"Cookie has another killer move: Sapien's second list is giving away a big giveaway"
Sapien has just completed a $10.5 million financing, and it can't wait to launch Cookie's airdrop. It seems that it is well prepared!
The Sapien list has just been launched, and everything is in the early stages. I plan to build a Cookie community to help everyone spread content. 📌Throw your first#Sapientweet in the comment area, and I will interact first~
To understand what @PlaySapien is doing, we have to start with AI. In the past few years, AI technology has advanced by leaps and bounds, especially the wave of large models, which has made the scale and quality of training data a bottleneck. Model training does not lack computing power, but lacks high-quality, diverse, and continuously updated data, and the collection and labeling of this data is often the most expensive, most scattered, and most labor-dependent part.
If you surf the Internet frequently, you will find that many websites have a verification, which may be some animals, asking you to identify how many cats there are, or traffic lights, asking you to point out the location of the traffic lights. Do you think these operations are the website verifying that you are a human machine? Actually, it is not. Every click you make is providing samples for AI. You are free and cheap labor🌚
Individual users who participate in it have almost never received any rewards.#PlaySapienwants to dismantle and reshape this set of logic that was originally implicit, one-way, and value-grabbing by the platform - let the data production rights return to the hands of individuals, so that the labeling behavior itself can be quantified, recorded, and incentivized, and become a sustainable and tradable on-chain labor network
The most core changes are two points:
First, lower the participation threshold. The tasks are designed to be lightly gamified, such as identifying animals in images, labeling car VIN codes, etc., and ordinary users can participate in the fragmented time of their mobile phones. It does not require any professional thresholds, does not rely on skilled labelers, and maximizes the coverage of the population
Second, direct cash. Complete tasks to get USDC rewards, and the chain will be instantly credited. No points, no platform points, or delayed redemption of points malls. Users know what they are labeling and how much the data is worth, and the platform does not have to bear the lengthy credit endorsement and points economy maintenance costs
This combination of "light threshold + cash incentive" is naturally suitable for small training data markets with extremely scattered data needs and fragmented task types.It is difficult for large model companies to organize manpower to do such fine-grained annotation, but Sapien has formed a data supply system with high cost performance by distributing manpower to fragmented users all over the world, diluting production capacity.
Logically, @PlaySapien is actually doing two things: turning data supply into a traffic business; turning data annotation into a labor business
This is not a fantasy. Variant, Animoca, YGG, Folius, and several other funds are willing to bet in the early stage. In essence, they value this market with structural supply and demand imbalance: AI companies are in urgent need of long-tail data, and long-tail data is extremely difficult to solve with traditional employment relationships. Disassembling the supply-side structure is what Web3 can do
In the next article, we will introduce in detail how to use Sapien for free 🪂
#sapien#cookie#spark#cookiedotfun #snaps
