In its heyday, using Twitter as an aggregator of alerts in a place that I was already receiving casual updates from was incredibly convenient. Tying that to conversation, it was easy to discern if this was a global impact caused by, for example, a failed deploy, or local, say a failing CDN endpoint. I truly don't think this is a case of folks opting solely for convenience over accuracy, so much as it added both convenience and context. Not to mention it was a useful canary in the coal mine, as declaring incidents on large platforms trigger SLA contract provisions, so we often see quite delayed reports from first-party sources.