Hacker News Reader

Comment by wpietri | Hacker News Reader

And I'm just going to self-promote a bit and say that the bet was registered via a project I worked on, the Long Now Foundation's project Long Bets:

http://longbets.org/

We've been going since 2002: https://www.wired.com/2002/05/longbets/

We are happy to host bets of long-term significance, and the minimum bet is only $200/side. I am glad to personally help shepherd people who are serious about bets to make sure you get through the process.

I would especially like to see HNers making bets about the things we argue a lot about. Bitcoin! VR! Uber! If you're tired of people posting waffle on some topic where you have a strong opinion, then challenge them to put money down.

View on HN

Long Bets is a good site to make a public statement but PredictionBook[1] or Metaculus[2] may be better to track predictions about things we argue about. Both sites calculate your Brier score, which is how you can know how well calibrated you are.[3]

Since prediction tracking and aggregation sites use reputation instead of money, they scale to a higher number of predictions. As proof, there have been only 747 predictions and bets on Long Bets, while gwern alone has made[4] 2275 on PredictionBook. Furthermore, the Metaculus community prediction provides valuable forecast information,[5] much like prediction markets, while predictions on Long Bets don't count as Bayesian evidence of their own.

[1]: https://predictionbook.com/

[2]: https://www.metaculus.com/

[3]: https://en.wikipedia.org/wiki/Brier_score

[4]: https://www.gwern.net/Prediction-markets

[5]: https://www.metaculus.com/questions/track-record/

Sure. These things have different purposes.

The Long Now's mission is to foster long-term thinking broadly. We don't care as much how well-calibrated a particular individual is. The point of Long Bets is to generate discussion, argumentation, and consideration. That's part of why we require each bettor to make a case for their side: we're less interested in who was right or wrong than how and why someone was right or wrong.

I hope that people also use those sites. But I would love to generate some clear, dramatic bets about controversial topics here. The Buffett bet is a great example. People every day are putting money into one investment or another and have been for a long time, and there's no lack of data on results. But because of this bet a great number of people have thought about these specifics, as well as the many interesting related issues. I'd love to see that around more topics.

On https://longbets.org/register/, it says that "Long Bets registration is temporarily disabled. Please check back soon."

Why is registration temporarily disabled? Can I get an email notification when it's back up?

Good question! Somebody took over running the technical side of that some years back, so I haven't followed it closely. I'll ask.

But if anybody has a bet they'd like to get registered, just email me and/or them.

What email should they use?

From the website linked in his profile:

Contacting me

The best way to contact me is via email; try my first name at this domain.

You can also write to longbets@longnow.org - but we will have the registration open again shortly (within weeks).

Dan @ Long Now

If longbets.org was really serious about long-term bets, it wouldn't be restricting years to only 5 digits. Haven't they heard of the Y100K problem?

If they had only used 4 digits, then in eight thousand years or so, longbets.org would have felt really stupid about having to hire a historical programming re-enactor to show us how our ancestors used to adjust the width of a database column.

But luckily they used 5 digits. 5 digits should be enough for anyone.

The 5-digit display year is definitely in keeping with our mission, which is to get people to think longer term. But under the hood, I represented all dates as 64-bit longs of ms since epoch. So I believe the farthest date the current code can represent is 17 Aug 292,278,994 AD. By that point the earth's day will be an hour longer [1], so whatever sucker has inherited this project will have bigger problems than just the size of the field.

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

What is the Y100K problem. More importantly, who would they pay out at that point when both parties are dead

I'm pretty sure that's a joking reference to the Y2K problem: https://en.wikipedia.org/wiki/Year_2000_problem

Payout are to the bettor's selected charity. Or, if that's gone, the closest one we can find.

What if you want to bet they find a cure for death?

We would totally accept that bet. The closest bet we have right now is bet 11: “At least one human alive in the year 2000 will still be alive in 2150.”

http://longbets.org/11/

The tricky part is coming up with a clear agreement about what "cure" and "death" mean here. But there's a lot of good discussion around e.g., the extropians, so I think a solid bet could be created.

Does "Brain in a Jar" count? How about a Ghost in the Shell? More likely this is a bet on if Cyrogenics finally works out.

You mean for that bet? Good question, and not one I'm likely to have to answer. But from the arguments, the spirit seems to be about traditional meat-brain-in-meat-body life.

Or do you mean for your own? This is between you and the person you bet with. If you have a firm opinion, definitely find somebody on the opposite side and see if you can put together a bet.

this utterly rules. like, i want to fly to your current location to give you a fist bump. in my utopian dream world, every political pundit has to put their thoughts on here.

Next you're in SF, definitely look me up. A couple years back the Long Now opened a cafe and bar, so we even have a good place to meet for the fist bump:

https://theinterval.org/

Even if they get funded by special interest groups to spew the group's narrative, at least this will be a quantization of how correct or incorrect the pundits are.

this would separate the wheat from the chaff with regards to special interest groups.

Crazy question: how can/does longbets interact with people's estates? Meaning -- what happens when one of the bettors is dead when the bet is settled?

No, it's a great question.

We take the money up front and deposit it in a long-term investment account. Each bettor picks a charity, and the initial bet plus half the gains go to the winner's selected charity. (The charity thing is to get around US gambling laws.) If if that charity doesn't exist anymore (which is plausible given our longest current bet settles in 2150) then we pick a charity that's close in spirit.

> We take the money up front and deposit it in a long-term investment account

A minimal-fee S&P 500 index fund, I hope. I think Buffett just proved that anything else would be fiscally irresponsible.

That's a very cool project! Congratulations on keeping it going for so long!

Cool. Put your money where your mouth is.

I'm not sure why this is getting downvoted; that's exactly the point. We get people to stake money and reputation, which makes them very careful about making clear, specific claims. There is a whole lot of mouth around the future, but very few testable predictions. We could have tried creating an apparatus for rigorous predictions, but everybody understands a bet.

Reply on news.ycombinator.com