One problem with the "strict mode" solution is it's set on the client. It means you have to be careful about what clients access the DB, or one misbehaving tool could ruin it for everyone. Altho the improvements are what we've all been asking for, they kind of feel "bolted on", whereas they were fundamental design goals of Postgres from the beginning.
Ultimately, both are good choices though. Open source is about giving us more choice, not less. I think the only thing is some people see the success of MySQL as disproportionate to it's quality, when compared to Postgres. (But maybe it's just an example of "worse is better").