1 2014-02-24 00:00:31 <todamoon> so far, the best potential solution I've heard is the use of off chain transactions
 2 2014-02-24 00:01:15 <todamoon> or a hard fork that would address those issues if/when they become critical
 3 2014-02-24 00:04:58 <todamoon> gmaxwell: do you see those as potential issues?
 4 2014-02-24 00:05:36 <gmaxwell> There are many potential issues, and all of these things exist in a complicated space of many tradeoffs.
 5 2014-02-24 00:06:13 <gmaxwell> I am more concerned about the centeralization of Bitcoin, since centeralizing it moots its purpose. Navigating parameters that yield the best effective decenteralization is hard.
 6 2014-02-24 00:06:44 <todamoon> centralization by miners and large shared wallets?
 7 2014-02-24 00:08:00 <gmaxwell> among other causes— it's not like additional blocksize is free. or that difficulty can sustain itself to preserve POW as a working security mechenism without non-trivial fee income.
 8 2014-02-24 00:08:51 <gmaxwell> We've already violated the original security model with the situation we have now where single computers have commanded 1/3rd to as much as 1/2 of the hashpower.
 9 2014-02-24 00:09:58 <todamoon> presumably, it's in the best interest of miners to not exploit their power but I understand your concern
10 2014-02-24 00:10:11 <todamoon> do you think proof-of-stake would be a viable alternative to POW?
11 2014-02-24 00:10:43 <gmaxwell> thats like arguing it's in the best interest of the fed to not exploit their power, and it ignores the issue when someone compromises that infrastructure (which has happened and has been used to rob a bitcoin site)
12 2014-02-24 00:11:03 <gmaxwell> (part of the reason that I find things like that more concerning is because its a quiet problem; vs things like fees or transaction throughput are immediately self-evidence)
13 2014-02-24 00:11:33 <gmaxwell> todamoon: So far no one has demonstrated a PoS formulation which is a viable consensus mechnism on its own, I suspect it's not actually possible, which is indeed a bummer.
14 2014-02-24 00:12:33 <todamoon> I see
15 2014-02-24 00:12:55 <gmaxwell> todamoon: (The PoS mechenisms all suffer from the nothing at stake problem: PoW encourages consensus because you forever burn a valuable resource to mine it, so you want to burn that resource on the one unique chain most likely to survive. In PoS the optimal rational strategy is to mine all possible forks that you do not hate. Results in some funny attacks where you PoW-like grind the PoS to find the history where you are the lucky ...
16 2014-02-24 00:13:02 <gmaxwell> ... winner for every block.)
17 2014-02-24 00:13:11 <todamoon> it just seems inevitable that there is no technical means to avoid miner centralization though
18 2014-02-24 00:13:18 <todamoon> only social means
19 2014-02-24 00:13:45 <todamoon> I see
20 2014-02-24 00:14:02 <gmaxwell> todamoon: well there are technical means but they don't have the right tradeoffs, I expect. There isn't a lot of actual advantage in the centeralization we have today, it's just symmetry breaking that went wrong as far as I can tell.
21 2014-02-24 00:15:03 <jcorgan> i think it is an inevitable effect of differences in mining operation efficiencies and economies of scale, visible in many industries, that the market converges to a power-law based size distribution; the distribution of hash power seems to fit this.
22 2014-02-24 00:15:32 <gmaxwell> (it's perfectly possible to pool mining without pooling control of the netwrok, but thats not what got implemented. Changing the status quo is hard, esp now that you've got miners with >100k of mining gear running it off a rpi.)
23 2014-02-24 00:16:28 <todamoon> right now, pools "dictate" the block to miners?
24 2014-02-24 00:16:34 <gmaxwell> jcorgan: centeralization in mining today has relatively little to do with the centeralization of hashing. The largest entity I'm aware of has ~7% of the network's hashpower, compared to pools in the 30-50% range. 7% seems less concerning to me.
25 2014-02-24 00:16:46 <sipa> i'd rather say it's hashers who sell their right to vote to a pool
26 2014-02-24 00:16:59 <gmaxwell> todamoon: may be better to call the people who are running mining hardware "hashers" not miners.
27 2014-02-24 00:17:23 <gmaxwell> todamoon: for all pools except P2Pool, effectively.
28 2014-02-24 00:17:28 <todamoon> I see
29 2014-02-24 00:18:03 <todamoon> I was under the impression that pool hashers were free to build their own block so I didn't see the problem with pools controlling a large part of the hash rate
30 2014-02-24 00:18:17 <sipa> that's how it should be
31 2014-02-24 00:18:29 <jcorgan> gmaxwell: i'd including pools in that distribution
32 2014-02-24 00:18:33 <sipa> but no, hashers usually don't know what block they're really working on
33 2014-02-24 00:18:56 <gmaxwell> But thats not what got implemented, I don't think anyone really realized that was possible at first. Getwork sort of encourages thinking in terms of controlling the whole thing.
34 2014-02-24 00:19:28 <gmaxwell> jcorgan: sure, but thats not necessary, it's perfectly possible to pool for payments without handing over control. It just isn't what got implemented.
35 2014-02-24 00:20:07 <todamoon> how many core devs are paid for their work by the way?
36 2014-02-24 00:20:22 <todamoon> I understand Gavin gets a salary from the bitcoin foundation?
37 2014-02-24 00:20:30 <gmaxwell> todamoon: Gavin and Wumpus recieve paychecks from the Bitcoin foundation. Jgarzik works for Bitpay.
38 2014-02-24 00:20:39 <todamoon> gmaxwell: thanks
39 2014-02-24 00:21:55 <jcorgan> gmaxwell: agree; i think your symmetry breaking argument and my hash rate distribution argument are orthogonal
40 2014-02-24 00:32:55 <jgarzik> tedtalks, hooray for IPv6