The decentralization is the key element to avoid the acquisition of dominance of a certain group in the initial stage; the issue is that you cannot avoid it, just put a delay on it.

It can be a long term feature, like the ICE-Age of Ethereum, or it could lead to an ASIC viability like any other crypto, but just remember that the "centralization"concept means also volume on a daily basis, and volume means market capitalization.

You cannot acquire success without having a third-party infrastructure backbone, and it's simply called market approval.

In my experience the ASIC technology it's unfortunately a wrong step, because it leads to an excess of centralization  in the mining process (easy build-up of farms) and consequently a bigger issue related to 51% attacks (double spending).

A path with a CPU compatibility in the testnet, with a near-0 distribution with an early stage it would be fair, associated to a random distribution. Then, following the client's releases and patches, it could be scheduled a VGA startup phase, which would imply a near-immediate market approval, efficient step-up and distribution phase. Have a look at ZEC emission rate, and what brought to ZEC/BTC: it started with hype, and due the latter it vanished, like a leaf of ice in the desert.

Don't give so much importance to the algo, the emission plan it's one of the keys to the success of a coin.