HN Theater @HNTheaterMonth

Great points. An interesting addition is that Musk has also stated that the tunnels would be vacuums (to 5-6 atmospheres) in order to seal against the water table.[1] It's interesting to consider the impacts on fire safety, both in the reduction of risk of spread (a positive) and the difficulties it presents with evacuation from compromised vehicles (a negative).

[1] https://www.ted.com/talks/elon_musk_the_future_we_re_buildin...

⬐ URSpider94

Can't watch the video now, but are you saying they are vacuums or pressurized? If they are vacuums, the highest vacuum you can have is -1 atmosphere, I.e. No air.

5-6 atmospheres would be PRESSURIZED, not vacuum.

Also, a vacuum would pull in water from the water table, you'd need pressure to keep it out.

Vacuum would be good for fire prevention, things can't burn if there's no oxygen. Pressure, on the other hand, will greatly amplify fire, if you just pressurize air. If you want to avoid that, you have to pressurize with a neutral gas like nitrogen, which would be expensive if there is any regular amount of gas leakage.

In either case, you'd have to build compartments that were pressure/vacuum proof (think a submarine or spacecraft) -- and in either case, leaving the pressure vessel would pose dire risks to occupants.

I'll try to watch the video when I'm back on WiFi.

⬐ trendia

The minimum pressure of a perfect vacuum is 0 atm, not -1 atm as you say [0].

[0] https://en.m.wikipedia.org/wiki/Vacuum#Measurement

⬐ URSpider94

Sure, but it's -1 atm gradient, which is how we commonly measure pressure. I.e., when you fill your car tires to 32 psi, that's 32 psi over atmospheric. When you're talking about pressure, it's the gradient that matters.

⬐ washadjeffmad

It was taught as psia, -g, & -d (or absolute, gauge, and differential) when I was in school, so I read your -1 atm as psid. Your car example would be 32 psi gauge assuming STP, or ~36 psi absolute.

The miners are below the surface and not at 1 atm, so any gauge measures taken there would be psid.

⬐ zkms

From the transcript, this is what the "5 atmospheres" thing meant -- it was about the pressure differential, and that whether the inside is at 1 atm or 0 atm isn't a big deal if the tunnel has to be underground (in which case it already has to stand quite a bit of outside pressure!):

> Exactly. And looking at tunneling technology, it turns out that in order to make a tunnel, you have to — In order to seal against the water table, you've got to typically design a tunnel wall to be good to about five or six atmospheres. So to go to vacuum is only one atmosphere, or near-vacuum. So actually, it sort of turns out that automatically, if you build a tunnel that is good enough to resist the water table, it is automatically capable of holding vacuum.

⬐ abalone

Thank you for clarifying. I misspoke earlier slightly, but the result is the same: when you build a tunnel strong enough to withstand the water table, it can hold a near vacuum. This helps transport and energy efficiency.. but also presents interesting issues for fire safety.

⬐ None

None

Elon maintains that with just cameras you can achieve super-human (10x) performance with self-driving cars. Fact I nabbed when he spoke about Boring at TED. https://www.ted.com/talks/elon_musk_the_future_we_re_buildin...

⬐ lloydde

Sure, but what is the reduction in accidents for just cameras vs all three? Super-human might only be reglatorial approved in the absence of technology being available with high safety.

⬐ agildehaus

I wouldn't ever bet against Musk, but Musk says a lot of things. Yes we should be able to make self driving cars with just cameras because, after all, we do it with just two eyes. But in what world is computer vision perfect? Tesla's Autopilot already has a fatality on its record because it couldn't differentiate the color of the sky with the color of a truck.

LIDAR solves a lot of problems.

⬐ drcross

>Tesla's Autopilot already has a fatality on its record because it couldn't differentiate the color of the sky with the color of a truck.

