The foundation of ChatGPT's development is model training, which requires a large amount of training data and meticulous model design to ensure that ChatGPT can understand and answer users' questions during interactions. Model training needs to be constantly adjusted and improved to gradually improve ChatGPT's accuracy and response speed.
We consider early ChatGPT as a "child", and its creator OpenAI as a "parent" needs to carefully cultivate this "child" and provide "good education" - accurate annotated data, so that it can become "smarter".
If OpenAI as a "parent" wants the child to become an expert in a certain field, then it needs to hire professional teachers for "teaching", and this process requires providing targeted data annotation for personnel in the field. For example, if I want ChatGPT to learn about biology, we need biology teachers to label the relevant knowledge in biology in order to make ChatGPT better understand biology.
The question now returns to the beginning. Who will do the training and why is training necessary?
A Label-to-Earn platform to improve AI using the global workforce and provide human jobs in the post-gpt era. AI-related industries could launch targeted annotation tasks through a task publishing platform on PublicAI, cultivate smarter AI, and provide generous rewards for the "teachers" who help with the annotations.
This system is a win-win situation for AI companies with annotation needs and users participating in the annotation.
How to become a teacher and participate in Mark-to-Earn:
- After the publisher posts a reward task
- Markers can accept the reward task on the PublicAI platform
- After completing the annotation, submit it to Validator and Fisher for review
- After approval, the system will automatically distribute rewards.
Four decentralized roles form a mutual restriction. The Nash Equilibrium is achieved when the raw data is well labelled.
The solution is PublicAI
If you are an employee of a traditional labeling platform, you not only have to accept low labour wages but also may encounter difficulties with currency conversion (e.g. as an Amazon Mechanical Turk employee, your average hourly wage is between $1.20 and $5, and most people are paid via Amazon gift cards instead of cash. If you want cash, you likely couldn't as only US bank cards are supported）.
Anyone can perform data labeling on PublicAI, and each marker will be paid based on the amount offered for a single task, with users provided multiple forms of payment to further ensure their earnings.
We believe that ultimately, this revolution will be decentralized. While it may not happen immediately, we want to tell you that PublicAI will be the Amazon Mechanical Turk killer.
It is time to establish a decentralized labor market and provide a solution that everyone can participate in.
About PublicAI:
PublicAI is a decentralized marketplace focusing on AI data annotations, leverages blockchain technology to deliver a trustless and cross-border labor market being strategically incentivized by crypto economics and having instant cross-nation payment settlements. The vision is to realize the model 'PublicPilots' whose commercialization incomes will be shared back to the DAO. The team includes national academician, IEEE Fellows, professors/PhDs from top universities including Stanford, HKU, CUHK, and previous senior employees of Goldman Sachs and JP Morgan. The core team has rich expertise in the AI and blockchain fields.
Twitter: https://twitter.com/PublicAI_
Discord: https://discord.gg/avf6THqCQw
Telegram：https://t.me/+IC37WmTTpdYyM2Fl
MVP：https://www.youtube.com/watch?v=tiYd9ncb_7Y&t=14s
All Comments