The book "Range: How Generalists Triumph in a Specialized World" by David Epstein is also relevant. He argues that it can be very beneficial (across a wide range of careers) for a person to have gained some experience of several unrelated fields before specializing in one field. When it comes to innovations and problem solving, such people can sometimes reason "The problem I am trying to solve superficially looks like the [such-and-such] problem in another field, and that problem was solved by using [whatever], so perhaps that same technique might be applicable here". To an outside observer, such problem-solving skills might look like genius or spontaneous eureka moments, when in fact they are rooted in pattern matching or "making the connection" thinking.

From the top of my head...

Zip2, co-founded by Elon and his brother Kimbal (later, co-founder status was awarded to an early angel investor too), was a web server that combined maps (think Google Maps but years before that was invented), directions on how to get from one location to another, and a customer review service of businesses (think Yelp but years before that was founded). Elon did most/all of the web server back-end coding for the initial version.

What people know of as PayPal was the result of the merging of two companies: X.com and Confinity. The staff of both companies, including Elon (from X.com) contributed to the development of PayPal.

Elon played a key role in the design of Tesla's first car (the Roadster).

Elon is listed as co-inventor on several patents.

The original business plan for Tesla was written by Eberhard and Tarpenning, but Elon made at least one important contribution to it: deciding to sell cars directly to customers rather than use third-party dealerships. This has provided a significant competitive advantage for Tesla.

Nitpick: having read a few biographies of Elon Musk, I disagree with your claim that he: (1) "just funded existing ideas and hired the right people", but didn't (2) come up with any new concepts. He might have done more of (1) than (2), but it is incorrect to claim he didn't do any (2) at all.

It is relatively common for a person to be highly skilled at a core competency (programming, physics, art, carpentry, plumbing, cooking and so on) but to have mediocre skills, at best, in other areas, such as management/leadership, communication, marketing, sales, logistics, financial planning, customer service, and so on. This is one reason why it is common for start-up companies to fail: the product idea is great and the core skills required to implement it are present, but the other skills required to have a successful business are lacking. If you think of the Manhattan Project as being a start-up, then having enough people with the core skills (nuclear physics) was important, but the project would have failed if it didn't have the required non-core skills too. The tone of the interview questions in this posting fall into the trap of thinking that only core skills were important for the Manhattan Project and if Oppenheimer's skills in that area were not quite as good as those of others, then his contributions must have been negligible.

I agree there is some merit to the "Tesla is being irresponsible" accusation. However, here is a counter-argument which I illustrate with some made-up numbers.

Google suggests there almost 43,000 road fatalities in the US annually, but I will round that down to 40,000 to keep the maths simple, and let's assume that 90% of those deaths could be prevented by better-quality driving (assume the other 10% are caused by acts of god, equipment failures and so on). So that is 36,000 preventable deaths per year. Let's assume that eventually autonomous driving gets to be 10 times better than humans and becomes used in virtually all cars, so the 36,000 preventable deaths will drop to be 3,600 death. This means the 10X better autonomous driving will save 32,400 deaths per year in the US.

Now let's assume Tesla's approach to rolling out its buggy-but-rapidly-improving FSD results in an average of 100 deaths per year until the "10X better" goal is reached, and that Waymo's slower-but-safer rollout of its autonomous driving technology results in zero deaths per year. If Tesla's rollout approach results in the "10X better and widely deployed" goal being reached, say, 5 years sooner than Waymo's approach, then that will save 5 * 32,400 = 162,000 lives. 162,000 lives saved dwarfs an average of 100 extra deaths annually caused by FSD until it reaches 10X better goal, assuming that the goal is reached within the next few years/decades.

My understanding is that the above counter-argument is the rationale behind Tesla's approach to developing and deploying FSD. Whether Elon Musk is right in taking such an approach is as much a philosophical question as it is a technical one. Especially, when you consider that one dictionary definition of "irresponsible" is "showing a lack of care for consequences." The "trolley problem" (https://en.wikipedia.org/wiki/Trolley_problem) comes to mind.

My recollection is that when SpaceX started supplying Starlink free-of-charge to Ukraine, Elon stated that he/SpaceX was willing to do that for 6 months rather than indefinitely.

I look on Amazon and eBay for DVDs. I usually buy second-hand DVDs if possible, because they tend to be dirt cheap and 98% of them work fine even.

This link shows the history of Tesla's profitability: https://twitter.com/ICannot_En...

