I really want to touch on the subject if ASIC resistance, as it is counter intuitive.
ASIC resistance is very important to me, but this is an incredibly challenging subject to approach from a technical level. It is one of the main failures of Bitcoin, as I mentioned earlier. For some background, my “classical” training is in Silicon fab/design (specifically digital design). I started FPGA mining Bitcoin back when it was effective.
The silicon world is constantly, iteratively and rapidly changing; With the current node trajectories, Fab consolidation, possible transition to carbon, it is not intuitive what “ASIC resistance” even means in 2020, let alone 2025. While certain hashes may be challenging to compute on todays technology, that is not to day they won’t be easier with tomorrows.
This subject is going to need a lot of academic research, as well as a roadmap. I think key questions that need to be clearly defined are:
- What does an “ideal” mining ecosystem look like? (100% GPU? SoC-embedded FPGAs? Proof of space?)
- What are the probabilities that different Silicon/Fab technologies will become commonplace in 5, 10, 20 years, and what impact will those technologies have on different mining approaches
- Is there a planned schedule to update ASIC resistance as technologies evolve
Frankly, given my training in ASIC design, my gut is telling me that “true” ASIC resistance is likely impossible based on economics. If it becomes worth it to design an ASIC, people will fund it. Digital design engineers are incredibly good at what they do, and if you give them hard problems, they have a remarkable history of solving them.
Proper strategic planning, likely form a game-theory/economic standpoint is required here. The Silicon industry is on the verge of total transformation after we hit 3-5nm nodes. Planning anything past that is going to be challenging.