He's right that the current blocksize is coping with present demand, and that 1MB can serve near-term demand to some extent. We only have to look at what happens when the spam attacks stop; daily transaction volumes fall comfortably back down between 250,000 and 275,000 tx per day. Sometimes lows closer to 200,000 have been reached, although that never lasts long.
The fact is that the fee market for on-chain transactions is important. When blocks are not quite full, the miners can decide which transaction they process, and in turn the fees they are willing to accept. Today's 6.25 BTC bock reward helps the miners to decide in the users favour, but I think it's been amply demonstrated recently that miners have a penchant for taking decisions against users interests, in spite of likely damage to their own interests in the mining market (one wonders why they would choose self destruction to hurt the users' interests)
But, with permanently full blocks, the users must fight to compete for the blockspace with their fees. This creates a secondary incentive effect also; the miners lose profit margins to their competitors if they decide they don't care so much about fees. It's a positive feedback loop between user fighting for blockspace, and the miners fighting for profits.
When blocksize is so high that miners can afford to ignore transactions (because the profit margin is in the block reward, not the tiny fees they could earn), the Bitcoin system starts to lose this property of economic incentives dictating fees. Users don't need to fight for blockspace, because there's plenty of space going unused. But miners don't have to provide the space, they won't earn much from it anyway. The only reason miners might include adequate transactions in blocks under those circumstances is plain goodwill, and there's a danger a negative feedback loop of disincentives could develop. Even now, with fairly full blocks, certain pools still cynically produce entirely empty blocks, so there is evidence that maximum competition among miners is vital to the Bitcoin network's smooth running.
You can safely ignore what luke-jr believes. Nobody takes him seriously, not even other developers on Core.
Explain why, and which developers feel that way. Provide evidence, not assertions. Why should we take your unsupported proclamations as facts?
I understand why people criticised Luke-jr's 300 KB blocksize proposal.
Luke was trying to be too clever for the audience, and it backfired on him rather badly. His idea was that no-one would accept 300 KB today, but because of the effects of time on the size of blockchain, maybe people would begin to accept his point that even 1MB is too big for today, but maybe in 2 years or 4 years from now. Because in 2 or 4 years, Luke's blocksize increase schedule would have decreased to only 400-600 KB, and Luke expected the smaller decrease would become more appealing once everyone has had 2 or 4 years more to think about it. (Luke's proposal finally gives a bigger blocksize than 1MB in 2024)
So there is a lot of psychology involved in Luke's proposal, and I think that's where it unwinds. Like I say, he was simply trying to be too clever. But who knows, maybe Luke's judgement that people will begin to take his proposal more seriously once 2 years (~120 GB of full blocks) or 4 years (~240 GB of full blocks) passes will come true. That would mean a 240GB blockchain in 2019, and a 360GB blockchain in 2021 (and that's all assuming we still have a 1MB blocksize limit at that point), maybe that would motivate everyone (especially considering that Segwit will likely be activated by then, we may as well multiply those figure by 4 in which case).
Will more efficiency tweaks to the Bitcoin software, better internet infrastructure and cheaper/faster storage be enough, or will the hard limits of those factors be enough for people to consider reducing the blocksize? I guess only time will tell.
