Good catch, exchemist. Yes, it is academia.edu. Yes, it is a business.
So far as I know, they allow you to have a profile and post papers without charge. That was true at least when I joined the site. I think it's still the case. However they provide a paid membership option that enables more features, which includes more information about who is looking at your posted work, if they are also members. I signed up for that a couple of years ago. Apparently the AI podcast feature is one of these that applies to paid members only. It just appeared on my profile recently, and I thought it was talking about podcasts that I might or could make myself, the old-fashioned way, but turned out that it generates them in a few seconds if you enable them to be put on your profile.
Most indexed journals are wanting about $2K to publish a paper these days and those that don't are unwilling to review my papers now. My papers now are a lot more advanced than earlier ones, but it's harder to get them published. For example, Foundations of Physics reviewed three of my papers and published one in 2016. Now they don't send them out. Anyhow, the whole academic publishing structure today seems like pay-to-play and there are of course many truly predatory journals that want as much but aren't even indexed.
Academia.edu's AI does a nice job, I think, of pointing out that my papers are counter to conventional wisdom without dismissing them out of hand. It also takes pains to say that it isn't endorsing them or checking the math. It would be nice if an AI review checked the math, though. Then, peer review could be made mostly automated. At least, a journal could at least run an AI peer review prior to deciding to desk-reject. Hopefully that will come along before too long.