It has been the 4th year since Nebula AI token has been issued, there are lots of decision and development has been done, over the last couple of years.
We like to do a review about what we have done.
The beginning of everything
In 2017, ethereum mining is a hot topic, eth price increase 10x in one year, but at the same year, eth starts the research about how to switch from POW to POS, so we decide to find a way converting the GPU from mining to machine learning from 2018.
How to design it?
We decide to fork eth to build a side chain so we can solve the following issue:
- Less traffic than eth, it is an Machine learning only blockchain
- We can set the block reward according to the needs
- If ETH has some feature not good for the NBAI, we can choose not to merge it.
Things goes well, the blockchain setup quickly and the mining pool is up as well, it is http://aegispool.org it works well for 4 years and perform a good job.
In 2020 we see BSC,Polygon, Huobi chain use the same strategy, so proved to be a correct choice.
What is hard?
- The system is complicated when you are doing the data transfer and computing. If everything is rely on web3 blockchain, you have to wait 1 minutes for each action after it is onchain.
- The data protection is difficult. When doing computing you need to decrypted the data, which is extremely difficult to keep the data safe.
- Eth mining rig is not powerful enough. It is not good enough if we want to do cloud computing or deep learning, we need good CPU/Storage/GPU
What we do?
After successfully build the system with S3, machine learning container and container native service, we found that we have difficulties to make the system wildly accepted by existing industry user, the web3 technology and ecosystem is not grown up yet. The fund is almost running out, we need to find to proper way to solve the issue we mentioned above.
FIlecoin is a decentralized storage solution with extremely powerful machines. Working with them can let Nebula AI’s solution running on top of Filecoin system is a good fit.
What we have learned?
- Marketing is important.
While we are focus on development, we do not have spend enough time on marketing, which cause a low market influence, keep enough fund in marketing is something we definitely need to do in 2022.
- Fund raising is a continues work. It should be go though all the time
- Industry acceptation takes time, adaption in blockchain should be the first thing to do
- Working with other blockchain is good for the project
- Focus on one product not every product.
What to do in 2022?
Crosschain support will gain more attractions in the space, computing should be a module can be running on any chain, and increase the investment in marketing.
