I agree with him that the block size limit is an anti-spam feature and should be treated as such, but the question is why (for example) 20 MB of spam is considered acceptable now. I see no good reason to allow 20x as much spam, when little to nothing has been done to control spam in any other way.
according to the new XT code today, the max would be 8MB, with doubling every 2y if 75% miners up-version.
you'd have to believe, for all the same reasons we've argued against for yrs, that a large miner would be willing to risk paying 8MB worth of spam tx fees for a prolonged period of time to try for the fuzzy goal of driving smaller miners out of business over an unspecified period of time, which may never come, and which comes at great risk to themselves from orphaning.
Not necessarily. It could be more of a gradual process of spam being accepted into a system that isn't well protected against spam. The miners who are better equipped to produce somewhat larger blocks produce them (by including lower value, lower fee transactions aka spam), which puts more of a strain on smaller miners who drop out. As smaller miners drop out, the average block size then increases, which puts even more strain on the slightly-less-small miners, who then drop out, repeating the process.
The end result is larger spammier blocks, and less broad participation in mining.
The fact remains that as I said there is no more of a mechanism to control spam
other than a block size limit than there was when satoshi added the block limit, and I don't really see how spam is
20x8x less of a problem or potential problem now. I'd support a smaller increase though, since I do see technological progress as having reduced the threat of spam by a smaller factor.
Longer term I see the 8 MB and auto doubling every two years as even worse than 20 MB. There is no rational basis to believe with high confidence that gigabyte blocks would be acceptable in 15 years.