Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

[Link] The Copenhagen Letter

0 chaosmage 18 September 2017 06:45PM
Comment author: TheAncientGeek 13 September 2017 12:12:26PM 0 points [-]

No, we're in a world where tourists generally don't mind going slowly and enjoying the view. These things would be pretty big on the outside, at least RV size, but they wouldn't be RVs. They wouldn't usually have kitchens and their showers would have to be way nicer than typical RV showers.

And they could relocate overnight. That raises the possibility of self-driving sleeper cars for business travellers who need to be somewhere by morning.

Comment author: chaosmage 13 September 2017 03:41:58PM 0 points [-]

Yes. I wonder how hard it'll be to sleep in the things. I find sleeper trains generally a bad place to sleep, but that's mostly because of the other passengers.

Comment author: Lumifer 12 September 2017 04:21:02PM *  1 point [-]

Sorry, I'll bail out of the detailed debate, primarily because it's all handwaving and there are very few falsifiable assertions in there. You say that it's going to get much much cheaper, I say that it won't -- and there is no way for us to resolve this disagreement. As an aside, several statements of fact that you make here are wrong (no, you don't need a crew of at least five people to charter a yacht; yes, RV rentals are much more expensive than hotel rooms -- have you rented an RV? I have).

For what it's worth, my original opinion remains intact.

Comment author: chaosmage 12 September 2017 06:50:28PM 0 points [-]

I should be disappointed, but disappointment requires surprise.

Comment author: Rossin 08 September 2017 06:27:46PM 2 points [-]

I'm a little worried if it came across that way because that is not at all what I am trying to argue. The example was intended to show that if one sees something in the world that they think is bad (people dying in Africa of preventable disease) and yet they end up doing nothing by convincing themselves not to care, the mental process going on in their head is likely not very different from the one that occurred in people living under Nazi rule, who themselves felt uncomfortable about their Jewish neighbors being rounded up by the Nazis but did nothing. I am not comparing people who aren't donating to charity with the actual Nazis, and I'm sorry if it seemed that way.

Comment author: chaosmage 12 September 2017 01:12:00PM 1 point [-]

Don't worry, you didn't actually come across that way, Lumifer is just being a jerk again. You're fairly new here, so you don't yet know Lumifer prefers that kind of comment. Sorry about him, and about LW not having a mute button.

Comment author: John_Maxwell_IV 11 September 2017 10:40:40PM *  4 points [-]

From my point of view, most of the "content" is in comments, not in posts, and being able to participate in the give-and-take makes the conversation even more worthwhile.

If content is content is content, why are your nasty comments typically about top-level contributions rather than comments? However little signal there may have been in chaosmage's post, your reply contained even less. It was at DH3 at best, plausibly DH0.

If content is content is content, and most of the content is in the comments, shouldn't we also call everyone in this thread who responded to chaosmage's post in earnest stupid?

When you assess your own contributions, do you think they are stupid? Do you think they are good enough to be top-level posts? If not, why exactly should should we hold your comments to a lower standard than we hold top-level posts to? Especially if the discussion is where the value is supposed to be found.

Frankly, I find your contributions consistently uninteresting and banal. Just in this comment, you managed to

  • Misread my comment. "If we take it more reasonably as a trade-off" - that's what I did. I discussed the both the costs and benefits of harsh replies (cf "The cost of an occasional bad post is not very high: you read it until you realize it is bad and then you move on.")

  • Strawman my comment. "I am also not interested in providing incentives for any content. There are places on the 'net with LOTS AND LOTS of content -- Facebook, Reddit, Buzzfeed, etc. etc." That's not what I advocated. I advocated taking in to account the counterfactual impact of nasty comments. Which is a point you still haven't responded to. Maybe it would be sensible to make comments harsher if we start to have the "too much content" problem. But that's not the current problem.

  • Introduce a non sequiter: "As I said several times I dislike the high-priesthood view of science and I dislike the high-priesthood view of forums as well. I am not particularly interested in having a few high-status people bestow their wisdom upon me in exchange for adoration." How is this relevant to our discussion? The issue here is you being nasty without adding any signal. Despite their username, chaosmage is no "high priest".

  • Respond at DH3 again: "First, that's false." Why is it false? As usual, you express disagreement without adding anything to the discussion.

Consistent with my position, I might tolerate you if your only issue was low-signal contributions. I might tolerate your constant need to disagree with everything if you weren't a jerk about it: "Hey chaosmage, I'm skeptical that the future is this easy to predict." I might even tolerate your toxic behavior if you were actually providing contributions of value. But your combination of low-signal contributions and toxic behavior is too much. You seem to think that flatly contradicting everything in sight makes you some kind of bold maverick. I see you as more of a poster child for this essay.

Show me the data.

