Some practical thoughts regarding blockchain based science and engineering funding proposals.

I review one.

(My opinion as first shared on Discord.)

 

— 〈  1  〉—

Scandinavianlife - Today at 6:15 PM
``What do you think about it @tibra ? https://www.matryx.ai/''

Interesting. There're actually some good ideas in Matryx:

(1) The problem they propose to solve. It's important.

Most reward in research is winner takes all. Isomorphic to contest reward. Or to a bounty.

(1a) Contest based reward is specifically designed to get more production than rewarded. That's why for example, when Nature and some other groups did a bounties for problem solutions platform it failed.

The rewards offered were typically far too low for genuinely hard problems: create a solder for A to B in electronics without using lead. 15.000$ reward. (3.000.000.000$ worth if the problem, a long unsolved one in fact, was actually solved and the solver marketed it.) And the mode reader of, say, Nature, is perfectly aware that 15.000$, even 50,000$, is not sufficient to fund any hard research.

The mode reader is also perfectly aware that if they submit, and their proposal is not the one selected, even if it solves the problem, they did work without being paid. Did work, gave away value, yet while not rewarded, nothing prevents the bounty makers from using their proposal also: it was only not rewarded because it was deemed not the best proposal. In fact, all the proposals, besides the best, are free gifts to the bounty makers.

At most only one proposal will be rewarded. Indeed, that's the reason a contest is made: you give a reward to those who comply, and fail to reward those who do not comply, but this is costly, if only a few do not comply, so instead you reward the single best proposal. This still draws most to comply, but only one is rewarded. You pay for only a small part of the work done, yet the effect to motivate work is as if you paid for all work. At least for naive producers.

Well: many scientists are not naive. They see through contests. Few real researchers submit. Other than when bored.
The bounty or contest structure fails to attract the very best effort. Because it's based on most valid contributers not being rewarded for their work. Which is why contests are preferred: you get slightly less work than if you just rewarded work done. But you reward much, much less of the valid work that's done at your request.

So how to split bounty rewards amidst all contributers, to make a bounty platform that will actually get participation from no rush, relatively wealthy, and sophisticated producers, is a good use case.

 

— 〈  2  〉—

(2) Issues:

(2a) What exactly, other than a cool factor, does VR have to do with R&D and science funding? Anything?

(2b) Scientists are few: they collectively have expenditures far in excess of their collective resources, so the token, with them being the exclusive active users, will always have low value. Mostly they would redistribute from each other; but collectively they do not have enough to make that worthwhile. Companies will not invest much, however, in a low liquidity, risky, low value, asset. They have low risk, high liquidity, high reward assets: private placements. Investments in startups. Finance.

(2c) Most scientists that are currently around don't work for any reward. The ones who needed to be paid to do science, end up not doing science.

Science is the most expensive sport around.

It's almost pure positive externality.

There is almost no way to make any return on scientific labor; most return that appears to be from scientific work is primarily due to the finance, and is merely a single developmental resource. If present, the finance guys can earn a lot. If absent the finance guys can earn nothing. But the scientist almost never earns anything, because most value contribution with be from the finance and marketing side. Take the machine learning field for example. Valuable: but rebranded discoveries in cybernetics from the 1950's and 1960's, almost all marketing. The science itself cannot earn reward.

So the scientists who contribute mostly are those who don't really need any reward.

Almost all the ones who needed payment for their work were squeezed out of the field by poverty. The utterly poor cannot have the leisure needed to perform research. 99 % of conjectures are failures, 1 % are successful. High cost in time and resources up front.

You need years merely studying. Otherwise you cannot be competent at a really high level. Dangerous to say that. People get offended. They think that having a science related job is being a scientist. No: getting results is being a scientist. And publishing is being a scientist.

But then you can only solve hard problems by accident. So probably won't solve any, and therefore little reason to participate.

So the problem of rewarding scientists is for the next generation, if any, of funding science to scientists who earn a living doing science. But the current generation must have a reason to buy into the platform; they don't. So most will probably not buy into the platform.

(2d) Do they have peer review of peer review of ... it's needed. Peer review has a problem when some peer review is invalid. Which is actually quite often; but not a huge problem because many journals and not any money involved. But here there is money involved and one platform.

 

— 〈  3  〉—

(3) The good and the bad:

They might still get funding. One token for AI raised 30.000.000$ in their ICO. This month I just realized the same thing can be done easily on Steem; right now, no special token needed. 30.000.000$ not needed.

They need an institutional investor who'll invest a little, by mistake, and then convince a larger institutional investor to come in, and fund their platform. It's their only hope. And if they can get over that hump, maybe something interesting might happen. Then they overcome, by accident the problem of why ought scientists buy into the platform, and where is the token going to get initial value from, and maybe some will use it.

Somehow well known writers, I agree, must contribute to the system. But they have no reason to contribute to the system, rather than to submit to a prestigious but paywalled journal. Even open access for a large fee, if the journal is prestigious. The institution behind them will foot the bill. And libraries spend budget; the readers of journals do not pay for them, so they don't really care much how much journals cost. If they complain, they complain in principle.

(Though paywall is a problem in Europe, I'll admit. Not all universities have access to all publications of all publishers, while in US or Japan, everybody has free access to every paywalled journal.)

And monetary reward is too small to be relevant. Moot to the current generation of high levels scientists.

 

— 〈  4  〉—

So the platform can only get users by prestige, not by rewarding users, but without users, it can't get prestige. Catch-22.

The solution is that somebody networked with several well known writers and themselves publishing in otherwise prestigious journals must be the founder of the next open system. Or some large investment must come in by sheer accident, as discussed above and jump start the use case process.

So with Matryx you have a risky bet. But it's still possible it'll get 30.000.000 or something like that, because of the (i) cool factor, (ii) an ignorant but networked first investor, and (iii) a circle jerk by institutional investors. Just discount that future value by the frequency of such investments in the last 3 years (write off period) compared to all tokens whose hope is primarily i, ii, and iii.

That would be the theory.

If the expectation is below my minimal desired 3 year return, based on my taste for risk, I would sell, and if above, hold.

That is what I would do; not advice.

After 3 years, the return might be enormous, but it's considered a write off. Because if I was more risk averse, there are plenty of disproportionately less risky investments with same minimal 3 years return.

 

— 〈  5  〉—

All the VR stuff will either annoy the experienced large investor, or wow the incompetent large investor, so again, there is some indeterminism here. Like the famous example that nobody was doing much plain paper copying did not imply that Xerox would not make hundreds of billions selling plain paper coping machines.

There were no good plain paper copying machines.

So nobody did much copying. Which suggested Xerox would fail.

That when reason, contrary statistics, would suggest it would succeed. And it did succeed. Massively.

So the future is not like the past. Correct.

But since here the success will depend on initial big funding from a few large investors, drawn from the same population as in the past decade, rather than organic users, frequency of funding the like may be a proxy for probability of success.

We discount future value accordingly. Forming some expectations.

All this is theory. Not advice. One cannot give advice. Risk preference is not known. Budget to hedge is not known.

Risk is defined for a bundle of every investment. So be careful.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License  . . .   . . .   . . .    Text and images: ©tibra. @communicate on minds.com