Models with taste have a large (and constantly growing) potential market.
Written by: Alana Levin
Translated by: DeepTechFlow
In the past two years, new AI models have emerged in large numbers. These AI models can perform many types of tasks—from finding information and answering questions to providing customer support, proofreading documents, and generating content.
Many of these tasks are objective, with clearly defined optimization functions: finding the right answers, identifying the most relevant information, detecting any errors or anomalies, and so on.
However, there are also some models whose outputs are highly subjective, such as creating "excellent" artworks or developing "interesting" videos. I refer to these as "models with taste." Taste-based models are often more difficult to optimize because they are a hybrid of collective and individual decisions; there are no clear answers or outputs. Therefore, frequent feedback is particularly valuable in helping the models understand the latest cultural preferences.
Today, there are roughly two ways to cultivate the "taste" of models:
Based on user-generated content/data (such as summaries from Twitter or Reddit), which theoretically can reveal the latest trends of human attention (thus serving as representatives of taste).
Utilizing a community of human "taste makers" to actively train the models around their preferences.
The first approach has many less-than-ideal situations. The data may be isolated (for example, Reddit closing its API) or introduce biases (for example, sharing only partial data). The model may also become overly adapted to the algorithms of specific platforms, especially if its data sources are limited. This may not seem important until people start imagining a large amount of new media generated based on popular content from Twitter. This is not ideal.
The latter approach, a network provided with feedback from humans, avoids many of the risks mentioned above. There may still be biases, but only in terms of including the preferences of community members who choose to help train the model. Therefore, the key is to ensure that these community members, the "taste makers," are truly closely related to the model's cultivation of good taste.
The crypto track can help promote this consistency. Providing ownership/economic benefits to participating members in the model's output can motivate them to truly participate. Cryptocurrency also makes participation more open and accessible: anyone from anywhere in the world can contribute as long as they have a blockchain wallet and internet connection.
A notable example is the Botto project. Botto is an autonomous artist, and $BOTTO token holders have the ability to help train the model every week. The training is simple: participants vote in favor of or against various images, and Botto learns from the preferences of the members. At the end of the week, the most popular works are auctioned, and the participants who helped train Botto that week will be rewarded.
Art is just one category of models with taste. Others may include movies, television, other forms of storytelling (novels, short stories), comedy, and advertising/brand activities. Even just a few years ago, these taste-based models were impossible. These tools had poor performance, were slow, and could not reliably produce cohesive or (in the case of videos) realistic outputs. Only today has this become possible.
It is important to note that models with taste have a large (and constantly growing) potential market. The art market is worth billions of dollars. Online content consumption occupies tens of trillions of hours of attention each year. If people are already planning to spend time and money on these forms of entertainment, it seems reasonable to give them a certain share in production, which will not only create a more active user base but also a more satisfied user base. Imagine if in the Oscar for Best Picture, the main participants were the audience who helped train and develop the storyline, or a completely new award was set up for movies created by the community— that would be very cool.
I believe this is creating a new category of content rather than replacing existing creations. This is similar to how smartphones and Instagram have enabled everyone to become a photographer, and the existence of these new technologies has not eliminated the work of actual photographers; in fact, it may have made more people appreciate the work of photographers. Models with taste are the same: they have created a new form of participation by utilizing new technologies, in this case, the crypto track for consumer ownership and economic alignment, thereby expanding each of the above categories.
In the past few years, we have seen the emergence of thousands of new models. In the next few years, there may be millions (or even more) of new models, and at least some of them should strive to attract stakeholders in new ways, from greater openness and accessibility to new ownership structures for incentive measures. Models with taste are particularly suitable for this kind of innovation, but they are unlikely to be the last.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
