance. Fission reactors produce little radiation until they start operating. So you launch into space, then separate your main payload from the reactor using a long conductive tether. Then fire up the reactor.

Several tools are being developed that can look inside your brain while you are being interrogated. These include EEG to measure your brains electrical activity, and fMRI to measure metabolism. So far these are too bulky to be used surreptitiously, but "they" are working on that. Invest in tinfoil.

It is funny that people bring up examples where split second response, and accurate steering control are crucial. These are exactly the situations where autonomous systems are the strongest, and will do better than a human 99% of the time.

No, it would be irresponsible to continue to trust humans once computers can do better. The computers don't have to be perfect, they just have to be better than humans. That is not a high bar.

None of these are particularly difficult for current autonomous cars. They have multiple sensors, including GPS, radar, camera, inertial sensors, and rotation sensors. Snow may interfere with cameras, but have no effect on the others. An autonomous car also has access to far more information than you do, such as exactly where the road is, the location of other cars, and the exact location of signs and mileage markers (this data is collected and saved as the cars drive).

Yes. Once autonomous cars are on the road, the advantages will be obvious, and the objections will fade away.

More likely the exact opposite: As preventable deaths are reported, that were caused by humans interfering, there will be a demand to get people out of the loop. Soon insurance companies will void your policy if you drive your own car.

Only a complete retard would think that it works this way. It doesn't just "kick out" of auto drive. It will always make a best effort to drive safely. There are, however, certain situations, such as dirt roads and construction areas, where it will recommend the human take over. If the human fails to do so, the car will continue to drive safely, or pull over and stop.

Autonomous cars have driven hundreds of thousands of miles on existing roads. So why do you think additional infrastructure is needed? It seems to me that the opposite is true: less infrastructure will be needed. Parking spaces can be narrower (passengers will exit before the car is parked), parking lots/garages can be smaller and remotely located, lanes can be narrower, road construction can be reduced as road capacity increases, traffic lights can be phased out, etc. Public transportation will become more more popular as it shifts from big, infrequent, inconvenient buses to small, on-demand, direct-to-your-door vans. The result will be fewer cars on the road.

You make it sound like they hire a salesperson to go around and market their patent to potential customers. Or maybe you think the customers search for useful patents to license and then contact the inventor. Neither of these scenarios is common. What is common, is for the NPE to just sit on the patent, wait for someone to independently come up with the same innovation, and then demand payment. This is not contributing anything positive to the process.

And an even worse example if you want to show that NPEs are useful today. That fact that independent inventors were useful a century ago is irrelevant. They play very little role in modern innovation. Many companies refuse to even talk to independent inventors, because knowledge of their patents can expose the company to liability. What most NPEs do is sit on the patent and wait for someone to independently come up with the same innovation, and then demand payment. They are just parasites.

It is easy to fathom if you look at how the program was structured. All the incentives were inverted: nearly everyone involved actually benefited from cost overruns (the contractors got more money, the bureaucrats had the prestige of managing more resources, and the politicians had more patronage to dispense). There was no accountability (no one is being disciplined or fined). There is not even any political fallout because the blame is smeared out over multiple administrations (Conservatives can blame Labour for starting the project, while Labour can blame the Tories for mismanaging the implementation). It is like it was designed to fail. A decade from now you will be reading about some other project that failed in the exact same way, for the exact same reasons.

As they say in China: Kill the chicken to scare the monkey.

Decades ago many people predicted that only the "super wealthy" would be able to own computers, and decades before that the same thing was said about cars.

“Automobiles are a picture of the arrogance of wealth ... nothing spreads socialistic feeling in this country more than the use of the automobile.” -- Woodrow Wilson

So far, there is no reason to believe that future robots are going to be particularly expensive. I already have a Roomba and a 3D-printer, both easily affordable by any middle class family. Prices for manufacturing robots are high, but falling quickly.