I've been waiting for evidence that this happens in non-neglible numbers for months.

Wikipedia's Covid-19 article links to ongoing studies:

https://www.itv.com/news/2020-08-12/long-covid-what-is-it-wh...

https://www.webmd.com/lung/coronavirus-complications

From WebMD: "Some people -- about 1 in 6 -- will have complications, including some that are life-threatening."

From ITV: "a study by King's College London found an estimated 10% of people with the virus take at least three weeks to recover, with 250,000 people in the UK alone thought to experience symptoms for 30 days or more. Many of those experiencing long tail coronavirus were fit, active people whose lives have been completely turned upside down by the ongoing symptoms."

We are talking about disabilities, i.e. permanent (or at least long-lasting) and severe damage. Taking a few weeks to recover does not qualify.

I could end up being very wrong, but it's probably uncommon or else it'd surface somewhere other than support groups for those people and anecdotal accounts in local news stories. Enough celebrities and politicians have gotten it that if this were widespread, they'd be sounding the alarm en masse and we'd be hearing all about it.

Off the top of your head, do you recall any stories of "I had a bad case of Covid and am 100% now"?

I think there is an assumption that if you contract something that puts people in the ICU, you should expect your life to be permanently altered by it. I suspect that assumption is helping mute attempts to collect and quantify that data.

Pretty much anything will put someone into the ICU in the right circumstances, which is why quantification of risk matters. Flu puts people into the ICU. Drinking too much water can put you into the ICU. Those are much smaller risks though.

Absolutely. Waiting for evidence is what one does when stats aren't being compiled.