#fallacies of cdns

i feel like cdns are both good and bad.

on one hand, cdns provide image files, media files, and anything in between at a fraction of the server load and latency than coming directly from the origin server.

on the other hand, these are giant treasure troves of data being touched by ai scrapers and bots, and while thats not really bad, many times, important documents are stored in cdns, especially in multi-region companies.

you CAN use something like sharepoint, or dropbox, or gcp buckets, but when you have data like shareholder info, it is important to have fine grained ACLs such that only the people you want to can get it, and the people you don't, well... don't

while yes, things like cf (cloudflare) zero trust, etc, do exist, they are often clunky, and when they get the file, what then? they have full range and ability to share it with anybody they want.

ai scrapers in browsers (such as Dia or Atlas or Perplexity) also can get access, and since we don't know if the data is being used to train, then oh no, you're cooked.

\therefore, i built (ahem, claude opus 4.7 built the backend, and cursor composer 2 built the ui, please pardon it) a fine grained acl configuration, and while you CAN screenshot it, at that point, you're cooked anyways.

after all, its important that your files only end up in the hands of people YOU want. (ahem, ahem)