I see your point now. Tesla has a fully autonomous car in the lab with a different sensor suite as well. The current Model S isn't that.

There are other companies with more or less the exact same technology: Mercedes and Infiniti that I've driven or owned and I think one or two others. The only real difference between the systems is (1) Tesla doesn't require you to hold the wheel all the time and (2) it actually works. The Infiniti system ping pongs back and forth between the lane markers like a drunk to such a s degree that no one actually uses it. The Mercedes system is better, but still far less capable than Tesla's. Are Honda and Daimler also desperately trying to prop up their stock?

I've driven a couple thousand miles up and down the East Coast on autopilot without experiencing those kind of problems. Its not unusual for it to drive for me for an hour at a time without touching the steering wheel. If you look at one of the polls on TMC, you'll see that the percentage of owners who report being satisfied or very satisfied is somewhere north of 85%

Tesla has a better level 3 system in production than anyone. In terms of the speculative future car with full autonomy, they have one in the lab (same as everyone else) and engineers I know in the AI world think that their data collection has the potential to be enormously significant.

I've actually met some of the engineering team doing the work for Tesla. They have some very serious folks. The same phenomenon that attracts fanboys to Elon appears to attract young engineering talent.

They have the most advanced system in the field and I'm pretty convinced at this point that you have no special insight into the competition in the R&D world. They might win or they might lose. I do agree with you that the fact they have gotten there first with some features doesn't prove they will get to full autonomy first and I agree that the press and public aren't always very well informed.

If a competitor gets there first with full autonomy, that doesn't really destroy their edge altogether. There is still the whole stupid-fast electric car thing.

You guys might find Chris Urmson's TED talk interesting on subject. He strongly believes that the Tesla approach of attaining automating through iteration is fatally flawed and makes a pretty good case for why. The interesting thing that except for the highly secretive Apple program (staffed in large part with ex-Tesla guys who are apparently the B team) everyone else (Tesla, Daimler, GM, Ford, Nissan, Toyota, Honda, Volvo) are going with the iterative approach.

I don't see it that way. He says a SDC can do things a driver assist/pseudo autonomous algorithm cannot.

The identification/driving algorithms are still iterative in nature.

Not sure I agree with this. I think it has to be close to perfect because while people accept tens of thousands of driving fatalities each year now, I suspect that they'll have a much bigger problem with even a single serious SDC accident.

Software commoditizes hardware; and data commoditizes (ML) software.

Definitely this. People are very willing to blame individuals in mistakes made by human drivers, but one fatality from a computer driver will lead to massive lawsuits no matter the relative safety rates.

Yes, we have ToothSayer's armchair analogizing to Donald Trump ITT---an important data point.

On the other hand, we have veteran journalist Ashlee Vance, who spent hundreds of hours interviewing Elon, his (ex)-wives, coworkers, etc. and thousands more researching Elon's life while writing the (as of yet) definitive biography, who thinks Musk is best captured as, "Profoundly gifted---in the clinical sense." [180+ IQ, 1:10^6 population prevalence, overwhelming internal drive to tackle existential-scale problems].

Who can say which is the more reality-based take? Is there information about the world not loss-lessly encoded in the relations of stock tickers and press releases? Who really knows?

I don't see how there is any disagreement. The legal framework is already lagging and there will be a SDC before there is a full system of laws to regulate them.

Obviously there will be lawsuits. It is America. People are very stupid.

Sub,

You realize no one thinks Elon Musk is dumb? People claim he is full of ****, because he says things that are obviously bull****. Lets take SpaceX and his roadmap to a two order of magnitude cost reduction. It appears one of his cost saving measures was to outsource structural components without testing. This is great until your spaceship explodes.

Elon is a weird guy in the sense that he is brilliant, but holds profoundly idiotic opinions (AI viewpoint) and speaks nonsense frequently. People struggle to rejoin the fact that he knows when he is lying and does so for an explicit purpose. If I'm not mistaken he'd have been bankrupt now with Tesla/SpaceX both folded if he has not made a completely outrageous lie that was believed by a VC. With his record of success at dishonesty, why stop?

well if Elon's hagiographer has a high opinion of him who are you to disagree?

you're reinforcing TS's point about the cult of hero worship around Elon. somehow you twisted very specific criticisms of Telsa to a referendum on whether or not Elon is smart.

yea, he's smart. TS said that. what about everything else he posted?

