2) Bitcoin mining is not ASIC resistant
There is no such thing as ASIC resistant. Some algorithms try to do this by being memory intensive but that will fail because GPU farms can still provide sizeable hashing power.
You can make ASICs harder to manufacture, i.e scrypt ones are more difficult than SHA256 but still... These are not the only flaws, purists wanting to keep 1MB blocks forever or scale to unlimited size or have a band-aid fix of a one time increase are in my opinion deluded, but that's my own opinion. One of my indian clients is struggling to give me daily payments of my articles due to TX fee size, this is becoming a problem for Bitcoin. No one has considered countries which do not have a lot of money, a 20p TX fee is too much for example. Now we are switching to every few days which I am fine with, if this carries on we may move to Litecoin for a more clear blockchain.
Segwit is the best option for those who wish to run full network nodes even though it takes more power, imagine unlimited having 64MB blocks? This wont even come close to visa or mastercards transaction volume but it would make blocks insanely hard to propagate and store.
ASIC resistance is a myth, you can make a dedicated chip for anything if theres enough desire and someone with the cash decides to do it, design it and have a foundry make it for them. You can make it harder and less cost effective (i.e scrypt and cryptonight) A CPU is simply a chip designed to do Multiple tasks, a GPU is like a CPU which can do many SINGLE tasks of one's choice, one at a time, an ASIC is a chip designed for a SINGLE task, no other task can ever be done or adapted for it, but it can do it better than the other two. For example if Bitcoin crashed and no sha256 coin took its place, all the ASICs become a doorstop. This can be designed for any particular task, some ASICs can be used for video decoding if desired for a single algorithm. Cheap MP3 players would have an ASIC to decode the MP3s. Someone designed an ASIC to crack 64 bit DES keys.
But slowing down the research is he point not to stop it. If you make ASIC chips very expensive to research and manufacture then the centralization will halt.