I assume you failed to read the what
I posted, for you seem to be unaware that (both I2P and Tor admit on their websites that) timing attacks only require the adversary to control the router at the entry and exit nodes in order to de-obfuscate the IP address of the user (or possibly for several users in the mix depending on the statistics of the low-latency and other factors), there is no need to invade the user's computer with exotic methods you mentioned.
Even hackers can (sometimes) do that.
I understand perfectly well what you've written, you seem to have misunderstood what I have written.
First, stop talking about Tor, we've never mentioned Tor and have already rejected it due to the problems inherent in their use of exit nodes.
Now: we are not aiming to prevent a global adversary who has control of all bordergates at all ISPs on the planet from making assertions and observations. We are aiming to prevent them from knowing
whether Monero is running or not. They can know I2P is running, that we don't care about, but they should not be able to determine whether the person is harmlessly browsing the web or using Monero without completely decrypting the traffic.
Why? Because a wannabe "global police state" (without resorting to the exotic methods I described) cannot possibly go after millions or billions of people around the world under the control of regimes and governments of all sorts merely for using a communications system.
I understand that you're a terribly clever and opinionated chap, but don't assume that we are ignorant or that we have made decisions for no good reason. I2P integration is for a specific purpose and is fully expected to be a stopgap if its shortcomings are not solved over time.