This is obviously false. The last time tesla was in a major bind was during delays with the model s ramp, and Musk all but inked a pretty impressive deal with google allowing him to continue on the same path and having google pony up for the gigafactory. He walked away from it when good reviews started pouring in and the stock bounced. So when all the shorts were hyperventilating about tesla being on the verge of collapse/bankruptcy during the model s ramp issues, the reality is Musk had already set up a pretty decent worst case scenario via a google buyout.

So just to clarify, your position is that if a person invests large amounts of time researching someone else in order to write a book about him (in journalistic circles, this is called a "biography"), that person is necessarily a sycophant?

As of his last several posts he's transitioned into asserting obviously false characterizations of Tesla autopilot/autonomous R&D as facts, so I have no idea what an appropriate response is.

Mikhel05 -

And same with you, I cannot imagine a way to penetrate your Dunning-Kruger shroud.

For example, I don't know what "profoundly idiotic" AI viewpoint you think that Elon holds, but the fact you feel qualified to judge his views as such makes further discussion sort of irrelevant.

Anyways, when Alan Mulally came to Detroit in '06, he summarized the situation as: "[Ford, GM, and Chrysler] have been slowly going out of business for eighty years. They were insulated and had great success early on...But they were arrogant. They made fun of the Japanese. And then they made shoddy products, all of a sudden they had some competition, and their arrogance caught up with them." (Vlasic, Once Upon a Car, p. 170).

I don't see any evidence these companies will ever achieve a product-focused culture. Along with their misaligned incentives relative to electrification and autonomy, they're walking dead.

Tesla and others (Google/Apple/Faraday?) who should have never had the chance to get even a fingerhold of market share, will instead have plenty of entry points to leverage for capital deepening. Anyone who sustains a product-focused culture with the right incentives to innovate on modern technology platforms is going to do very well.

http://www.bloomberg.com/graphics/20...n-musk-spacex/

You should brush up on your history before commenting again. While the story refers to Musk's wholly fabricated story as a "bluff", we can quibble about the semantics of what is making something up whole cloth.

I generally think any view, that has no theoretical support, in which you have limited experience, and which the entirety of the field of experts believes the opposite is profoundly idiotic. (If you are, in earnest, not aware of how ******ed his views on AI are. You are probably just not familiar with the subject in any serious way. There is no theoretical roadmap to AGI from ASI. If you really believe that Amazon and Google are destroying the fate of humanity by helping to recommend products and show me highly targeted ads, then I guess you'd fit right in with Elon.)

However, you seen to think because Elon said it. It is true.

"This system I have in mind, how would you like something that can never crash, is immune to weather, it goes 3 or 4 times faster than the bullet train... it goes an average speed of twice what an aircraft would do. You would go from downtown LA to downtown San Francisco in under 30 minutes. It would cost you much less than an air ticket than any other mode of transport. I think we could actually make it self-powering if you put solar panels on it, you generate more power than you would consume in the system. There's a way to store the power so it would run 24/7 without using batteries. Yes, this is possible, absolutely."

Would you characterize that as profoundly idiotic? Totally insane? The ramblings of a person who knows that his massive army of sycophants will ignore any semblance of reality when defending the Church of Elon?

C'mon man.

I've read the book that article was based on. And you don't know what you are talking about. Spacex was never at risk of bankruptcy post successful falcon 1 flight and subsequent ISS contract. And the book mentions Musk sort of "bluffed" one hold out VC during a emergency fund raising round, but that VC was holding out because he knew they could flip tesla for a good profit to early investors at that point - as soon as Musk threatened to personally borrow money and fund the round all by himself that VC backed down and joined the funding round.

Mihkel05, ok...I can't quit you man. Let me ask you this. Are you aware that a majority of the world's leading AI researchers expect to see ASI in our lifetime? (In the sense of "outperform humans on all cognitive tasks.")

If you are not aware of this, why are you not aware? If you are, why do you write things like, "There is no theoretical roadmap to ASI through AGI?"

Why would I need to characterize it that way? Wikipedia lists a number of serious attempts to assess hyperloop feasibility. And apparently there are two companies trying to commercialize the hyperloop.

But I don't know if there will be a definitive proof-of-concept/proof-of-impossibility in the near future.

I'm posting quotes from Tesla users. I'm pointing out the facts that demonstrate they are obviously NOT ahead of anyone, and are in fact years behind.

Frequently following off ramps is such a simplistic idiotic bug that it proves their software is **** and they've done next to no research. Proper spatial mapping, proper use of mapping/GPS, or just simple road surface mapping or look-ahead boundary mapping would make this problem disappear. They've done none of those simple things that others have solved many, many years ago.

