A faster way to improve ROI
After power, there's the question of unlocking rewards faster.
On Ethereum, it is the "uncle rewards" system.
With WaltonChain, it's "grandmaster" nodes.
With distributed ai/render:
It's about having access to the right programs, offering enough compute power/bandwidth, and creating a customer demand pipeline for increasing profit.
A better place to learn.
The benefit of a genuine community is mentorship.
- Accountability (having a job to do)
- Friendship (meeting new but like-minded people)
- Being Professionally "Nourished" (gaining new marketable skills)
Our near-term vision is to host various training classes for community members who wish to study computer science, graphics, ai, & various programming/scripting languages.
Running multi-GPU rigs is not just a hobby; it's your livelihood...
We're working together to protect it.
The future of GPU "mining".
Why the airquotes?
Because simply put, PoW mining cannot continue to scale at the same pace as crypto/blockchain usage, if it is to acheive mass adoption.
And frankly GPUs are not the best hardware for PoW consensus.
Don't misunderstand: we love blockchain and crypto.
But for all the distributed problems that GPUs are uniquely good at solving, we want to prepare & work together.
WHY ARE WE DOING THIS:
Graphics Processors Were Not Designed For Blockchain
What Were They Designed For?
oil and gas
ai model training
virtual material testing
So Why Haven't Former Miners Been Able to Sell Their GPU Power?
TL;DR "Because tech reasons".
The full story:
A few of the industries listed above will actually pay very good money for the use of advanced graphics cards like the ones miners employ for blockchain.
There are several major hurdles:
1. The buyers of CGI rendering have major security concerns.
What works for blockchain (which, as you're aware, should require little or no trust) won't work for them.
Computer generated images (CGI) for films, television, and web series all contain time-sensitive intellectual property.
A data center rendering the scenes of a Hollywood film featuring popular comic book characters cannot have images "leaked" before the movie premieres.
A long running TV show cannot have the season finale spoiled by having the end scene posted to YouTube in a security breach.
While watching those spoilers might seem "cool" to a 12 year old who watches at home, they are not much fun for the major entertainment companies who employ 10's of thousands of people internationally to create these properties.
2. Both ai & computer graphics rendering require far more host bandwidth than mining rigs provide.
Most proof of work mining algorithms have been discovered to be more constrained by memory speed than memory bandwidth.
This has allowed mining motherboards to feature only a single lane of PCI Express running at only first generation speeds (PCIe 1.0 at x1 lanes) and still maintain profitability.
For miners, this is a tremendous value!
However, for ai and render both high bandwidth and high speed count, and have been factored into the hardware design by AMD and Nvidia.
Motherboards used for mining can never acheive competitive performance for ai and render if they use only one lane at 1st gen speeds.
3. Most distributed render and ai clients require specific GPU models and minimums which exceed the average at-home miner's inventory.
This point has been a frustration to many at-home miners (and even mid-scale miners).
While a single ai researcher may only request a single GPU for model training, an ai school may wish to purchase in bulk.
And for graphics rendering even a "small" TV visual effects or animation studio will often need many GPUs- upwards of hundreds, or even thousands- all in high bandwidth host devices, with impeccable network security.
While distributed ai and rendering "brokerage programs" are growing in popularity- they do not fit the upper scale of today's market demand in a perfect 1:1 ratio.
There are many "big" clients who, if given the choice (and price being equal) would prefer to send their jobs to a single data center with more GPUs than they need, rather than splitting their jobs to multiple separate locations, each with fewer GPUs.
Why not use multiple separate providers on each job?
Each new "ultra-micro data center" brings uncertain turnaround times due, and potential security threats (see also; points 1 and 2).
Enter: the IDC Co-Op
What Is The MYB Co-Op?
Who Is MYB Co-Op Right For?
Like so many in the community, we're holding out hope that GPU mining will return to massive profitability in the near future.
After all, GPUs still strike the best balance between flexibility & performance for mining.
Sure, FPGA mining has returned, and shows some promise- but not at the same scale that GPUs have: There are tens of millions of GPUs currently in the wild, hooked into operating systems, with drivers installed, and ready to mine.
Use the form below to sign up as a future community leader for the co-op.
Join The Co-Op
A cheaper way to host your GPUs
Let's face it, the bottom line consideration when creating a mining farm is the cost of power.
The less expensive the power, the more profit you earn.
Our collective buying power gives the MYB Co-Op better rates than any mid-scale mine can achieve operating solo.