Do you think in your wildest dream that this exact scenario would ever be allowed to happen again by Tesla engineers? It may have crashed causing a fatality in this instance but it doesn't mean the prospect of camera only vision systems need to be written off. What people often overlook is that these systems get stronger with more and more edge cases and objectively the Tesla team are doing outstanding work with the sensor package that they have.

⬐ dllthomas

> But in what world is computer vision perfect?

I'd guess the same world where human vision is perfect...

⬐ virgilp

There's another aspect of it though. As soon as Waymo makes self-driving cars more popular, and there are less humans on the road - you just need inferior technology to be able to operate. Bootstrapping this is the hardest - after self-driving cars are a thing, normally we'll see accelerated move towards fully automated driving.

In the extreme case, if Waymo cars are 80%, Tesla cars can easily be 20% just based on the fact that Waymo cars will avoid most accidents made by Tesla software errors

⬐ galdosdi

You might be right, but I'll devil's advocate a bit, with a few possible predictions (none of which I believe in all THAT strongly, just speculation)

- If there's a period where new self driving cars are popular but aftermarket self driving kits for older cars are unaffordable or nonexistant, then during this period, the price of non-self-driving cars will plummet dramatically. Who knows, you might be able to buy a 2 year old camry for $1000. The market for these cars will exclusively be (other than oddballs/enthusiasts) people whose wages are so low that driving yourself in a non-self-driving car is still cheaper in an overall sense than using self-driving cabs, so that will cap the possible value of such cars at a pretty low point.

- As long as (possibly as a result of the above) the number of human driven cars on the road is at all significant (like, at least 1% or 0.1% or something), the benefits you mention will be hard to attain. In a given trip a car interacts with hundreds of other cars.

- You forgot about pedestrians and cyclists. In many places, there are more of those than there are cars, anyway, so even if all cars are self driving,

- There's going to be weather and construction and falling trees and runaway skateboards and runaway trucks on ice covered hills and wildfires and so on. This kind of software will operate at scale, so improbable events become important to handle.

- Uber is probably dumb enough to hope your argument is true and think they can get away with half-assed self-driving, which is why they're going to fail miserably and go bankrupt.

- There's going to be regulation. As a made up example, if the "Waymo patent tax" is "just" $5000 per car, it's going to be pretty hard for a car manufacturer to justify to regulators that they'll instead develop competing software that is somewhat inferior in safety, but "acceptable", for the sake of saving, say, half of that $5000 (it still costs the thrifty manufacturer money to develop or buy their lower quality software!). All it'll take is one fatality and people will be up in arms. If the "acceptable" technology was the first one out, it'd be different, but it'll be hard to convince anyone it's worth the risk when there's better tech already out there. BTW, Waymo management would obviously be clever about setting a good price point that captures lots of value but still is low enough to discourage this kind of competition.

- My vague software and driving intuitions tell me it'd actually be easier to design a more general system anyway (massive handwaving). It's actually HUMANS (specifically those accustomed to easy suburban western driving conditions) who have a harder time driving when they can't assume everyone will follow the rules, not computers, and I think that's why so many people mistakenly (IMHO) imagine it'd be so easy to make self driving cars if only everyone followed the damned rules. The simplified laws of physics applicable to driving are simpler than the DMV manual anyway.

From [0]:

> Musk is without doubt a dazzling salesman. Who better than a guardian of human welfare to sell you your new, self-driving Tesla? Andrew Ng—the chief scientist at Baidu, known as China’s Google—based in Sunnyvale, California, writes off Musk’s Manichaean throwdown as “marketing genius." “At the height of the recession, he persuaded the U.S. government to help him build an electric sports car,” Ng recalled, incredulous.

Based on the TED talk [1], Musk seems to be proposing a scarce luxury transport solution. The Boring Company is an answer to the question, "What inconvenience do Los Angeles millionaires suffer daily?" And for this he is being elevated to no less than savior of humanity?

[0] http://www.vanityfair.com/news/2017/03/elon-musk-billion-dol...