During the period where LW declined, you have consistently been at or near the top of "TOP CONTRIBUTORS, 30 DAYS" on the sidebar. You spend more time participating than anyone else, so you set the tone for the forum. This was during a period where almost everyone agrees that things moved in the wrong direction.

At this point, due to all the time you spend here, Less Wrong has undergone evaporative cooling. Many of the users who remain are those who have an affinity for your brand of uninspired, disagreeable obstinacy. I know of at least 4 other people who agree with me that you, specifically, have been toxic for Less Wrong's culture. Some of these people I have a lot of respect for, in the sense that I learn things from talking to them. (I never learn things from talking to you.) These people spend a lot less time contributing to LW as a result.

What if you are the harbringer of Eternal September? Have you ever considered that?

I probably won't respond further in this thread, because I've found that arguing with you results in an exponentially growing tree of tiresome objections. You never seem to change your mind on anything and you seem to disagree just for the sake of disagreement. I don't ever get the sense that you have an underlying model of the world that informs your comments. It seems like you are more about disagreeing with everything in sight. And you seem to think that because you are disagreeing with people, it's OK for your comments to be held to a lower standard.

I will say that I really, really wish you would find somewhere else on the internet to hang out.

Comment author: chaosmage 12 September 2017 01:06:38PM 0 points [-]

I completely agree with everything you said here.

Comment author: Lumifer 11 September 2017 04:13:33PM *  5 points [-]

The two root problems in your post are that you treat self-driving cars as cost-free instant teleportation devices and that you don't understand which costs drive the particular forms that businesses take.

Diversification of vehicle types

Somewhat, but much less than you expect because contemporary car design is driven by law-mandated safety requirements. The same requirements put a floor on the cars' weight.

There is also the fact that a general-purpose car will spend less time sitting in a parking lot doing nothing while waiting for someone to require it. Renting specialized equipment is expensive partially because of this -- there is a lot of idle time.

Services at home

Nope. It's not the case that the doctor doesn't come to your house because she can't afford a driver. The doctor doesn't come to your house first because her time is more valuable than yours and second because it's hard (=expensive) to bring along all the nurses and assistants and the medical equipment that she has around her office. And, by the way, the doctor doesn't fill out the insurance paperwork -- she has a much cheaper assistant who does.

Of course you can get a doctor (and a hairdresser, and a tatoo artist, etc.) to come to your house, even without self-driving cars. It's just going to be very very expensive. I don't expect this to change.

Rent anything

The cost of a driver is a minor component of the cost of renting large, expensive, luxury things. Taking it out will not make them suddenly affordable. And, by the way, who will unload, set up, dismantle and load back into the self-driving truck all these jacuzzis and huge sound systems?

Also, about the "stuff that previously only millionaires or billionaires would afford" that your median-income person would be able to rent if only you take the truck driver out of the equation -- literally nothing comes to my mind.

Self-driving hotel rooms

They are called caravans or camper vans or RVs. They exist. Have you tried renting them? They are quite expensive to rent, much more so than hotel rooms.

Rise of alcoholism and drug abuse

" for a large number of people, driving is their only reason not to drink or do drugs" -- that's, um, wrong. I have no idea how you came up with such nonsense.

Autonomous boats and yachts

There is no such thing as a "sailing license" (in most countries that I know of) and renting sizeable boats is quite expensive. It will not become less expensive if the boat has an autopilot. Recreational boating doesn't want fully autonomous boats, anyway, and commercial shipping already uses autopilots and still finds out that it needs people to run ships.

Mobile storage

I think this already has been mentioned -- storing things in trucks is much, MUCH more expensive than storing them in warehouses. Also, just-in-time (JIT) manufacturing was all the rage a couple of decades ago. The enthusiasm has cooled down considerably since then, mostly because people have found out that JIT production is not robust to disruptions and does not degrade gracefully if something goes wrong. Hint: something always goes wrong.

Renting a self-driving truck is not going to be cheaper than renting a regular truck (the rental company does not supply a driver in any case). Go ahead, look up how much renting trucks costs and then see if you want to store stuff in them.

Comment author: chaosmage 12 September 2017 12:49:09PM *  0 points [-]

contemporary car design is driven by law-mandated safety requirements

I don't know about your country but in mine (Germany) the car industry has so much influence they basically write their own laws. (That's how we got those safety requirements: They're defense against cheaper cars from abroad.) If their business model stops being focused on general use cars, the laws will change very quickly.

a general-purpose car will spend less time sitting in a parking lot doing nothing while waiting for someone to require it

Sure! Not a problem if its TCO is pretty low. Most of those more specialized vehicles will be cheaper than general purpose cars, because they have sharply reduced capabilities. A great number of them will be pretty small, just enough to carry a single person or (in an even smaller, windowless car) a piece of cargo. Others will have to require paying a premium, but they'll mostly be doing things general purpose cars cannot do, like transport a horse.

