Thoughts from someone who may not matter due to not being much of a “non-professional” video watcher (I’m one of those older & casual viewers, albeit slightly more technically knowledgeable than many of those you’re referring to):

1. If this achieves the success for PT that you hope for, the cost of being the gateway to that content would still become significant.

2. Not being a big video watcher, I don’t know if this actually is as much an issue as I seem to experience on Lemmy, but the slow loading of the list of available posts to view that I assume is endemic to how the Fediverse works may prevent this from taking off among the casual YT viewers you’re looking to hook. Well-done caching strategies may help minimize this, but now we’re increasing the costs of running the gateway again.

3. Eventually some small portion of those users may want to be able to upload videos while also having them tied to their account’s activity history (comments, public playlists, etc.), and will be put off by the fact they can’t. What then?

I’m not saying it’s a bad idea per se - I could certainly see myself going for it since I’m unlikely to ever want to upload videos to share with the world at large - but if you’re going to make yourself the public face of PT, and yet prevent some of its functionality then you might just wind up damaging the brand you’re trying to promote.

One way out I can see is to make uploading a paid feature rather than preventing it entirely. I know that theoretically goes against the desired appearance of “openness” in the platform, but the costs are a reality that can’t be denied, and shouldn’t really be a major stumbling block if kept as reasonable/cheap as possible. Those who philosophically are against it (or just can’t afford it) can still make accounts elsewhere that they can refer their followers (if you support such a feature) to.

Just my somewhat uninformed two cents.

I just had a couple other thoughts on covering server expenses to allow uploads.

1. You could publish server maintenance costs, and solicit donations and/or “memberships” that give priority access to available bandwidth. Sharing your actual costs and revenues from the above would help prevent fears of unreasonable profit-taking abuse among users. Any excess profits could be placed in an interest-bearing account to generate more funds to support activities, and act as a “slush fund” for future upgrades, repairs/replacements of hardware, and/or to cover periods where funding shortfalls occur due to things like a souring economy.

2. You could give users the option of viewing ads to earn credits towards hosting their videos. Again, many people won’t like this, but costs are a reality they can’t really argue, and making this just an optional method of support means they simply don’t have to use it if they don’t want to. Finding an advertising company willing to do such small-scale non-targeted advertising while still sharing some of the revenue for it is admittedly probably going to be difficult, however. I think starting by looking at who the smaller-scale search engines such as Qwant and StartPage work with might be useful here. Overall, this approach honestly isn’t likely to generate much revenue (at least initially) so it may not be worth the effort, but I don’t know enough to say for sure - it’s just an avenue to consider that might work better if & when you achieve some significant growth.