[1] https://www.ted.com/talks/elon_musk_the_future_we_re_buildin...

⬐ crush-n-spread

Forget about this argument. USA needs tech companies, and they need attractive ones that get investment. You think our tax money that builds roadways and schools materializes out of thin air? No, we need great, powerful, sexy, innovative companies. And guess what guy, those are expensive as all hell! But in the long run the tech gets cheaper and we all get it.

Let's take a trip back to the 1980s. You're looking at a desktop PC company like Apple, tearing them down how their first computer costs $5000 in today's dollars. Should anyone have paid you any mind at all? Of course not! Because of Apple pushing the tech so hard for so long, we now have another behemoth tech company on our country's soil. Have some perspective before you go cutting down tall poppies.

"Yeah, so this is using only cameras and GPS. So there's no LIDAR or radar being used here. This is just using passive optical, which is essentially what a person uses. The whole road system is meant to be navigated with passive optical, or cameras, and so once you solve cameras or vision, then autonomy is solved. If you don't solve vision, it's not solved. So that's why our focus is so heavily on having a vision neural net that's very effective for road conditions." - Elon Musk, latest TED (https://www.ted.com/talks/elon_musk_the_future_we_re_buildin...)

⬐ maxerickson

I'd love a really well implemented augmented reality windshield though, one that added sensor data to my field of view.

Imagine being able to see heat signatures at night, or through obstructions.

⬐ dwringer

The real question on everyone's mind is when he's gonna start talking about his 4D ones.

⬐ praving5

pretty soon :)

⬐ praving5

I don't know where he gets this "rich" ideas from? I mean either he is watching a lot of cartoon movies or reading lot of fiction.

⬐ RichardHeart

I love how often he must stare into space because he's actually thinking about the answers to complex questions, and not solely regurgitation of canned answers. The world needs more real thinkers being asked interesting questions in large public formats like this. Maybe some people can think deeply while locked in eye contact with someone else, but I think there's a overhead to the thoughtful gaze into another's eyes.

⬐ imartin2k

This guy is a phenomenon. Could become this century's most important person, if he can pull everything off that he plans to.

An interesting takeaway: He wants to dig tunnels because he doesn't believe in flying "cars" which would make feel people uncomfortable. Interesting antithesis to what his peers in the tech world are doing.

⬐ noir_lord

Flying cars beyond the cool factor make me very uncomfortable, aside from his joke about a hubcap becoming a guillotine, how do you solve the problem of maintenance? If you require a regime like light aircraft they'd be ferociously expensive, current land cars are already showing remote capabilities that are worrying, shutting down a car remotely is bad enough, shutting down a flying car is catastrophic and then who controls them, if they are autonomous they become programmable aerial weapons and if you human operated they become both weapons and machines for screwups (and drunk drivers).

I just don't see a future for them outside of billionaires play things.

⬐ njarboe

I heard my first drone the other day that I was not expecting. Just a loudish buzzing sound in the sky while walking in a park. Looked up and it was pretty high and not very large. I was hoping flying personal transport could solve some serious problems like traffic jams, but now I'm fearful. Like Elon mentioned, noise will be a problem (barring some new levitation system).

⬐ guftagu

Watching this interview made me feel like a useless bastard...

⬐ OrwellianChild

Interesting takeaways here around the logic behind the Boring Company that weren't clear from the spec video released earlier this week:

Cost savings breaks down by:

     1) Narrower tubes - less cross-sectional area means cheaper tunneling

     2) Continuous tunneling/reinforcing - no stops to build tunnel walls. (Seattle's Bertha did this.)

     3) More power = faster execution (not clear on details of this one)

Seems not to address different materials/geology in different depths. Also, no mention of regulatory burdens to approve digging underneath existing property rights. (two things I know nothing about)

⬐ zuron7

A lengthy discussion, but should answer some of the skepticism that HN put forth when the announcement was made.