The function of the bureaucracy in this context is a filter for effort. It naturally filters out people who don't care enough to battle through the bullshit, which is arguably quite an effective way to distribute limited resources to the public.
In general, our economy relies a lot on these natural filters for effort. A company with a nice website doesn't necessarily mean the company is good, except historically it kind of does, because it's a proxy for effort (and budget), both of which correlate roughly with reputability and a good product.
The same applies to well-written blog post: typically, a very good writer generally is also someone with something interesting to say, even though in theory the two things don't have to be connected.
Similarly, the entire moat of many companies is that switching to a competitor is a lot of effort.
AI is effectively sending the effort required to do any of these things to zero, which means all our existing implicit filters are going to stop working.
Exactly this. The overall productivity of not doing something should never be underestimated.
>which is arguably quite an effective way to distribute limited resources to the public.
I would argue in a lot of cases it isn't: especially in cases where the resources are intended to go to people who are struggling in some way. People who need the most help are often the least equipped to deal with the bureaucracy, while someone who is exploiting the system can happily jump through as many hoops as it puts up (especially when they're arbitrary rules that are encouraged to be applied without critical thinking and have very little correlation with the thing that is supposed to be being assessed). It might limit the number of people using it but it selects opposite to the desired result.