And if people leaving creates a deficit the reward automatically rises until enough people are tempted into joining or rejoining.
There will be a domestic market that might be more fickle (as you suggest), which leaves space for others to ‘arbitrage’ if the reward rises as fickle people leave but don’t return. The latter will though also learn that it is costly to move in and out, so they will think carefully before shutting down. Some will be fickle, some will be sticky, some won’t even care.
I suspect it is how the world works and the finance system attempts to mimic. If there is not enough grass then grazing animals move on or reduce population etc. In commodities exchanges farmers are (should be) told whether there is too much wheat and not enough barley for instance. That allows supply/demand balance. That balance is important.
Then if not enough folk like a product it dies as it should, we have seen many in the Internet so far that do that. So the supply/demand thing is inherent to an awful lot. I don’t think it in itself is sub-optimal.
I hear you, @neo. I held a similar view until very recently. Both strategies want to achieve the same outcome but are approaching it from different ends, with advantages and disadvantages to each side. Some factors to consider:
(Spirit of) centralization vs decentralization
Each individual vault being able to set its own price jives more with the ethos of decentralization versus the network holding all the power of price setting, which is a form of centralization.
Omnipotence
The network might have to deal with a lot of complexity to appropriately set the put price (e.g., oracle system to be aware of fiat price of the coin)
Latency
Even if the network could reasonably set the correct price to attract vault operators, there will be a lag time between when these operators become aware of the more attractive price and are able to put resources on the network. That’s assuming they haven’t already assigned those resources to other purposes as the current model might be assuming that serious vault operators will just wait idly by for prices to become attractive.
Uber example
Uber sets the price of car rides and changes them on a whim to attract more drivers. The end result is that Uber rides actually tend to be more expensive than normal cab rides. And during inclement weather, forget about it. Yet, they don’t necessarily bring out more drivers during these inclement weathers, potentially as a result of latency between the price going dramatically up, drivers noticing, wrapping up whatever they’re doing to get on the road. In inclement weather, I’ve often managed to find a normal cab before an Uber ride reached me.
Many of the drivers often complain about prices being too low. I suspect they would complain less about the network if they had the power to set their own price and were forced by the market to lower it.
Additionally, Uber has introduced various ride tiers over the years to account for the reality that different drivers offer different resources and would want a commensurate price for it (e.g., Black, X, XL, pool, etc.) Does the Safe network want to have the responsibility of eventually setting different pricing tiers?
It would be useful to find more examples today of networks setting (i.e., centralizing) the price of services for their decentralized operators (e.g., Uber) and studying what has worked well and what hasn’t.
AirBnB example
AirBnB is a network where individual providers set the price for the accommodation they provide. Therefore on that network, you can find palaces and hovels. You can find hosts who are super popular hosts (reasonable prices, great services) and hosts who host only once in a blue moon (high prices or terrible service). But everyone is represented. The network also outsources the complexity of having to know the economic realities of each locale and setting prices accordingly.
It would be useful to find more examples today of a central network letting the price of services be set by its decentralized operators (e.g., AirBnB) and studying what has worked well and what hasn’t.
Goodwill and psychology
Goodwill and psychology is something to keep in mind. With the elevator psychology, pressing that close button often does nothing but give people the impression that they can do something about their situation and that they aren’t just at the mercy of somebody else’s dictate. As mentioned, Uber drivers often resent Uber and bemoan how prices used to much better early on. A similar feeling might be at play with AirBnB hosts, but if it is, it must be much less pronounced because I haven’t heard of it with regards to pricing.
Regulators/failsafes
The network can always employ a failsafe to mitigate black swan events (e.g., social manipulation to collapse/spike the put price). In fact, the neighbor price proposal by @oetyng and @mav is a good step on that front.
Manipulation could happen here too. Could enough vault operators collude to quit the network and stay off thus forcing it to raise prices to an unreasonably high point? How does the network prevent that?
It is costly to leave and return. The rewards accrue to those who stay.
So vault operators don’t really have free choice to leave if prices are too low then. In effect, the network would be holding a metaphorical gun to their head (i.e., leaving and losing seniority) to accept the network’s price.
I think you’ve strayed from reason. There’s not a point there for me to respond to as far as I can tell.
To clarify, having a cost to joining the network acts as a break against various abuses. These include attacks such as, but not limited to the one you put forward.
I think it’s perfectly reasonable to point out that vault operators don’t have full free will to leave if prices are low because they have to be aware of other network constraints. But let me leave this discussion right there. I’m not sure what just happened.