And that's just one of the bugs. The car following algorithms they have (which are as easy in the SDC space as coding "hello world!"), appear so simplistic as to be dangerous, occasionally lurching toward other traffic. There are far more.
It's an apt description. I contend it bothers you so much because it has strong elements of truth. If I compared him to Putin, or Ellen, would you care?
Just curious, but how does this look different from:

I know a couple of sociopaths, and many of the people they know gush about them, despite no genius or talents despite manipulation.

The truth is that if Musk had a 180+ IQ, and genuinely wanted to save the world, he wouldn't have Paypal, a solar company or a car company. None of them make a lick of difference to the world.

Musk's modus operandi is in entering fields where he can avoid commercial realities and suck up government money. His recent talks about establishing a factory in Germany are about getting handouts.

You can see he's not genius level smart by lots of things, but an interesting example is his plans to colonize Mars, which are simply nutty self-aggrandization (his nuking of Mars idea is the height of assclownery and indicates by itself his IQ is sub 140). Nothing he does is going to make a lick of difference to the colonization of Mars. That will be decided by materials science creating new materials suitable for establishing a colony; robotics, and new energy sources such as fusion, something Musk has nothing to do with. The real heros and geniuses are working on these technologies right now (and better batteries), and changing the world without the spotlight.

Musk has made similarly dumb comments/opinions with solar (the wrong energy source to bet on, that makes the world a worse place), and electric cars (Musk will make zero difference to the time or rate of mainstream adoption, which will be decided by battery science operating under existing commercial and research pressures, to which Musk makes no difference).
His frantic management of the PR cycle, including by dishonesty, shows what kind of person he is and how real he is. He's a brilliant businessman and self promoter. He's talented at making things happen. He has a decent physics brain that's been honed in real world businesses and engineering efforts. That's all there is to him. He's no genius; his many idiotic statements and beliefs make that obvious.

As for his following, he has a kind of formerly-beaten-child, downcast-eyes type humility (with a core of strength and narcissism and a sprinkling of self deprecating intelligence), a combination that's as powerful and sexually stimulating to tech lovers in search of a hero, as women with tiny bound feet were to Chinese noblemen. All critical thought goes out the window.

Um, ok. You describe Autopilot as "crapware" and "****". In order for this to be a reality-based characterization, there would need to exist competing driver assistance systems that are much better than Autopilot.

But oddly, no such superior systems exist. To learn this, you could have read NoSoup4U's account above; or any of the scores of highly favorable reviews on TMC; or just ****ing googled "Tesla autopilot review". Here is one of the first review summations I got, from the Car and Driver Blog:

And even more oddly (for your "crapware" thesis), Tesla is the only company that has built the correct framework for rapid iterative improvement of its driver assistance software: Fleet Learning + Over-the-Air Updates.

So while Autopilot is, right now, only "venturing beyond Mercedes-Benz S-class Intelligent Drive" in a few areas, within a few product cycles it will vastly outclass every other driver assistance offering.

You are basically operating in a fact-free zone of Trump-ian proportions here.

Ok, you seem to have extensive firsthand experience with con men, sociopaths, and demagogues. Such individuals are quite frequent and it's unsurprising you know some of them. Profoundly gifted people are, by definition, incredibly rare and you probably have no experience with them. In any case, in reality, there is considerable evidence that Elon Musk is the latter and not the former.

In reality, for example: "The most striking part of Elon's character as a young boy was his compulsion to read. 'It was not unusual for him to read ten hours a day,' said [Elon's brother] Kimbal. 'If it was the weekend, he could go through two books in a day.' The family went on numerous shopping excursions in which they realized mid-trip that Elon had gone missing. Maye or Kimbal would pop into the nearest bookstore and find Elon somewhere near the back sitting on the floor and reading in one of his trancelike states...Elon, in fact, churned through two sets of encyclopedias [around age 10]...The boy had a photographic memory, and the encyclopedias turned him into a fact factory. At the dinner table, Tosca would wonder aloud about the distance from Earth to the Moon. Elon would spit out the exact measurement at perigee and apogee." (Vance, Elon Musk, pg. 34).

Characteristic behavior for the profoundly gifted.

Well in reality he does have a 5/6-sigma IQ, and those are the fields he decided to pursue. As for basic research, he was accepted into the Stanford PhD program for applied physics (intending to concentrate on battery research), but then decided he would have a bigger influence as an entrepreneur.

I definitely agree that no one person can change the world by themselves. And I also agree people have a too-strong tendency to canonize their leaders, especially charismatic leaders.

But I'm also increasingly confident that you do not have a good sense of how smart the best people really are. You need make an authentic---painfully authentic---effort to try to do good work in hard field (say research-level math/physics/science), for thousands of hours, before you can get a real handle on the difference between a 3-sigma smart person like yourself, and a true genius. (Such as Elon Musk, like it or not.)

