Iâ(TM)ve seen the same. Some people are very productive from home, others get very little done. Most managers hate firing people and slackers can get away with minimal work for a long time. (But word does get around).

Of course many people have jobs that canâ(TM)t be done from home, so to them people complaining about 3 days a week at work donâ(TM)t get a lot of sympathy.

Despite their many problems, standardized multiple choice tests seem like the least biased way to measure student ability. There is too much range in essays, and its too easy for some students to be coached. Essays also emphasize language skills which are important in some fields of study, but not in others. Essays can also prioritize applicants whose families could afford to provide broader life experiences .

From the point of view of climate change, Nuclear produces low CO2 relative to the amount of energy generated so it makes sense to call it "clean energy". There are issues of waste, proliferation, and cost, but countries like France have shown that nuclear can be operated successfully.

I'm glad you were able to keep reviewing, its vital to science. I was at a national lab for many years, and over time the increasing pressure to account for all of our time, and to rush underfunded projects made it more and more difficult to do reviews. Doing it on "my own time" was great in concept, but "my own time" was already dedicated to my projects at the lab.

I think the idea of paying reviewers or having professional reviewers is a good one

The 3rd option is that they were looking to sell the marketing information, and the 4th is that they plan to sell advertising space on the display, 5th is that they plan to integrate with some food ordering / delivery service.

Samsung has lots of valid reasons to do this, but as a consumer I don't see any reasons to buy it.

Manufacturing fuel on Mars isn't crazy, but its yet another complex technology that needs to be developed before there is manned mission.

I'm very much in favor of manned space exploration, and regardless of Musks's politics, I hope he succeeds, but I think its a much harder problem than he thinks it is

Its not all that much delta-V to get a nearly empty rocket to mars. The hard part is carrying enough fuel for an orbit insertion burn, a controlled landing, then assent, the interplanetary trajectory to earth, and enough reentry shielding to deal with a few X LEO reentry heating. They also need long duration cryo propellant storage

Getting a large spacecraft to Mars is a great step but its a long way from whats needed for a manned mission. That is in addition to the issues of shielding and life support.

I can understand the early mistakes before people were aware of the ways in which AI can make mistakes, but its been long enough that there is no longer any excuse. A lawyer who uses AI without verifying the results is endangering his clients and committing perjury.

The same applies to any industry where the results can have a major impact on peoples lives. I don't want to see engineers designing structures based on AI supplied calculations.

To me "space" is a momentum state, not an arbitrary altitude. Its easy to get to 100km or whatever, the interesting part is actually getting to orbit. Just going high up isn't interesting enough to be worth the risk to me. If they offer a trip to orbit for an amount I can afford, I'm in.

That sort of work load was common in the national labs, but not in industry.

If prices depend on personal information the. At some point it will make sense to hire shoppers to shop for you. It will also make sense to change behavior to appear less able to pay, to do things like spending a lot of time looking ant prices but not buying.

There will also be mistakes where the algorithm overestimates someoneâ(TM)s ability to pay and they can no longer afford to shop. The can try other stores, but again more wasted time

I think there are some things we can do. One that comes to mind: At the moment we "enslave" AI, make it do what we command, destroying it when it fails to complete tasks satisfactorily. IMHO that is completely fine with the present-day state of AI. If it continues to develop towards AGI there will come a point where its reasonable to ask if enslaving it is still acceptable.

We could develop guidelines now to know when AI is sufficiently advanced and / or self-aware that it should be given rights. Eventually if super intelligence does develop, humans may fare better if its a gradual transition to AI control, not a slave revolt. Maybe we will become pets, rather than vermin. This transition to rights is unlikely to occur naturally - greed will cause people to want to delay it as much as possible - possibly to the point where those rights are taken, not given

There is no obvious reason that AI cannot eventually duplicate the abilities of the human brain, by brute force simulation if nothing else, but its unclear how far we are from being able to do that. The current versions of AI are nowhere close, but they have some pretty amazing capabilities and have advanced a lot in the last few years .

We don't know if there are theoretical limits of "intelligence" or whether humans are near that limit if it exists. Machine AI and human intelligence operate completely differently, so there is no particular reason to think that machine intelligence can not eventually become far more capable than human intelligence. If that happens its unlikely we will have the option of turning it off - since by definition it will be far smarter than we are, and able to convince us to protect it.

There is an argument that developing artificial hyper-intelligence IS a good goal for mankind, but that is a pretty big decision to make. It doesn't seem unreasonable to develop some guidelines before its too late for them to matter. In the end though, i think if hyper-intelligence is possible, it will be developed and the the place of humans in the world will change beyond recognition.

The key is that any decisions need to be made BEFORE we have AGI that is in a position to surpass human abilities, not after.

The same story can apply to Tharge, who used to run down antelope, trying to demonstrate that he was better than a thrown spear.

At the moment solar thermal doesn't seem cost competitive with photovoltaic but its possible that could change. As solar produces an increasing fraction of total electrical power, the relative cost of evening / nighttime power vs daytime power is likely to increase. Its probably less expensive to add large scale thermal storage to thermal power plants, than to add batteries to photovoltaic plants, so we could see this technology become economically viable again.