There’s a false dichotomy here.
Neither the @JoeSmithJr bidding, nor PAC, is void of the supply / demand phenomenon. Not at all. Not in any way. Let’s just all be clear about that.
It’s only a nimbler, more direct, reflection of the dynamics within supply / demand, that these alternatives provide. Naturally many considerations on all parts and ends above and below that. But in essence, that is the case.
I don’t think @JoeSmithJr is unaware of this, but I think his words were interpreted as if this dichotomy existed.
One of the worst enemies of productive discussion is not knowing what we are talking about.
I’m pretty sure that what @JoeSmithJr means with
a scaling strategy summed up in “just leave if you don’t like the price” is terribly sub-optimal for adoption.
is that there are unevitably consequences and inefficiencies as a result of a slower, more inert and unresponsive supply / demand mechanics, which are:
terribly sub-optimal for adoption
…something that he has (as well as us, @mav and me, have) described concretely in this topic.
We have seen in most countries that the economy starts off that way then they get taken over by big companies after time passes. Either bought out or just pushed out by big companies.
Being able to set the price takes away from the network the ability to attract or dissuade people from supplying vaults.
In other words doing this will cause centralisation because greedy companies will be what they are.
Uber is a bad example as they are now changing for the bad. They started off without any insurances or protections for passenger or driver, and now they are having to apply the proper safety concerns and uber are controlling them too.
The low prices are often at the expense of others
Now the other side of your examples is comparing SAFE to the money grabbing companies out there. You have tried to divide it into farmers. But it does not apply since the network NEEDS to be able to attract or dissuade farmers for the health of the network.
Safe is more like an organism rather than independent farmers doing a task then stopping. EG Uber, AirBnB and the others.
Uber and AirBnB in 10 years though will be charging higher and higher prices once taxis and many smaller holiday places are out of business. History is a good teacher.
In my mind supply/demand will in most of the time work well and reflect a fair equilibrium between supply/demand.
External shocks like sudden increase/decrease in aggregated supply/demand due to factors, for example sudden popularity increase through media exposure or negative popularity by political policies. Sudden shocks would impact price positive/negative beyond fair value but only for a short limited time until supply and demand adjust and finds a fair value.
It might not be an sub optimal solution but it is a solution that works quite well for many goods and services all over the world with the positive feature of not needing any or limited intervention, algorithms or others, to work.
I will argue that it works for blockchain mining and that it would probably also work for the SAFE-network in a similar way.
What if many choose to host vaults just so they can surf the SAFE-network for free, they might not care much of making money?
This don’t seem like a fair comparison to me, Uber and AirBnb are dominant players in a market with high entry barriers for ordinary people to compete with them, compared with the entry barriers of the SAFE-network which will be relatively very low.
I agree. And there must be a risk reward built into the network for those farmers who believe in the network and will keep the their nodes running no matter what the price bidding might be.
Whatever the design decision we must focus on improving world economics. Right now things are way out of balance.
An example of the fix:
Having money from economic growth flow to poor people rather than the rich feeds into a lift in the rate of economic growth and lower unemployment. Conversely, as income inequality increases, the potential for economic growth is constrained. If 1bln is giving out. Those with lots of money just absorb the profit, whereas the poor people spend it back into the economy, creating growth and jobs. With the latter, the rich get richer through true economic growth as poor people buy more of their products. A balance is struck!
What this means for the safenetwork: poor people equal the majority of farmers, not just the rich people, the potential colluders. A balance, weighted toward the majority, makes a better network.
So to design a reward algorithm to push behaviour away from wealth inequity makes sense for everyone.
But Uber and AirBnB also need to attract drivers and hosts respectively for the health of their network. If they don’t they will die as they don’t own a single car/lodging.
In any event, those examples were meant as thought starters to find and learn (economically) from other networks where independent agents provide a service on a network. There are several others I can think of right now, but would rather mention them when I have more time to clearly write out why and how they are an apt comparison.
By the way, a potential attack vector to manipulate the service price in safecoin set by the network would be to manipulate the fiat value of safecoin on the markets, up or down depending on the attacker’s goal. I realize that price set by vault operators would also be vulnerable to such an attack via fiat markets, but would it be less so or more so?
The process is completely voluntary. No one here is an anti-freemarket or centralization advocate. The airbnb and uber examples were good analogies, but you turned them into quasi-strawmen with descriptions that had a lot of subjective bias.
There are three ways that a transaction between the network and an operator can proceed.
