I still think you're going to struggle with the ROI issue, if delegators need to share with authors of old posts then they're likely going to be getting lower returns than the de facto 100% self-voting of the currently-established bots. I think this is potentially even worse if some of the old creators aren't even active on the platform anymore, I think delegators will find it hard to see the value proposition there. Maybe have the selection criterion focus on old posts from currently-active creators, so there's a clearer link for how the incentives could work?

On Substack Notes, the platform's social-media layer, they recently implemented a change to have it add a little "from the archives" tag when it surfaces older content into your timeline, that changed my subjective reaction from something like "why is it showing me these years-old posts!?" to "that's cool, they're trying to get people to not treat the platform as ephemeral as other social media". So I think that supports the notion that showcasing older content can make sense, but figuring out how to do it smoothly can take some work. They've said that on Substack they're trying to optimize their algorithm for paid subscription signups. Maybe tracking follower growth for Thoth-featured creators would be a useful metric for you to track to judge how well its working.

First, if Thoth is good at picking out content, the investors should get back rewards from other voters - not just from their own delegations.

I'm not sure there's enough organically-voted SP for that to work out.