Perpetual Auction Currency

[oetyng]

@jlpell

I did my best to summarize it in one sentence. Give me a simple definition in a single sentence and I will add it to the post.

I’ll give it a go at first best time to ensure it’s succinct, I should be sleeping now. But the two systems (PAC and Joe Smith’s Free Market) are very different in fundamental ways.

[Bogard]

Regrettably, I’m too pressed for time to lay out the issues I see in your reasoning. So let’s just agree to disagree. Two small points though:

@jlpell

You are now appealing to the benefits of an algorithmic approach and contradicting yourself.

I’m on the record previously advocating for a potential hybrid approach. Should be in this thread if memory serves. On further thinking, I do think something like the PAC does provide guard rails, but it doesn’t mean that I’m against other types.

@jlpell

Apologies if reality is a cruel master.

You don’t have to apologize. I’m quite aware of market realities and very much enjoy unvarnished truth

[jlpell]

@Bogard

Regrettably, I’m too pressed for time to lay out the issues I see in your reasoning.

Please do when you get the chance.

[mav]

Seems worth putting this quote here since it’s relevant to the ongoing debate about markets vs algorithms, and seemingly not biased toward one way over the other (emphasis mine)…

@dirvine

…understand natural processes (nature does not do absolutes, loves error and works wonderfully as it evolves ). I want to try and inject much more of that and try to get away from the more typical computer programmer, “everything must be absolute” approach.

The problem with absolutes is it pollutes right to the inputs. so absolute systems only work with absolute inputs. A bit like adding generics in rust or templates in c++ those mechanisms can circle round to requiring absolute inputs. So rather than ordering nature they require nature to be ordered, therein lies the rub. (there are places for absolutes of course, but not all places all the time). You see AI now works much in this loose way, try / fail (repeat until success) approach and decentralised networks must pay more attention to that thinking (I don’t mean use AI for marketing, but think deeply why it works). Instead of always blockchains perhaps thinking about perceptrons and the like would expand the mind

So basically folk experienced in decentralised networks as well as being great programmers. I would also like to see some genetic [EDIT cheers @happybeing ] algorithm/neuroevolution (open ended) thinkers get into our teams.It is all related.

[mav]

Some more reading about loosely related topics for PAC. Mostly about prediction markets and oracles.

SchellingCoin - SchellingCoin: A Minimal-Trust Universal Data Feed | Ethereum Foundation Blog
“you have the incentive to vote the truth, because everyone else will vote the truth and you only get a reward of P if you agree with them” (from https://fc17.ifca.ai/wtsc/Vitalik%20Malta.pdf)
“The fundamental problem is this: cryptography alone has no way of finding out that much about the outside world. You can learn a bit about computational power through proof of work, and you can get some market data between one crypto-asset and another by having an on-chain market, but ultimately there is no term in mathematical algorithms for something like the temperature in Berlin.”

Hivemind (previously truthcoin) - http://bitcoinhivemind.com/

Witnet - https://witnet.io/

[JoeSmithJr]

@mav

“you have the incentive to vote the truth, because everyone else will vote the truth and you only get a reward of P if you agree with them”

It’s fundamentally flawed for more than just that one reason. Once a version of “truth” is out, the incentives are immediately biased—and it doesn’t depend on whether that “truth” is actually true. Even worse, motivated groups can push their version with such decisive initial force that, together with the opportunistic joiners who just want to make a buck, it would be impossible to shift to the actual truth.

[oetyng]

@JoeSmithJr

Once a version of “truth” is out, the incentives are immediately biased—and it doesn’t depend on whether that “truth” is actually true.

Of course it’s biased - that’s a feature, not a bug.

If the value of USD was x yesterday it is very likely to be ~x today as well. Any random number is not as likely.
Same applies for a repeated Keynesian Beauty Contest (which PAC and Shellingcoin are variations of), i.e. it is biased for a good reason.

@JoeSmithJr

motivated groups can push their version with such decisive initial force that,

This is a pointless argument.
“Motivated groups can push [attacks on the network] with such a decisive initial force that…” etc. etc…

I.e. there is nothing about this network that is 100% resistant, it’s only a matter of cost.
For every single thing, it’s about making the cost prohibitvely high. With a system such as this network, or PAC, it’s hard to prove both that it is and that it isn’t, which has been a long running theme on this forum.

Therefore, talking in terms of “it’s possible with enough force” is totally meaningless, because it will more or less always be true for practically everything we do in the network.