On July 28th 2015, just two days before the ethereum genesis block launched, an etherean said: “I was under the impression that the switch to PoS would happen rather soon (~1 year).”
In reply, Buterin stated at the time: “Yeah, the plan is 9-12 months, taking into account Hofstadter’s law.”
That law says: “It always takes longer than you expect, even when you take into account Hofstadter’s Law.”
However, the ethereum development team was not quite working on Proof of Stake in 2015 or 2016.
The latter year had the Slockit DAO and a bit of a mess, but even in 2017 there wasn’t much focus on PoS.
As a hybrid version of it was about to launch in mid-2018, the whole thing was completely scrapped in a surprise move.
A new blockchain was then to launch by the end of this year or by January the 3rd, but that’s still just hybrid PoS and, more importantly, it hasn’t launched.
The estimate was then Q1 next year, which has become Q2, with full Proof of Stake unlikely before circa 2021.
That makes it months not years, with another interesting revelation from the time on the obvious relationship between block times and issuance. Buterin said:
“We promised in the ether sale agreement to keep the reward to below 26.00% of the amount of ether sold per year. At 60m ether, that means 29.71 ether per minute.
Assume 5 ETH per block, 1 block per 14 seconds, and an uncle rate of 0.4, consistent with conditions under high load. That gets us to 30.0 ETH per minute, pretty much the same as that maximum.
The ether sale docs specifically said a maximum of 26.00%, and not exactly 26.00%, and I felt it was important to hold to that promise and target 26% as a reasonable upper bound and not an average.
Under conditions of lower load, yes the reward is going to be a few percent lower.”
They were planning 13 seconds per block, but because of the above ICO promise, they changed it to 15 seconds per block.
Coincidentally earlier this year block times were reduced to 12 seconds per block from the 15 seconds.
They are now kicking out the difficulty bomb through a hardfork on New Year’s Day for two years. Making that block times decrease a stealth increase of ethereum’s inflation rate.