Major lessons learned:
1) I was playing Amph bots as a land faction that could occasionally retreat into water, and failed to analyze the terrain early in the match. For example, I put way too much emphasis on the fords even though my army could have climbed onto the central island almost anywhere.
2) Minute one and two energy should focus on defensible placement. But defensibility gets much better after the raider phase. I gave up on large scale tidal farming after one demolition by GoogleFrog's Ducks, and it crippled my economy in the long run.
3) When advancing bot raiders against skirmishers, I should do a lot more attack move and less right click.
4) Don't 1v2 on offense without a massive advantage! I lost a Duck army pointlessly that should have disengaged after it saw both Scalpels and Venoms present.
5) While I did build some air defense this game, it usually lacked a critical mass to actually get kills. If 1/4 of my opponents are air, about 1/4 of my fighting budget should be AA.
6) I should have set up my Grizzlies to use a retreat zone.
7) The first nuke flew shortly after 24:00. This is a good timing to keep in mind for Antinukes in the future.
8) It's quite possible to triangulate missile silos from their attacks' trajectories.

giga don't care, micro-optimization for the win (saying as a fun subgiant)

Point 2 is very much about expansion. I lost my initial worker and almost half my non-commander energy in a way that I wouldn't've if I'd just stuck with safe solar instead of being greedy with the tidal generators and trying to grab an underwater mex too soon vs. GoogleFrog's ducks. This also cost me any early game pressure because I lost my own Ducks trying to defend the energy.

I disagree that the game was in any way over at that point, though.