Bananas are a bad example of this. They don't follow the ordinary selection processes during domestication - the ones we commonly consume are hybridized cultivars with triploid genetics; they're infertile. They reproduce asexually, so all bananas belonging to the same cultivar are clones of each other.

More generally, yes, over time you are going to get more of what you're selecting for. Which may or may not be flesh:seed ratio.

They've been working on the same with açaí with some success.

You make that sound like merely an observed behaviour, but isn't that because that's what we're selecting for? We eat banana flesh, not banana seeds, so we're cultivating varieties that are fleshier (among other desirable properties, such as productivity, and taste of the flesh of course).

But isn’t that what he said? To put it in your words: we eat dates, not date seeds, so we‘re cultivating varieties that are fleshier. Less seed, more date.

Perhaps I misread, but to me 'tends to go up' makes it sound like an observed incidental effect, rather than a deliberate action.

Essentially I was saying 'sure, there's a correlation, isn't there also causation, aren't we the cause?' (which isn't the way around that's usually said!)

It was my suspicion that the change is a result of deliberate human action, but I am not a biologist or a farmer so that's not something I'm certain of. The trend, not the cause of the trend, is what seems more important in this case.

I think everyone misread, including me.

Everyone thought he was saying "are you sure it suggests that? It seems an odd to me because flesh grows over time, so it is a really bizarre conclusion if true!"

But OP didn't mean that. Instead this is what OP meant:

">> The successful seeds were all several centimeters long, 30% larger than modern date seeds, suggesting dates that were significantly larger than modern varieties.

"Actually, maybe it's not that the flesh was larger - but rather, the seed was! The flesh could have remained the same size."

OP then justified why he thought this was just as plausible, by mentioning wild bananas (Google image search "wild banana seed") as an example of large, ugly seeds in a same-sized plant.

i.e. the seeds could have been cultivated to be smaller over time.

It's possible larger varieties previously existed and fell dormant, then later the domestic varieties started with a smaller fruit and the genetics for larger fruit were no longer in the active gene pool of known varieties.

Strictly speaking, we aren't cultivating banana varieties any more. In the last 100 years there have only been two that have been commercially successful, Gros Michel and Cavandish, and they are grown from cuttings rather than seeds. The Gros Michel was nearly completely wiped out by fungus, and the Cavendish is living on borrowed time for the same reason. Hopefully someone will come up with a new variety soon.

Possibly, but ...

The cool thing in this instance is that the trees are growing, so we may get to observe them first hand.

It'll be a while before we see any fruit produced from them to know for sure. Makes sense to me that modern date seeds would be smaller, especially considering how huge they already are.