The point that changes happened doesn't surprise me - and if I was on the developer side I'd push for the same. The old cards, regardless of intent, in practice funneled players into the practice of dismissing a good half of the cards (if not more). Even if the Ewer and Spire weren't useless, I would say that most players saw them as "failures" when drawing. That's probably why the new cards are so homogenized in their effects. SQEX want all cards to feel useful, for players to always feel like they've drawn a prize.

I firmly believe that their current implementation is bad and misguided.

Variance in card effects is important to convey the feeling of choice and agency. Yet such variance is what high-end play strives to avoid; having a DPS or mitigation buff randomly available is terrible in the scripted encounter environment we play in. The solution to this isn't homogenization of effects. Rather, we should have been given tools to manipulate the draw results beyond the simple "redraw" and force the result - albeit at a cost further down the timeline. In physical card games, you are sometimes allowed to draw from your deck until you come across a card you're searching for. The associated cost could either be a game-specific resource such as "mana", having to discard the cards you've moved past and risk emptying your deck, or revealing information to your opponent. In the same way, I believe it suitable for an AST to force a result and pay in some way down the line: having to play the card you're overwriting without a chance to reroll, take a resource penalty, or suffer reduced effects in what you're forcing.

Overall, the approach taken by SQEX seem to point towards a preference of homogenization in an attempt to "balance". Whether this is due to lack of resources or a lack of vision is unclear.