So bottom line, I think if you realized that Elon is actually much, much smarter than you, you would be a lot more circumspect in how you evaluate his strategic decisions*.

But then you'd have to relinquish this fiction of him being a Trump-like figure who just happens to resonate with the sexual proclivities of the geek set instead of NYC supermodels.

And maybe that's a negative utility narrative swap for you.

* Which, suboptimal or not, will probably be good enough to win against the US auto incumbents---in a big way.

You're way off base here. According to Subfallen's Ashley Vance ( the great biographer ) link, Elon does indeed have talent. He can pick up a 350 pound sumo wrestler. See 24:50 into interview.


You're again not seeing the big picture. Elon wants to put "a million people on Mars, not just five." The numbers make the difference. See 15:55 into Subfallen's (Vance) interview. Also, Musk has read many comic books and every science fiction book known to man. See Vance interview at 17:00

Well yeah of course, but that's what makes him great - his horrible father. See Vance interview at 22:00.

Now in all fairness, Vance admits he sounds like a fanboy, but hey, Musk has "stamina and resolve, and an intensity of will," according to Vance. So it's ok to be a fanboy. If you don't believe this, then see the Vance interview at 25:40

You're way off base here. According to Subfallen's Ashley Vance ( the great biographer ) link, Elon does indeed have talent. He can pick up a 350 pound sumo wrestler. See 24:50 into the Vance interview.


You're again not seeing the big picture. Elon wants to put a million people on Mars, not just five. The numbers make the difference. See 15:55 into Subfallen's (Vance) interview. Also, Musk has read many comic books and every science fiction book known to man. See Vance interview at 17:00

Well yeah of course, but that's what makes him great - his horrible father. See Vance interview at 22:00

Now in all fairness, Vance admits he sounds like a fanboy, but hey, Musk has "stamina and resolve, and an intensity of will," according to Vance. So it's ok to be a fanboy. If you don't believe this, then see the Vance interview at 25:40

For those who missed it, here is Subfallen's, or I should say Vance's "best captured" essence of Musk, and Subfallen's evidence that Vance is a biographer and certainly not a sycophant.
https://www.youtube.com/watch?v=EEDDVbnYoP8

Holy crap, I ended up with two back to back almost identical posts. Ha, it was due to my editing. Sorry guys. The edit time ran out when I realized the mistake so I couldn't delete one of them.

yukoncpa -

Never apologize or acknowledge mistakes; not Trump-ish at all.

Actually the evidence that Vance is a really standard, experienced journalist who wrote a biography of Elon Musk is here. Enjoy!

(Or am I the only person ITT who still reads books? Starting to feel that way.)

Though BFI is clearly but a vehicle for onanistic performance art, so if you get off on apologizing, push through those beta feels and grovel for us.

Subfallen, the autopilot thing is just hilarious. The point is that if other companies put out what they've been researching and refining for years, they would crush Tesla in features. Their cars certainly wouldn't try to take ramps, or swerve toward passing traffic, thinking it should be followed(!), or fail terribly at parallel parking unless there's huge amounts of space. Look at where Nissan is at in their research. The car companies don't put these features out, because the tech isn't ready, it's dangerous for mass consumption, and unlike Musk they're not desperate to keep the stock levels up, or insane enough to use their customers (and other road users) as crash test dummies such that hands are allowed off the wheel when the tech is far from ready.

That Tesla fails at the most basic things - following a highway without taking erroneously taking the off ramps, not lurching toward other traffic!! - that were solved 10+ years ago in multiple ways, not just one, is simple and definitive proof that they are many years behind.

This is not even up for debate. It's simple fact.
When has Musk ever done research? He's a businessman, like Trump. He puts time into understanding the background of what he's working on, and thinks big, but he doesn't do the research or the hard work of actually making rockets fly or batteries work. He motivates and enables people who can, just like Donald Trump.
Bezos is a 1 in a million genius. Gates is a 1 in a million genius.

Musk is very obvious a sub 140 IQ with Aspergers and a strong drive.

I enjoyed your post, but to be serious, this is certainly what Musk has. He is deeply driven with an almost religious zeal (typical of those with profound narcissism). Strong drive + a physics background + narcissism + Aspergers + deluded grandiose thinking and 140- IQ perfectly describes Musk. People often make multiple attribution errors when it comes to driven, passionate people, just as they with good looking people. Subfallen seems to have actually fallen in love though, which can't be good for your objectivity.