If you read some books on the topic of "emergence" (sometimes called "emergent behaviour"), you might be less sure in your prediction that Artificial General Intelligence (AGI) will never be achieved. Emergent behaviour has been encountered numerous times in the field of AI, and the assumption is that enough iterations of "build a bigger and better neural network to achieve yet more emergent behaviour" will eventually result in the emergence of something that is (or is indistinguishable from) AGI. I do not work in the field of AI, so I am not qualified to confidently state if/when AGI will be achieved. But having read some books on emergence, I know enough to keep an open mind to the possibility that AGI will be achieved.

The linked article does not mention the following, but another article does... The woman failed to turn up for the court case so a default judgement was awarded against her.

A book I am currently reading ("Solar Trillions" by Tony Seba) suggests that Concentrated Solar Power (CSP) is a better technology for desalination than photovoltaic panels. The logic is as follows... CSP is more efficient than PV in converting sunlight into electricity. In addition, a side product of CSP is a massive amount of waste heat, and this waste heat can be used for desalination.

A benefit of PV is that the technology can scale up/down to suit whatever the intended application is. For example, a desktop calculator I have is powered by a solar cell that is just 3.3 square centimetres in size. In contrast, CSP is feasible only at large scale. However, in places where desalination is needed, the need tends to be large scale, so the "does not scale down to small sizes" limitation of CSP is not relevant.

In the short-term (and where you live), you might be right. However, in the long-term (and more globally), virtual power plants could offer some nice benefits. The potential benefits rely on the assumption that the cost of solar panels and battery storage will get exponentially cheaper over time. If that assumption turns out to be correct, then instead of having centralised power plants that ship electrical power over hundreds/thousands of miles to consumers, each neighbourhood could have sufficient solar panels and battery storage to satisfy its own needs. This approach might provide some resilience in the face of natural disasters or war. Also, in a developing country, this might be a viable option for providing a remote village with electricity.

I think there are two reasons for Tesla cars not supporting vehicle-to-grid capability (yet).

First, the lifetime of a car's battery is related to how many charge-discharge cycles the battery undergoes. For example (using made-up numbers), if you get 200 miles per charge and you can recharge the battery 2000 times before it degrades significantly, then you have a battery lifetime of 200 * 2000 = 400,000 miles. But if you frequently use the car to power an off-grid home or a home during a blackout, then you will find that your car's battery needs to be replaced far earlier than 400,000 miles. That could spell a PR disaster, because many car owners would not realise that car-to-grid can shorten a car battery's life.

Second, I forget the technical name of the valve component, but Tesla vehicles use a one-direction rather than a two-direction electrical valve in the car, which saves on costs but prevents car-to-grid functionality.

In recent years, some significant improvements have been made to battery technology. For example, LFP batteries support about 3 times the number of charge-discharge cycles of lithium ternary batteries. And Tesla has been funding researchers who have been making progress towards a "million mile battery" and even a "100 year battery". As the expected lifetime of batteries gets to be significantly longer than the expected lifetime of a car, it is reasonable to assume that more and more car makers will support car-to-grid, since the battery will still likely outlive the car.

I don't know much about satellite companies, but TFA suggests that Eutelsat's geosynchronous satellites are not suitable for internet communication (I'm assuming they are in a too-high orbit to provide low-latency communication), and that OneWeb's ambitions are to provide only 648 satellites initially, which is less than one quarter of what Starlink already has in orbit. I know from elsewhere that SpaceX (the parent company of Starlink) has a reputation for being able to design, build and launch space equipment significantly more cheaply than every other space company in existence, so it seems reasonable to assume that Starlink is likely increase its lead over its already-far-behind competitors. As such, the answer to your question:

is that they should not enter (or continue to compete in) a market if doing so if likely to be a case of throwing good money after bad. Perhaps there is evidence to suggest this isn't a case of sunk cost fallacy but, if so, such evidence is not presented in the article.

I suggest you read Ashlee Vance's warts-and-all biography of Elon Musk. In that book, you will see plenty of evidence that Elon Musk is very intelligent, is a workaholic, and has played a hands-on engineering role in his companies, including SpaceX. In addition, a trivial Google search will provide you with a list of patents for which Elon Musk is listed as co-inventor. Elon Musk may be a jerk, but being a jerk does not negate the fact that he has played a hands-on engineering role in his companies.

From various sources I have read, it seems Elon Musk does have a high intelligence. However, the words of Thomas Edison are relevant: "Genius is 1% inspiration and 99% perspiration". Elon Musk's success owes a little to his intelligence, but a whole lot more to his dedicated hard work. The various sources I have read confirm that, aside from his intelligence, Elon Musk is also a workaholic.