4 ms·

I've been thinking the same thing over the last few months... I see all of these "prediction markets" and "gambling platforms" being launched and claiming that

I've been thinking the same thing over the last few months... I see all of these "prediction markets" and "gambling platforms" being launched and claiming that they are "fully autonomous". Well where do they get the outcome information from? It must be from a person. You cannot program a smart contract to know who has the most reliable sports scores or political information. AI does not exist yet. If I bet that Hillary Clinton won the election, and she lost to Trump, what's stopping a hoard of Clinton supporters from voting (or tweeting or however the DAO collects their "data") that Clinton actually won the election and stealing my winnings?

jackfraser 9y ago
> You cannot program a smart contract to know who has the most reliable sports scores or political information. You can't do that as a human, either. All you can really do is look at someone's record to see how their predictions have panned out historically. Given the right data sources, a program could do this no problem, and assign metrics to individual sports commentators as well as subsets of their consensus on any particular topic. Given the nerdy propensity of the fantasy sports guys, there's no reason to assume this kind of data won't be available eventually. As for politics, I'll agree with you there - neither human nor AI is going to be able to produce particularly reliable commentary any time soon! > If I bet that Hillary Clinton won the election, and she lost to Trump, what's stopping a hoard of Clinton supporters from voting (or tweeting or however the DAO collects their "data") that Clinton actually won the election and stealing my winnings? This is called a 'coup' when it happens IRL, and happen it does. The chief difference here is that in reality, pitchforks and torches are needed for these efforts; online it's a little bit physically safer.
- PeterisP 9y ago
  It seems like you're talking about predictions, but the point of "reliable sports scores or political information" is about a trusted source about past/current facts. People have no problem verifying the bet results of yesterday's game or last year's elections. However, making a blockchain that can't possibly be manipulated into thinking that Clinton won the election (no pitchforks needed, you're misleading a single automated system, not everyone else) is actually much harder than it seems - especially if the decisions are enforced by the system, final and irreversible and can't be overruled by courts if it's found to be fraudulent afterwards.
mthoms 9y ago
By participating you're agreeing with whatever mechanism has been coded into the contract (which you can read for yourself). Obviously, you'd be stupid to participate in a contract that relies solely on twitter for resolution. But, for sports scores the contract could be coded to check all three of espn.com, nfl.com and abcsports.com (just for example). Same with the election results: usa.gov, cnn.com, foxnews.com, bbcnews.com You make a valid point I think, but the examples you gave are pretty trivial. More complex ones are another matter.
- Slartie 9y ago
  Actually, it can't. In no existing blockchain-based smart contract system is contract code able to request random data from websites and act upon it. Websites as external sources are not deterministic, and determinism is necessary for the computations to result in the exact same result on any of the nodes that are part of the network in question. This also includes - and this is forgotten quite often - determinism across an eternal timespan! It can't matter when I want to recompute a certain contract transaction - it always has to return the same result. The only known solution for these problems is to have an oracle inject the necessary "knowledge" into the blockchain itself and then have the smart contract code work with this data, that is then stored on the chain forever and thus makes everything deterministic again. But the oracles come with their own sets of problems, like them just being another "trust anchor" (and they're one that you can't validate the code of, since the oracles themselves by definition have to reside outside of the blockchain, it's just the data that ends up within it) or the giant pollution of the blockchain with data that they cause. Even these very simple examples don't pass the reality check on close inspection.
  - jjn2009 9y ago
    this is actually incorrect you can verify a tls connection was made to a server and a particular output was given in that tls connection with oraclize it.
    - Slartie 9y ago
      Which is just an implementation of precisely the oracle scheme I talked about. Introducing precisely the problems I mentioned: blockchain bloat, additional trust anchor which can never be 100% validated and thus destroys the idea of requiring zero trust in any particular individual person or system.

jackfraser 9y ago

> You cannot program a smart contract to know who has the most reliable sports scores or political information. You can't do that as a human, either. All you can really do is look at someone's record to see how their predictions have panned out historically. Given the right data sources, a program could do this no problem, and assign metrics to individual sports commentators as well as subsets of their consensus on any particular topic. Given the nerdy propensity of the fantasy sports guys, there's no reason to assume this kind of data won't be available eventually. As for politics, I'll agree with you there - neither human nor AI is going to be able to produce particularly reliable commentary any time soon! > If I bet that Hillary Clinton won the election, and she lost to Trump, what's stopping a hoard of Clinton supporters from voting (or tweeting or however the DAO collects their "data") that Clinton actually won the election and stealing my winnings? This is called a 'coup' when it happens IRL, and happen it does. The chief difference here is that in reality, pitchforks and torches are needed for these efforts; online it's a little bit physically safer.

PeterisP 9y ago
It seems like you're talking about predictions, but the point of "reliable sports scores or political information" is about a trusted source about past/current facts. People have no problem verifying the bet results of yesterday's game or last year's elections. However, making a blockchain that can't possibly be manipulated into thinking that Clinton won the election (no pitchforks needed, you're misleading a single automated system, not everyone else) is actually much harder than it seems - especially if the decisions are enforced by the system, final and irreversible and can't be overruled by courts if it's found to be fraudulent afterwards.

mthoms 9y ago

By participating you're agreeing with whatever mechanism has been coded into the contract (which you can read for yourself). Obviously, you'd be stupid to participate in a contract that relies solely on twitter for resolution. But, for sports scores the contract could be coded to check all three of espn.com, nfl.com and abcsports.com (just for example). Same with the election results: usa.gov, cnn.com, foxnews.com, bbcnews.com You make a valid point I think, but the examples you gave are pretty trivial. More complex ones are another matter.

Slartie 9y ago
Actually, it can't. In no existing blockchain-based smart contract system is contract code able to request random data from websites and act upon it. Websites as external sources are not deterministic, and determinism is necessary for the computations to result in the exact same result on any of the nodes that are part of the network in question. This also includes - and this is forgotten quite often - determinism across an eternal timespan! It can't matter when I want to recompute a certain contract transaction - it always has to return the same result. The only known solution for these problems is to have an oracle inject the necessary "knowledge" into the blockchain itself and then have the smart contract code work with this data, that is then stored on the chain forever and thus makes everything deterministic again. But the oracles come with their own sets of problems, like them just being another "trust anchor" (and they're one that you can't validate the code of, since the oracles themselves by definition have to reside outside of the blockchain, it's just the data that ends up within it) or the giant pollution of the blockchain with data that they cause. Even these very simple examples don't pass the reality check on close inspection.
- jjn2009 9y ago
  this is actually incorrect you can verify a tls connection was made to a server and a particular output was given in that tls connection with oraclize it.
  - Slartie 9y ago
    Which is just an implementation of precisely the oracle scheme I talked about. Introducing precisely the problems I mentioned: blockchain bloat, additional trust anchor which can never be 100% validated and thus destroys the idea of requiring zero trust in any particular individual person or system.