Just trying to keep up with demand could explain it alone. The strict quotas (even for Plus users), the occasional errors and long waits etc. etc. indicate that

Just trying to keep up with demand could explain it alone. The strict quotas (even for Plus users), the occasional errors and long waits etc. etc. indicate that the hardware performance has clearly been a bottleneck all throughout. And demand has exploded since, so it's plausible that they had no choice but to sacrifice quality, if they wanted to keep the service running.