Hacker News Reader

Comment by sschueller | Hacker News Reader

This one https://www.youtube.com/watch?v=6AV2NcyX7pk is insane. Elon's Boring Company is a joke in comparison.

They use a tunnel boring machine to bore a tunnel with a 45° slope.

They do go into the mechanics of how they make this insanely massive machine drive up a grade that steep, and how they ensure it doesn't slide backward.

I was glued to the screen more than with most movies.

If you like channels like Practical Engineering, you will enjoy this.

I watched this at 2x expecting a mildly interesting educational video. This was a mistake. Almost forgot to blink.

I was glued to the screen more than with most movies

i was going to say something similar. this was better than the action movie i saw earlier.

Agreed - most interesting YouTube video I’ve seen in many months.

Elon's company is targeting different things - cheap and fast, unlike the company in the video doing technically hard but probably not cheap.

Is The Boring Company actually targetting anything?

Or was it just a silly idea that Musk threw some money at and they've been struggling to justify their existence ever since?

Their website has:

>The mission: solve traffic, enable rapid point-to-point transportation and transform cities

via cheap tunnels. Maybe they won't get there but it's an interesting goal.

Move fast and break things runs into problems when “break things” translates to collapsed tunnels.

The boring company isn't move fast and break things. They literally just bought a drill and use it like any other construction company. There's no innovation at all.

They started that way, just buying a drill, but are now trying to innovate and build their own one. They are just starting using Prufrock-3. The Prufrocks are machines the Boring company have made themselves. They say "Prufrock's medium-term goal is to exceed 1/10 of human walking speed, which is 7 miles per day." which is way faster than anyone else. The Swiss machine in the video did 400m in 4 months. They are also experimenting with evacuated tunnels (https://youtu.be/nV07jqwCy0A?t=879) which may not go anywhere but is at least an attempt at innovation.

"7 miles per day" is all talk and no proof it can actually be done. The Swiss dig a lot of holes from easy to very complex terrain. If there was an easy way go that much faster they would.

There was a derailment of a cargo train the the Gotthard base tunnel last year [1]. It caused a huge amount of damage and one tunnel will be close until September this year. They need to replace 7km of track. If there way anyone to do this correctly faster they would as it is costing millions not having this route open. Concrete needs time to cure etc. some things you can't just make faster because someone said so.

[1] https://www.swissinfo.ch/eng/business/gotthard-base-tunnel-t...

It’s incredible how defensive people get around others’ ambitious goals.

Every Elon company sets super high goal targets. They are often completely unrealistic with current technology but it inspires certain types of people to innovate and it works quite well.

> The Swiss dig a lot of holes from easy to very complex terrain. If there was an easy way go that much faster they would.

Nobody said anything about “easy”. Also, The argument of “the established players don’t work on X so X isn’t possible” doesn’t work.

If the Swiss are happy with their industry the industry isn’t going to risk really capital intense experimentation to do better. See: Innovator’s dilemma

> ... medium-term goal is to exceed ...

According to the roadmap, another Elon company should have attended fully automated driving years ago.

    > The Swiss machine in the
    > video did 400m in 4 months.

No, a lot of the time the machine was stationary because the needed to manually reinforce what they were about to drill through, so the machine wouldn't get trapped and buried in gravel. Wouldn't Prufrock-3 be similarly slowed down!

In addition to that I assume that the 7 miles a day claim assumes 24/7 drilling, whereas I wouldn't be surprised if the swiss were doing 8/5 drilling.

Well the innovation appears to be, drill smaller holes faster, and put less traffic through them.

Brilliant!

Nobody said “break things”. Making one specific thing reliably, quickly and cheap is a completely different approach to the “move fast and break things” approach.

There's no collapsed tunnels. It's the same technology, just a smaller surface area / volume of displacement due to not needing exhaust mitigation.

Is fast and cheap always better though.

Fast. Fast always wins. Trying something, regardless of outcome teaches something about the world. The more trials, the more you roll the dice, the more you learn.

I really hate being wrong, but it is much better to be wrong a lot, and quickly understand why. The alternative is to try nothing. It’s kinda sad.

Fast wins for something like mobile apps but not infrastructure where safety matters and you can’t just shrug off liability. The Boring Company is a great example: fast to market themselves but almost all of their projects have fallen through.

Which nascent technologies have became common, dominant by innovators unilaterally adhering to the precautionary principle?

The cliché "Worse is better" describes this (turrible) phenomenon. (The flip side being "safety regulations are written in blood.")

