If you been around the crypto community for at most a week now then the phrase "Bitcoin Gold" should ring a bell. The August 1st Bcash hard fork has taught us a lot especially the more technical stuff like replay protection and wipe out protection. Bitcoin Gold (BTG) claims to address these fears as well as provide the community with some scaling solutions.
At the very core of Bitcoin Gold is ASIC resistance. Instead of reliance on processing power, this Proof of Work algorithm called Equihash depends on memory. This is the same algorithm used for mining Zcash. Yes, the proponents of this fork believe this would be a solution to mining centralisation. By making the protocol incompatible with the current status quo, miners would hopefully be forced to return to General purpose processors such as CPUs and GPUs. This is an excellent idea.
The big question now is; whats the guarantee that even if the Bitcoin community overwhelmingly accepts the fork, we wouldn't begin to see Bitcoin Gold ASICs coming into mining? It toke Bitcoin about three years before we started to see deployment of ASICs in Bitcoin mining. Wouldn't we in future begin to see equihash ASIC miners?
Secondly, if people are desperate enough, miners could still pill up large number of GPUs on top of large Memory and we'll be back to centralisation.
There are certainly trade offs for every technology. One side would have to suffer at the expense of the other.
Bitcoin is trading security for decentralisation. All this trillions of harsh power expended to mine is a security measure for protecting the blockchain. The greater the harsh power, the more safe the blockchain is from a 51% attack which is the likeliest we could face.
The second most significant upgrade they are adding is a Difficulty adjustment after every block unlike the adjustment done every fortnight on Bitcoin. This also is an wonderful idea as it would checkmate the computational power of the network and make adjustments immediately. The down side here is that this consistent adjustments might just slow down the network as they'll be having regular adjustment intervals between blocks and this might not scale if they intend going mainstream.
The team lacks well established developers. For now, they claim to have implemented a two way replay protection.
The upcoming fork is expected at block 491407 (October 23, 6am UTC) however, trading with the tokens would commence on November 1st.
Here is the project roadmap.