I am happy to discuss the POW change and the motivations for it. What I don't want to get into is a long discussion of "you shouldn't change the POW", because that is the direction for this project.
Fair enough, which is why we are starting a new project in parallel that will also fork but not change the PoW algorithm.
However I don't think it is practical. If the branch is much smaller than the main chain (which it will be initially) then doing so opens the branch up to attacks that are easy for a main chain miner to start. For example if the minority branch has 1000th the value of the main chain, then any 0.1% miner could successfully 51% attack the branch.
I think the security of any chain will simply always reflect the value of the block reward. Therefore a minority branch that has 1000th the value of the main chain. Will be just as expensive to attack whether you are using GPU/CPU or ASIC's. It does not actually matter, security scales with value. Which also means that the reward for attacking such a blockchain tends to be in line with how much it costs to attack it, in most cases making fifty one percent attacks irrational. Since they are also ineffectual, these type of attacks can not be sustained, in most cases the blockchain survives as long as the community of people behind it believe in it enough. Therefore it does not matter whether ASIC's or GPU/CPU's are used the actual cost should be the same, not including some of the more intrinsically secure aspects of ASIC's when compared to consumer hardware.
Furthermore a genesis fork that does not change the PoW does not represent as much as a threat to Bitcoin miners compared to what you are proposing. Since the can easily point their already existing hardware to the new genesis fork and profit as it gains value, possible even bringing in more value overall for SHA256 mining. I think most miners will be very agnostic about the chains they mine just like they are in the altcoin world because they will just chase the best profitability they can find, multipools can also help to bring about this equilibrium in profitability.
Yes there are some alt coins using SHA256, but none of them are real threats to Bitcoin. If one did become a threat it would probably be attacked. So far out of all of the alt coins that have real traction (i.e. the top 5 alts) none use SHA256 mining.
I would consider many of the top five coins to be real threats to Bitcoin. Ethereum's recent rise might even be able to be attributed to the failure to increase the blocksize in Bitcoin. Another example might the cryptonight family of coins, they are all in competition with each other yet they do not attack each other. Miners are self interested, instead of wasting money on attacking another coin they can just join a multipool instead and profit.
This project directly attacks the main branch, so it is likely there would be more interest in attacking it. Changing the POW is one level of protection.
Changing the PoW does not add any protection since the level of the security is based on the value of the block reward, or in other words the value of the coin. I would argue that ASIC's are actually inherently more secure because it is less liquid, more dependent on the success of that specific algorithm compared to consumer hardware, placing stronger incentives on those miners.
While also increasing the security of the network more by increasing the hashing power significantly compared to using consumer grade hardware which would already be in the possession of any would be attackers, there also lies an extra cost, therefore greater security.
2) Changing the POW is meant to create a risk to the miner community. If a new-POW branch starts to gain traction, hopefully that will motivate more to increase the block size, rather than to keep blocks artificially small and risk a different branch taking over.
I do actually agree with you on this point, and I do respect your efforts.
3) I believe the optimizations that were possible with SHA256 make for a very unfair and unbalanced mining community, and has been a main factor in the centralization we have seen.
I do not see how the algorithm has anything to do with what we are seeing now. I think the early transitional stages when we go from consumer grade hardware to ASIC's is the most dangerous because of the limited access and competition for the development of these ASIC's chips. It is the centralization of manufacturing that is one of the greatest pressures of mining centralization, along with economies of scale. Making ASIC's harder to develop would actually increase this centralization pressure.
The situation now is actually improving, since there are more competing companies that are or will sell to the larger public, and I even think that the SHA256 ASIC's themselves will be further commodified into different types of consumer hardware further increasing decentralization. This process would take much longer, and be greatly hindered if the PoW algorithm is changed to a highly ASIC resistant algorithm. Increasing the centralization pressures during this critical transition period in the cryptocurrencies evolution. If we also see this an inevitable evolution towards ASIC's then by changing the PoW we are setting that chain back significantly in terms of this evolution.
We could even imagine a future where there are multiple chain forks of Bitcoin, each using the same hashing algorithm, however all combined contributing to the decentralization of manufacturing, helping to ensure a more healthy ecosystem for all.
I believe we can do better than SHA256 and other alt coin attempts, to create a mining algorithm that after full optimization still has more balanced and fair economics. I am basing this view on past experience in porting many software algorithms to FPGA and ASIC implementations (which is a very specialized skill set) and having an intuitive understanding of what types of algorithms are difficult and/or not worthwhile to fully convert to power hungry ASICs.
I do not see how changing the algorithm makes the economics more fair. It does not matter, at the end of the day all that matters is equipment cost and electricity cost. Does making the equipment more expensive the economics more fair? I do not think that is the case, I think cheaper ASIC development cost would lead to greater decentralization and commodification as I mentioned before.
The POW I am creating will after fully optimized have characteristics similar to the FPGA era of Bitcoin mining, during this era capital costs were much higher than electricity costs and low cost electricity was not as much as an advantage. This POW will also likely not fully optimize to an FPGA, but to a low power CPU core connected to external low power DRAM. This will produce a better environment IMHO.
I am not convinced that this was the case because of the hardware that was used. I think you might be mistaken in the economics that brought about that situation. The only time that cryptocurrencies become very profitable to mine, meaning electricity cost are not as important, allowing for continues expansion, is when the infrastructure is still catching up to the rising value of the cryptocurrency. It does not matter what type of equipment is being used, whether it is ASIC's, FPGA or GPU/CPU it will always "centralize" towards cheap electricity. Even in the altcoin world GPU/CPU mining had become unprofitable for most home miners exactly because much of the hashpower had moved into large GPU farms that have access to cheap electricity. This only changed with the introduction of Ethereum, now it is profitable again for home users but the infrastructure is already catching up again. As you can see from this example, this "centralization" towards the lowest electricity cost is inevitable just like the development of ASIC's.
Most people probably doubt this and that is fine, all I can say is I will put this out there and we'll see what happens.
I feel the same way, and that is very much in the spirit of open source, we will also release a genesis fork that we think is a better option for the market compared to changing the PoW.