I really wish this wasn't the world we live in. I want to live in a world with consequences (and justice). I've been railing against it my whole career. And judging from my meager savings, failing in my efforts.

Iteration always wins. There may be other constraining factors, but if you can violate those to iterate more (i.e. cheat), you’ll beat your competitors who honor those constraints.

This has nothing to do with the success or failure of any given company.

> but if you can violate those to iterate more (i.e. cheat), you’ll beat your competitors who honor those constraints.

Until people die. Then suddenly, you can't cheat any more. And everybody else has to live with your decisions.

https://en.wikipedia.org/wiki/Morton_Thiokol

I wouldn't call that winning.

You seem to have misunderstood what I’m saying. Iteration is how you learn. Learning through iteration, which is known under various names like Wright’s law or more generally experience curve effects [1], goes hand in hand with economies of scale in driving down costs.

Yes, iteration requires you to survive. Not sure how that’s relevant. Cutting corners also isn’t necessarily a bad thing - you’re focusing on the extreme example where people die. Cutting corners can also be a careful evaluation of what processes are and aren’t relevant to a given situation but that becomes trickier when it’s enshrined in law. Imagine if guidelines from the 80s about how to write software were enshrined in law.

Regulations more often than not do ignore the flip side in terms of the cost of compliance because it’s difficult to show the counter factual universe in which a regulation may save 10% more lives (or maybe even 0% more lives) but drove up costs by 100x.

[1] https://en.wikipedia.org/wiki/Experience_curve_effects

> Imagine if guidelines from the 80s about how to write software were enshrined in law.

If Edsger W. Dijkstra's wild ravings from the 80s were enshrined in law, perhaps we wouldn't have quite so many fatalities attributable to crap software. https://www.cs.utexas.edu/users/EWD/transcriptions/EWD10xx/E...

> A number of these phenomena have been bundled under the name "Software Engineering". As economics is known as "The Miserable Science", software engineering should be known as "The Doomed Discipline", doomed because it cannot even approach its goal since its goal is self-contradictory. Software engineering, of course, presents itself as another worthy cause, but that is eyewash: if you carefully read its literature and analyse what its devotees actually do, you will discover that software engineering has accepted as its charter "How to program if you cannot.".

It does depend on the domain though. Sometimes moving slowly and carefully can result in a faster successful outcome than throwing shit at a wall and seeing what sticks.

> The more trials, the more you roll the dice, the more you learn.

If you have infinite dice rolls this is obviously true. If every dice roll costs you something or indeed everything... heh. Maybe don't just roll it to see what happens?

Yes, sure fast wins when building a bridge. Or a tunnel. Which then collapses. Safe wins, this is not Facebook where people share some holiday pictures.

"but it is much better to be wrong a lot"

I disagree, "Oh the bridge collapesed, I was wrong! But this is much better than being right" - Nope.

If you're opposed to fast and cheap, I have great news for you about the cost and pace of construction in the US.

... which has its own reasons, namely regulations (you know, every time there was a new disaster they added another one) or maybe some corruption. You don't see bridges falling in US regularly killing tons of folks, do you.

I am not saying its ideal and there is no room for improvement, nothing in real world is, but please consider other, 'fast' scenarios for long term (100+ years) existence when not only many lives are at stake.

> You don't see bridges falling in US regularly killing tons of folks, do you.

Falling yes. Killing is hit-or miss. I would be surprised if the reasons bridges are failing due to lack of maintenance is entirely unrelated to the cost of building.

There are many ways things could be better, and many ways things could be worse.

Attacking something only because it's cheaper and faster seems silly.

Please share a single recent example from US, I am only aware of one relatively recent catastrophic case from northern Italy, where maintenance was subpar.

Cheaper reflects many things, overall quality testing, attention to detail, how much effort went into design etc. Its not a silver bullet but oh boy does it always show on result, without exception.

https://en.wikipedia.org/wiki/List_of_bridge_failures

https://en.wikipedia.org/wiki/Francis_Scott_Key_Bridge_colla... was not 2 months ago.

> Its not a silver bullet but oh boy does it always show on result, without exception.

Permit me to show you an exception!

This watch is USD59: https://www.target.com/p/men-39-s-timex-expedition-scout-wat...

This watch is GBP225: https://mrjoneswatches.com/products/a-perfectly-useless-even...

There are even more extreme examples, e.g. the watch from Richistan.

Depends on the application, but commodifying (easier instances of) things that previously required bespoke engineering is a legitimate pursuit.

Fascinating. Thanks for that link.

Reply on news.ycombinator.com