Back when wardriving was a more popular pastime (and before the ISM bands got clogged completely up with everything), I wanted to play around more with long-range 802.11b/g signals.
I scored an old Primestar dish and LNBF. This was the easy part: I just talked to one of the guys at a small local company that dealt with things like consumer satellite, and asked nice. He was curious what I wanted to use it for (Primestar was dead by that point), so I told him, and he thought that was a fun idea so he gave me one or two that he had kicking around in the shop.
The idea was to toss the LNBF and put a biquad antenna made from wire at just the right spot where the parabola focused. It seemed easy enough, but my fabrication skills 20 years ago were lacking, and I didn't have any measurement gear beyond a multimeter and a tape measure, so the project never went anywhere.
But nowadays, with modern accessible CAD and 3D printing and sendcutsend and JLC and SDR dongles and Harbor Freight and everything else? With the RF gear I either own these days, or have access to? Yeah, I'd probably be able to make it work. I might even be able to turn a new feedhorn on my buddy's lathe to increase efficiency. FFS, there's probably already parametric models out there for OpenSCAD that just do this thing.
It seems much, much more do-able with today's resources than it was back then.
Except, also these days: If I wanted a big parabolic wifi antenna, I'd probably just buy one that was made in a factory...and use it. :)