I don't think the replies from @PlebeiusG were "useless". I think you didn't run the numbers as he did and thus you didn't realize yet that what you are proposing ("let's run enough USB miners to "decentralize" bitcoin mining") is not realistic.

It seems like the current best USB miner can reach a hashrate of 200 GH/s [0]. An Antminer S9 can reach a hashrate of 14TH/s (which is old tech since it was released in 2017) [1]. Current global bitcoin hashrate is 356.6 EH/s. [2]

So this means the current best USB miner is as good as 14/0.2 = 70 S9s. This doesn't sound too bad, right? Wrong. (I included S9s because of this comment from you where you asked about 200,000 thumbdrives compared to S9s.)

Looking at the global hashrate, this means that the best current USB miner is only (0.2/(356.6*1000*1000)) * 100 ≈ 5.61e-10 % of the current hashrate.

This means to reach 1%, you'd need around 1/5.61e-10 ≈ 1,783,000,000 ≈ 2 billion USB miners. That's more than how many iPhones were sold since it launched in 2007 [3]. There are 8 billion people on the earth. So every fourth person on earth would need to run of these USB miners to get bitcoin mining "more decentralized" by 1%.

And that's without assuming that the hashrate will continue to go up. I don't think USB miners will be able to keep up with ASICs. At some point, thermodynamics comes into play and you just need more power than a regular USB port can deliver for increased hashrate and probably a bigger chassis for cooling. At some point, you are better off to buy an actual ASIC, I guess.

Therefore, I think your efforts are better spent at looking into Stratum v2, for example.

[0] https://linuxhint.com/best-usb-bitcoin-mining-asics/ [1] https://www.asicminervalue.com/miners/bitmain/antminer-s9-14th [2] https://mempool.space/de/mining [3] https://www.demandsage.com/iphone-user-statistics