- Since the biggest cost (and barrier to entry) is the die. We will instruct the fab to release the die to "public domain". That means, any company can place an order from the fab without having to invest in any NREs. This will lead to a healthier competition.
By instructing the fab I understand that this will be your obligation to do so, being specified in a contract of some sort.
You see, trust is very hard to earn and stuff like "we will do x and y" means nothing if it isn't clearly stated in some form of legal binding contract.
I am interested in this project, I think it's fair and good to the community to have access to cheap hardware in order to have a more diverse ecosystem. Right now the markups are huge because there is a huge demand for these products, they are scarce and everyone, from Avalon to Bitfury, CoinCraft, CoinTerra, CoinWhatever has to make the choice of selling this chip at a higher price or mine for themselves (personally I would mine for myself - too much hassle) yet neither of them, so eager to "help the community" has come up front and said "we will make our chip public domain".
There are kickstarter projects that raised more than 2 million yet we buy custom made hardware because we are looking for the same thing: ROI in 4-5 weeks and a fairly fat profit 6-9 months down the road. Greed is natural. Fortunately, technology limits that.
Also you will only get one shot at this so you know... you have to make it count.
I will be watching this thread.
We are open to suggestions about what steps to take to make our operation more transparent. One option is to put all the "we will do x and y" in the company bylaws and making the Executives/Board personally liable if they fail to honor the commitment.
Interesting idea! I'm very fond of the transparency around your NRE expenses and community elected treasurer-group suggestion. A few questions though!
Assuming the 28nm AND 14nm projects prove successful (probably a ways to go on the 14nm) what about smaller chip-architecture?
Any plans for election processes of treasury agents? Length of terms of service?
What kind of background does your company have in ASIC design? You obviously know what you're talking about, I'm just interested as a miner in your "portfolio".
Pubic domain for the die on the ASIC chips sounds wonderful; do you plan on open-source firmware to run the chips too?
Best of luck and I will certainly be following!
14nm is still a dream. We will be talking to some companies this week to find out if it is even possible anytime soon. So far we know that we won't probably see 14nm ASICs until 2015.
We will let the community nominate and then vote for the treasurers.
We do not have any particular experience with ASIC design. However, majority of the work will be contracted to companies specializing in this. SHA-256 is not a very complicated algorithm, and designing an ASIC for this is relatively easy. We will also hire experienced people to help us make the right decisions.
Everything will be open-sourced, since it is funded by the community.