Have Modern Blockchains Lost Touch With Bitcoin's Early Vision?

I came across this post on Twitter and felt the platform didn't provide the same opportunity and depth of discussion as Reddit. I've made some edits to the individual's post to focus on their concerns, but you can find the original post and poster via his Twitter Handle @HMD2V

________
I've been reading about the new Mysticeti update. I found some concerning points, particularly regarding its scalability as the validator node set increases.

uses a DAG structure where each block references over 66% of all blocks from the previous round. If you double the validators, you get more than 66% additional blocks to hash and include in your block for the next round.

The second concern I have is Sui's leader selection. has leader slots that are deterministically selected (fixed or round-robin) and wait their turn. While the new update changes this from a single validator to multiple ones (the exact number set for each epoch, decided every 24 hours), and if a validator fails to show up, they can more easily skip it, this determinism remains concerning.

Some blockchain networks use cryptographic sortition with VRFs to avoid deterministic selection. VRFs can be run locally to prevent an adversary from having any suspicion of who the next winner will be. There is no validator with a massive target on their back the way proposers have for the 24 hours each epoch lasts.

Then I decided to examine what it takes to be a validator: 24 CPU cores, 128 GB RAM, and $33.3 million worth of SUI.

claims TPS in the 100k-400k range, but all benchmarks in their paper involve either a "small committee" of 10 or a "larger committee" of 50 nodes. Currently, there are 106 active validators on Sui's mainnet.

This raised my concerns about 's ability to scale with an increasing number of validators. It doesn't seem like there are any intentions to allow smaller participants to join.
_______

Interesting? Are blockchains now commoditised based on functionality and capabilities for DAPPS, or is the accessibility and ability to become decentralised crucial for Public Blockchains?

Essentially, I'm questioning whether the focus should be on what blockchains can do (their features and capabilities) or on how accessible and decentralized they are.