Side note: Some of those specialized cars will also be sold, not rented. I imagine rich parents gifting their seven year olds their own car. And as soon as someone makes what is basically just a bed on wheels, a small minority of people will live in those things.

The doctor doesn't come to your house first because her time is more valuable than yours

That's a good point. So maybe it starts with hairdressers.

Of course offering services at home will be most attractive to those who are new to their field and haven't sunk costs into an office.

and second because it's hard (=expensive) to bring along all the nurses and assistants and the medical equipment that she has around her office.

Most doctors need very little equipment most of the time. Some types of doctors (psychiatrists, dermatologists) need very little equipment period.

The cost of a driver is a minor component of the cost of renting large, expensive, luxury things. Taking it out will not make them suddenly affordable.

The main cost is insurance and autonomous vehicles means that one drops hard. The lack of a driver mostly means you can rent things out in a very large operating radius.

And, by the way, who will unload, set up, dismantle and load back into the self-driving truck all these jacuzzis and huge sound systems?

Most of the things never leave the "truck". The vehicle is built around them, on a standardized flat chassis. Some of them will have staff, sure. For example, you'd have a bartender if you were renting out a highly specialized mobile bar that has casks of twenty different excellent whiskeys and might be popular with bachelor parties.

Also, about the "stuff that previously only millionaires or billionaires would afford" that your median-income person would be able to rent if only you take the truck driver out of the equation -- literally nothing comes to my mind.

Alright. I'll leave it at that.

They are called caravans or camper vans or RVs. They exist. Have you tried renting them? They are quite expensive to rent, much more so than hotel rooms.

You're simply wrong about rental RVs: their prices are not much more expensive than hotel rooms anymore. (They do remain more expensive than small rental cars.) Check for yourself at places like http://www.apollorv.com/ . They'll get cheaper by going electric (like all cars will) due to less moving parts and less repairs. They'll get cheaper again by going autonomous (like all cars will) due to less mass for the driver cab and less accidents. So even if it was just self-driving RVs, they'd be an opportunity to disrupt stationary hotels.

But RVs are lower class than most hotel rooms, they're cramped, they're optimized for carrying lots of supplies and they have kitchens. A lot of people wouldn't want to use an RV even if it was half the price of a typical hotel room. If you have a self-driving RV and you want to really tear into the market share of stationary hotels, you throw out the kitchen and most of the cupboards and put in the best bed that you can make fit and a great entertainment system.

" for a large number of people, driving is their only reason not to drink or do drugs" -- that's, um, wrong. I have no idea how you came up with such nonsense.

Couple of years in psychiatric research.

no such thing as a "sailing license" (in most countries that I know of)

So you don't know a lot of European countries? I don't know a lot of non-European ones, so you may be right this isn't a factor there.

renting sizeable boats is quite expensive

Exactly. And why? Because the risk of accidents, and the insurance to cover that, is a larger cost factor than with cars. And the main advantage of vehicle autonomy is the sharply reduced number of accidents.

Recreational boating doesn't want fully autonomous boats, anyway

Maybe many don't. But recreational divers and anglers and people who just need to get across the water will be happy with it.

commercial shipping already uses autopilots and still finds out that it needs people to run ships

Yes but it needs way fewer people. If you charter a yacht now, you have a crew of at least 5 people. With an autonomous yacht, you can go down to one crewmember who mostly cleans the place, and maybe a cook.

Renting a self-driving truck is not going to be cheaper than renting a regular truck

Your assertion is ludicrous. Yes it will be cheaper, and a lot. The self-driving truck doesn't need to carry all the mass that the driver needs, including fragile points of failure such as windows, it doesn't have mandatory stop times, it gets into way fewer accidents, it basically cannot be stolen. If you don't think a self-driving truck company can undercut a traditional trucking company, I hope you don't run a trucking company.

Comment author: fortyeridania 11 September 2017 01:40:37AM 3 points [-]

This was also explored by Benedict Evans in this blog post and this EconTalk interview, mentioned in the most recent feed thread.

Comment author: chaosmage 11 September 2017 09:44:09AM 2 points [-]

Wow, this is amazing! Thank you!

He talks about various general effects rather than specific business opportunities, so the overlap is very small, but his vision and mine seem entirely compatible.

Comment author: Jiro 07 September 2017 09:29:20PM 2 points [-]

The reason it's too expensive to just leave them in the truck all the way from the factory to the buyer is that if you leave them in the truck, you are denied the use of the truck for other deliveries until you have a buyer. Having the truck be self-driving doesn't help, unless driving costs money but trucks are free.

