Here a summarized version of the AMA held by CEO Owen Tao (twitter: https://twitter.com/Owen_Matrix) in the Matrix main telegram channel (https://t.me/matrixainetwork). People got the chance to ask questions directly.
Im curious to know how the Matrix AI network will handle parallell AI work. How many parallel processes will the network be able to handle? Lets say we get 100 customers wanting to use Matrix AI 24/7 for different realtime services, will the network be able to handle that (assuming we got enough AI miners)? Please elaborate around this topic and explain your visions of how this will work. Thanks!
It’s possible to do that. All transactions will be set to a data centre. The centre will dispatch the tasks to different miners instead of handling the task directly. Currently, we can handle 1000 task simultaneously.
how many apocalypse miners do you pan to have online by the end of the year?
Our plan is to have roughly up to 2000 Apocalypse miner by the year end
When do you believe clients can start using and renting power from Matrix computing power platform platform? Do you have a time frame goal in mind?
Right now we have a number of contracts. We expect to start compute renting service by the second half of 2020.
- Is Matrix generating revenue already and if yes, how?
- Who are the customers and how much revenue is generated on average per month?
- How much are your operational costs per month in order to continue with the further development of Matrix Chain, looking for new clients etc.?
- How much funds does the Matrix team have left from the funding rounds in the past? Does the team plan to raise more VC/investor funds in the future? If yes, when?
- What is the current status of the OBOR project? Is this partnership a potential revenue stream for the project or is it more to be seen like a playground where OBOR can experiment with blockchain tech?
- Is there an updated roadmap for 2021 – 2022? What are the next milestones you are looking to achieve in the coming years?
1. The Matrix Foundation itself is nonprofitable. Its purpose is to create a platform to maximise the community’s ability to make a profit. Our focus is to build a ecosystem, rather than to make a profit.
2. This question would be similar to the previous one.
3. The operational cost for all of this is several hundred thousand USD per month.
4. Details as to this will be revealed when the Matrix Foundation publishs its financial report. As for now, we don’t have plan a new round of funding.
5. We’re planning to build a AI mining farm in Xinjiang. Once complete, this mining farm will provide support for the OBOR project. All profits from this will be for community members who have invested in their mining machines, rather than for us.
6. The roadmap will include our compute distribution platform, our data centre as mention in Whitepaper 2.0 and the creation of more mature algorithms on our mainnet.
Will gpu and apoc be released end of this month, or will it take a little bit longer?
The Apocalypse machines are en route to Xinjiang. If the logistics go right, their installation will be complete by the end of this month.
We still recommend you to use Apocalypse rather than GPU, as GPU is no match for the Apocalypse.
Can you share any information about “MDN” (http://mdncenter.com/index_en.html)?
MDN is a project that’s seeking a partnership with us on the data level. Currently we’re still in the process of investigating their program.
Would you be willing to share the demo of the AI applications platform you spoke about in last AMA?
I’ve got it, there are 2 parts :first is an introduction [but it is still in Chinese, we are gping to translate soon] second is the app itself, but i just tested, due to the break-down of the register sms service, i fail regitering. We are going to solve it. Before the release, we will post the introduction first.
I will have @Joker follow this up
What AI framework is Matrix build upon or supporting? Tensorflow, Torch, CNKT, MX? How much different will it be for a user to use Matrix AI compared to any if the above?
Also, can we at some point get some youtube videos with tutorials that shows some basic examples of how to use Matrix AI?
There are two types. First is basic compute output. This is packaged Tensor ports and programs. Users can directly develop their programs using this. The second is AI developed for specific purposes. For example, computing tasks for companies, like weather forecast and road condition predication. The first is like a finished platform. The second one is more for reference. Based upon this, people can develop their things directly, or they can enlist our help in doing so.
What three partnerships did you talk about on Twitter?
2. Tsinghua Unigroup
3. Theme Tech
When will the AI market place launch (AI app platform) ?
The second half of the year
Will GPU and APOC launch at the same time? Will the fastest GPUs be able to compete with APOC? Are non-Apoc GPUs not part of the plan so everyone can contribute?
We’ll consider opening up the network to GPU miners after the Compute distribution Platform goes online.
What would you like too see regarding marketing after the first clients begin to use matrix around the end of 2020?
Our focus then would be to find people or entities who have demand of our compute. Universities can be a good target.
Owen, how exactly will onboarding clients to Matrix AI services affect the price of MAN? dont give me the polished version, I want to know what gateways are being used and how much/little MAN is required for services, as well as how this translate against the enormous miner inflation in MAN that is happening as we speak
1. When using the services, people need to pay MAN. This will lead to a increase in the demand for MAN, which will have a positive influence on the price on MAN.
2. There are two part to the service fee. First is payment for AI scientists. Those who design the services will dictate the charge. The second part will be the cost of compute. That will depend on the supply of and demand for compute. The cost won’t be fixed.
Owen, what are you guys doing to increase the trading volume and get more listings? If you already are spending 100s of thousands per month, what is the limiting factor to why we aren’t listed on more/better exchanges?
Owen, what happened to Cybernaut, are they still invested?
We can increase trading volume by creating more demand. The way we create demand is by building our compute distribution platform.
We often get invitations from lots of exchanges, but everyone asks listing fee and promoting fee, it is expensive, further more, if we pay in man coin it will lead to inflation. .
Owen, why did a gigantic amount of MAN get moved from team wallets and split up into multiple other wallets?
Owen, has team sold any amount of MAN yet? if so, how much and why?
This is out of security concerns. We are transferring our token to cold wallets. We haven’t begun selling MAN on exchanges.
1: how exactly will onboarding clients to Matrix AI services affect the price of MAN? dont give me the polished version, I want to know what gateways are being used and how much/little MAN is required for services, as well as how this translate against the enormous miner inflation in MAN that is happening as we speak
2: what are you guys doing to increase the trading volume and get more listings? If you already are spending 100s of thousands per month, what is the limiting factor to why we aren’t listed on more/better exchanges?
3: what happened to Cybernaut, are they still invested?
4: why did a gigantic amount of MAN get moved from team wallets and split up into multiple other wallets?
5: has team sold any amount of MAN yet? if so, how much and why?
1. All of your questions except number 3 have been answered. Cybernaut was one of our earliest investors. They’re no longer invested, but they may be our partner in the future.
Hi Owen, what is the main source that allows the team to spend hundred of thousands US dollar per month? And how long this source is able to sustain till the team complete all tech development?
The budget is mostly spent on salary, rent and firmware. What we have got won’t last us till 2.0. We will get another funding or find new partners.
so the compute distribution platform seems to be the cornerstone of Matrix ecosystem. When are you planning to release the platform to the public?
We are testing it already. It will go online for business use in the second half of 2020.
The launch of GPU and compute system is one of the main goals of this season. What are the next main things to be worked on after compute ecosystem is done?
Our next goal is a data platform. Please refer to Whitepaper 2.0 for details.
How will the service remain affordable and how will you keep your competitive advantage if the token value increases?
The price of the services will be dynamic. It will depend on supply and demand. When the value of MAN increases, less MAN will be needed for the same services.