Network sets an buy offer price for resources based on all section information it is aware of with regard to network status/operation and the vault operator can accept it or not participate in the network. (aka Algorithmic Price Optimization or more affectionately the Take It or Leave It Method).
Vault operator sets a sell offer price based on its local needs or “feelings” and the network is forced to buy resources at that price if enough operators send the price in a certain direction. (aka Perpetual Auction Currency or JoeSmith’s Free Market)
Network sets it’s buy offer and vault operators set their sell offer for resources. When the rates match then a vault stores a new PUT or returns a GET, if the rates don’t match then stalemate and nothing occurs. (aka The Unregulated Exchange Method)
Of the three options, it is self-evident that case 1 is superior for smooth network operation.
This is not necessarily true. I don’t see slowness or lack of “agility” as you describe it. From a mechanical/systems engineering perspective these are features, not bugs. Damping and inertia are not always bad, since they offer smooth and stable price dynamics. Try riding your car down a bumpy road with no shock absorbers. As long as the price update (sampling) frequency is above the Nyquist limit, vault population dynamics can be resolved and prices can adjust at network time scale.
I don’t think I accused those advocating for network control of service pricing of being anti-freemarket in the economic sense. That should be obvious especially since I was such an advocate myself until recently. I’m not doing any of this to get into arguments. I’m only trying to contribute, dispassionately and when I can, to a network that I think could help secure certain data types.
Additionally, do you mind being more specific as to how the uber/airbnb analogies were straw men? I realize that I might not have explained well how they are analogous given some of the response. The insight I shared with regard to those companies were certainly anecdotal, but I’m not sure it was subjective.
I’m not sure it is self-evident. The only way to find out which would be best would be through testing–and looking at other networks with decentralized participants might help get there earlier.
If you read closely, you see that I don’t paint a black and white picture.
While the statement “not necessarily true” is true, it is quite unlikely that there would be no negative consequences or inefficiencies of the inertia. The question has always been about identifying the effects of those inefficiencies, and wether they in reality accomplish the task better than some other way.
I mean, arguing that inertia isn’t bad is like saying friction isn’t bad (i.e. a feature), and I’m certainly not saying it is bad (I.e. a bug). It’s not the question at hand, it’s not really a meaningful question at all (life would be quite hard without friction). It’s all about how much friction, where, how, when, why.
And then the perhaps way too subtle:
This means to say, that it is not a simple good/bad question, which has been shown in the very long texts, discussions analysis written on the subject in this topic only, but others as well.
The actual subject here has been hashed a couple of times before, so my post was actually not intended to do that again.
Sorry, but this is not Perpetual Auction Currency.
You could say it’s close enough, but the devil is in the details.
Edit:
I just had to address this as well
I think it is pretty clear that nothing in this regard is self evident (which way to organise the SAFENetwork economy).
Here are your straw men:
Relative to all the other complexity involved with the network operation this is trivial.
Pro vault operators will be intimately away of the price in real time. Casual operators won’t care what the price is. Vault diversity and population statistics is a health metric up for consideration by the network, but uncontrollable by independent operators. If resources are dwindling and the network is becoming unhealthy, raise offer prices. If resources are too plentiful, lower offer prices. How much simpler do you want it?
Uber Example:
Yes. This allows them to grow and compete with existing alternatives. SAFE will need to do the same.
Not applicable in the early stage due to network trust fund. Not applicable long term due to network adoption and algorithmic cost minimization.
Not applicable. Everyone likes to complain about everything. If they quit and find a different job, then you actually have some proof that the prices really were too low. The clients don’t complain when the prices are low.
Not applicable. A chunk is a chunk. The fastest to serve a chunk wins, regardless of what tech these use to accomplish the task. Longer term in the future, tiered pricing might be useful for for archive and computation nodes.
AirBNB Example:
It’s the SAFE Network, not the SRFE Network. Secure access implicitly requires cost minimization. It is not possible to know the true minimum cost for resources unless some operators leave the network. That’s bare metal truth.
The network doesn’t need to know or care about the economic realities of each locale when using an algorithmic approach or any other approach. The outsourced complexity exists in the external Fiat:SAFE exchange market, and is not necessary for the internal SAFE:Resource market.
While this does have some interesting applications for a pricing model, SAFE’s dictate is a mathematical and reality based one. Apologies if reality is a cruel master. SAFE is basically telling the operator that if they don’t accept the current offer, then the current outlook based on network metrics is that the network is heading for failure. Rinse and repeat at next price update. It’s a form of no nonsense bargaining.
You are now appealing to the benefits of an algorithmic approach and contradicting yourself.
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.