The Vitus frames. When you don't weld, you can use much lighter weight (thinner wall) tubing which will be more flexible. The French stayed with standard tube diameters and short lugs which added to the flex.
IMO the 1930s Caminargent is one of the most beautiful bikes of all time. The art deco lugs, brass fork ends, hex shaped tubes. If I owned one, I'd polish it shiny and hang it like art. No way I would dare ride one.
Gary Klien sold frames which buyers had to build up (or have their local bike shop do that), Cannondale sold bikes. A few thousand Klein vs millions of Cannondales. One infamous topic was the US Patent office granted Klein a patent for the standard double diamond frame. Klein specified a range of diameters for the main tubes and stays. So Cannondale went huge to get around the patents. The flat oval stays, the massively large down tubes (beer can) were the result. The double diamond frame has been the dominate design since 1900 or so for good reason. There was a a lot of discussion on the usenet at the time about whether the patent should have been granted. I've since learned the patent office issues crazy number of patents, like tens of thousands just for golf club grips. To be clear, the US Patent office doesn't always know what they are doing. Klein was making frames long before Cannondale but in small numbers. There was a lot of legal wrangling and lawsuits. Trek bought Klein and later shut it down.
Merida of Taiwan was an early maker of bonded frames. I talked to an ex-Trek guy who told me about the prototype they ordered from Merida, then decided they could do that themselves. I have 2 Merida bonded bikes from the mid 1980s. I think Merida made the early Specialized carbon frame road bikes. The Trek 1000 series aluminum road bikes and the 2000 series 3 carbon tube main frame models were bonded. Raleigh USA Technium frames were bonded and most of those were a combo of aluminum main tubes and steel stays. You can glue steel to aluminum but you can't weld those metals together. All these frames used cast head tubes, seat lugs and bb shells that did not use mitered tube ends, the tubes were square cut and fit over internal nubs on the castings. Glued, then baked to set the glue. Once the production was going it only took a couple minutes to assemble a frame. It didn't take expert frame builders to make a lot in a hurry. Trek had a hydraulic jig that would squeeze everything in to place prior to baking the thermal set epoxy.
All welded aluminum frames require heat treating or annealing to remove the stresses caused by welding. Cannondale used sturdy steel jigs that held the head tube, bb, seat lug and rear end in alignment when annealing. But the main tubes would warp like crazy. We measured some that curved as much as 1/2". The shop I worked at had display racks where the bikes were 'backed in" to slots so anyone would see the curved down tubes. When you see them from the side you don't notice but when you sight down the tube, it was very obvious.
In the 1970s bike boom, there were a lot of diy products. Frostline kits for clothing, tents, bags, etc. The Hyper Cycle recumbent kit. Bikecology and Bike Nashbar catalogs sold components so you could assemble your won bike. Proteus sold a frame build kit that surely got a few hundred custom frame builders their start. There were a couple of mail order catalogs just for frame builders. Aluminum bikes had been around since the 1930s. Welding aluminum frames got started in the 1970s by an unknown number of diyers as school projects or just on their own. There is a massive gap between making one and going into production that most makers/inventors never overcome.