The voting does cut down how many sets they have to review. There's a ton of just absolute drek among the IDEAS proposals.
But the voting based on a 10k threshold no longer cuts the volume down enough. The number of proposals hitting the mark is seemingly beyond LEGOS' capacity to really review all of them, let alone produce the sets. Other than separating obvious wheat & chaff, I argue the voting to a simple threshold doesn't impart enough information. Thousands and thousands of people vote for every way-too-expensive megabuild, thousands and thousands of people vote for nostalgic IP, and potentially thousands of newcomers not really associated w/ LEGO crash the party if some IP proposal builds up a viral campaign in the IP's space rather than the LEGO space.
So the format has to change to stay meaningful & effective, and there's a couple mechanisms I could envision.
The one I mentioned above is to make voting a preorder commitment. That way people won't just vote for every ridiculous megabuild, etc., that they wouldn't actually buy, and the things people actually do want to buy, build, and play with will have a much better chance of getting through the voting.
Another would be to select the top X most popular proposals, rather than all of them that meet a static threshold. This way you can set the X based on capacity and the proposals are competing with each other to claim those spots.
Similar but more complex would be to have rounds of voting, either based on threshold or top X, so that again the proposals are competing to utilize given capacity.
Personally I prefer the preorder. Yes, it would absolutely slaughter the number of people voting and really change the dynamic. Some of that would be bad, e.g., a huge part of the current setup for LEGO is simple marketing. It's a plus for them to have tons of people coming in & voting. That would be vastly muted by this scheme. But then the voting would have real, tangible meaning.
That said, a major challenge to a preorder commitment is that the final sets are frequently very different from the proposals, and costing would be difficult & time-consuming. So maybe just a small amount escrowed that goes toward purchasing the set if created? Small amount based on a simple formula of proposed pieces and minifigs? I'm confident requiring any amount of deposit to vote would really improve the information quality of the voting.