Or, more broadly, using a truck for something else other than transportation, such as storage, a doctor's office, etc. means that you're replacing a specialized tool with one that tries to do two things (driving and storage, driving and doctoring, etc.) It's particularly obvious for storage, since storage uses up a lot of space and using trucks for storage is inefficient just because of the cost of all those trucks, but even for other uses, a tool which does two things is never going to be as good as using a specialized tool for each one.

(The post tries to suggest that the vehicle can be specialized, but a vehicle specialized to also be an X is never going to be as specialized as having an actual X.)

Comment author: chaosmage 11 September 2017 09:18:47AM 0 points [-]

What I failed to say is that I expect the time between production of a good and arrival at the customer to continue to shorten. I think additive manufacturing makes just in time production economical for a growing segment of goods.

Obviously the production and delivery specifics of different goods are very different. Wheat will continue to be warehoused long after jewellery and clothing have completely moved to just in time production and direct delivery. I misstated my position by not mentioning these important differences.

I say I think this transition will be slower than the others because I imagine it starts with goods that are precious, fragile or spoil quickly, i.e. the goods where you most want to minimize loading and unloading.

I don't think trucks will be free, but I do think they'll be much cheaper per ton per mile than they are now. Electric means the motor gets rid of most moving parts and becomes much more reliable. Autonomy means the rest of the vehicle gets rid of all the things that only a driver needs, especially the operator's cab and mandatory stop times. A warehouse will still be cheaper in many cases (especially in rural areas), but the fraction of things that it is easier to just leave in the car can only go up.

Comment author: Yosarian2 08 September 2017 08:30:29PM 2 points [-]

Sure, that's very possible. Just because it didn't work last time doesn't mean it can't work now with better technology.

I think anyone who goes into it now, though, had better have a really detailed explanation for why consumer interest was so low last time, despite all the attention and publicity the "sharing economy" got in the popular press, and a plan to quickly get a significant customer base this time around. Something like this can't work economically without scale, and I'm just not sure if the consumer interest exists.

Comment author: chaosmage 11 September 2017 08:54:17AM 2 points [-]

You make excellent points. I hadn't even heard of SnapGoods, NeighborGoods etc.

I'm imagining it not as a peer to peer service, but more along the lines of a car rental company that owns a fleet of things it rents out.

I think you're right about the need to build a significant customer base rather quickly. My guess is that it might be feasible to first offer big expensive things that people don't usually own already, like a fancy jacuzzi, a top end VR rig, a complete "wedding size" soundsystem and a bouncy castle. And once you're known for those, work your way down into more normal consumer goods, guided by the requests of your first customers.

Comment author: John_Maxwell_IV 10 September 2017 07:49:54AM *  7 points [-]

Note that neither Lumifer, nor Dagon, nor Brillyant have ever made a top-level submission of original content to Less Wrong. It's easy to be a critic.

Since Lumifer, Dagon, and Brillyant seem to want a site that never has anything new on it, may I suggest example.com? It hardly ever changes.

...what did people say they'd need to rejoin [Less Wrong]?

Feel free to read these yourselves (they're not long), but I'll go ahead and summarize: It's all about the content. Content, content, content. No amount of usability improvements, A/B testing or clever trickery will let you get around content. People are overwhelmingly clear about this; they need a reason to come to the site and right now they don't feel like they have one. That means priority number one for somebody trying to revitalize LessWrong is how you deal with this.

Source. Less Wrongers overwhelmingly want there to be more posts.

The problem with comments like Lumifer's is not that they are incorrect. It's that they create a bad incentive structure for content creators. Anyone who posts to LW is doing free labor in an attempt to improve the accuracy of the community's beliefs. I believe that lukeprog, Eliezer, and Yvain have all complained that writing LW posts is not very rewarding. If there's some probability that the Lumifers are the world are going to call your post "stupid" without offering any specific feedback, that makes the job even more thankless. And no, this is not necessarily something a person can predict in advance: a previous post chaosmage made got voted to +55, and the ideas in it were being used by a friend of mine years after it was made.

The cost of an occasional bad post is not very high: you read it until you realize it is bad and then you move on. But the cost of nasty comments like Lumifer's can be quite high. Most online communities suck, and nasty comments are a big part of the reason why. If I was selling a product you could spray on an online community to prevent anything from growing there, I would name it Lumifer.

Comment author: chaosmage 11 September 2017 08:35:12AM 2 points [-]

Well said.

To be fair, I didn't expect this to be voted very high. Not in these times. I was basically writing this down to have a public record of the talk I gave, and to look back at this years later and see how I did predicting things.

I do see your larger point though. I was clearly motivated to produce this by the large and friendly crowd at the European Less Wrong Community Weekends, where most people go by Crocker's Rule and give way better feedback than Lumifer is able to. The little extra effort of writing down would have been worth it even if all the comments would have been Lumifer level quality. If LessWrong was the only place I put these ideas, the current environment here would indeed not feel very much worth the work of spelling them out in the first place.

View more: Next