Partnered With:

AI and Web3: Exploring innovation and challenges in cross-disciplinary fields.

Read this article in 12 Minutes

The combination of Web3 and AI may bring new breakthroughs. Taking model assetization as an example, this paper analyzes the opportunities and challenges in detail.

Original author: Lao Bai, ABCDE Investment Research Partner

Last time I wrote about the BTC ecosystem. I was supposed to follow up with an article on NFT and NFTFI, but unfortunately NFT has been really cold lately. Not only is the secondary market cold, but I haven't talked about any NFT or NFTFI related projects in the primary market for two months. On the other hand, AI projects are emerging like a geyser, so let's continue to hold off on the NFT article and focus on the trend of combining AI and Web3 for now.

1. Let's talk about AI itself

The AI industry was actually supposed to be in trouble. You know the founder of Near, Yilong, right? He used to work in AI and was the main code contributor for TensorFlow, the most popular machine learning framework. People speculate that he couldn't see much hope in AI (before large models) and switched to Web3.

Finally, the industry welcomed ChatGpt3.5 at the end of last year, and suddenly the industry became active again. This time it can be considered a qualitative change, rather than the previous hype and quantitative changes. A few months later, the AI entrepreneurship wave also spread to our Web3. On the other hand, Silicon Valley Web2 is in a frenzy, with various capital FOMO, various homogenized solutions starting price wars, and various big companies and big models competing...

However, it should be noted that AI has entered a relatively bottleneck period after more than half a year of outbreak. For example, Google's search popularity for AI has plummeted, Chatgpt's user growth has slowed significantly, and the randomness of AI Output has limited many landing scenarios... In short, we are still very, very far away from the legendary "AGI - General Artificial Intelligence".

Currently, the Silicon Valley venture capital community has several predictions regarding the next steps in AI development:

1. There is no vertical model, only large models and vertical applications (we will mention Web3+AI later).

2. Edge devices such as mobile phones may pose a barrier to data, and AI based on edge devices may also be an opportunity;

3. The length of Context may cause a qualitative change in the future (currently using vector databases as AI memory, but the context length is still insufficient).

二. Web3+AI

Translation:

II. Web3+AI

AI and Web3 are actually two completely different fields. AI requires concentrated computing power and massive data for training, which is a very centralized thing. Web3, on the other hand, is focused on decentralization, so it's not that easy to combine the two. However, the argument that AI changes productivity and blockchain changes production relations is deeply ingrained in people's minds, so there will always be people looking for that point of integration. In the past two months alone, I have discussed more than 10 AI projects.

Before discussing the new combination track, let's talk about the old AI+Web3 projects, which are mostly platform-based, represented by FET and AGIX. How to say it, my friends who specialize in AI in China told me that "these AI projects that used to be popular are now basically useless, whether it's Web2 or Web3, many of them are baggage rather than experience. The direction and future is models like OpenAI's Transformer, which saves AI with large models." You can judge for yourself.

Therefore, the general platform type is not the Web3+AI model that he favors. The 10+ projects that I discussed also do not have this aspect. Currently, the basic tracks that are seen are as follows:

1. Bot/Agent/Assistant Model Assetization

2. Computing power platform.

3. Data Platform

4. Generative AI

5. Defi Trading/Audit/Risk Control

6. ZKML

Today we will mainly discuss the monetization of Bot/Agent/Assistant, which is the most talked about and homogenized field. Simply put, these projects are mostly based on OpenAI, combined with other open source/self-developed technologies such as TTS (Text to Speech), and with specific data, FineTune to create robots that are better than ChatGPT in a certain field.

For example, you can train a beautiful English teacher who can teach you English. You can choose whether she has an American accent or a London accent, and you can also adjust her personality and chat style. Compared to ChatGPT's more mechanical and official answers, your interaction experience will be better. In the industry, there is a DAPP called HIM, which is a Web3 female-oriented game that features a virtual boyfriend. It can be considered as a representative of this type of game.

Starting from this idea, theoretically you can have many Bots/Agents serving you. For example, if you want to make boiled fish, there may be a Cooking Bot specifically Fine Tuned for this field to teach you, providing answers that are more professional than ChatGPT. If you want to travel, there is also a travel assistant Bot that provides you with various travel suggestions and plans. Or if you are a project party, you can set up a Discord customer service robot to help you answer community questions.

除了做这种「基于 GPT 的垂类应用型」Bot，还有基于此的衍生项目，比如 Bot 算「模型资产化」，有点 NFT「小图片资产化」的意味，那现在 AI 里面流行的 Prompt 是不是也可以资产化，像是 MidJourney 不同的 Prompt 可以生成不同的图片，训练 Bot 时不同的 Prompt 也会有不同的效果，所以 Promopt 自身也具备价值，也可以资产化。

There are also projects such as portal indexing and search based on these types of bots. If we have thousands of bots in the future, how do we find the most suitable one for us? Perhaps we will need a Web2 world portal similar to Hao123 or a search engine like Google to help us "locate".

In my personal opinion, there are two drawbacks and two directions for the current stage of Bot (model) assetization in the encryption industry.

Drawback 1 - Homogenization is too serious, because this is the AI+web3 track that users can easily understand, which is somewhat like a Utility attribute NFT. Therefore, the primary market is currently showing a red ocean trend, but the underlying technology is all OpenAI, so everyone actually has no technical barriers, and can only rely on design and operation.

Drawback 2 - Sometimes things like Starbucks membership card NFT on-chain, although it is a good attempt to stand out, may not be really convenient for most users without a physical or electronic membership card. This problem also exists for Web3-based bots. If I want to learn English with a robot or chat with Musk, Socrates, who needs a chat, why not just use Web2's http://Character.AI ?

Two directions - one is short to medium term, and the idea of modeling on the chain may be a train of thought. Currently, these models have a little ETH NFT small picture meaning, and most of the MetaData points to off-chain servers or IPFS, rather than pure on-chain. The model is usually tens to hundreds of megabytes in size, and it needs to be thrown on the server.

However, with the rapid decline in storage prices recently (2TB SSD for 500 RMB), as well as the advancement of storage projects such as Filecoin FVM and ETHStorage, it is believed that it should not be difficult to put hundred-megabyte level models on the chain in the next two to three years.

You may ask, what are the benefits of being on-chain? Once the model is on-chain, it can be directly called by other contracts, making it more Crypto Native. There are definitely more ways to play, with a feeling of being fully on-chain game, because all data is native to the chain. Currently, some teams are exploring this area, but it is still in the very early stages.

Another direction is medium to long-term. If you think carefully about smart contracts, the most suitable interaction is not between humans and machines, but between machines. AI now has the concept of AutoGPT, which can create a "virtual avatar" or "virtual assistant" for you. It can not only chat with you, but also help you perform tasks according to your requirements, such as booking flights, hotels, buying domain names, and building websites.

Do you want an AI assistant to conveniently operate all your bank accounts and Alipay, or to easily transfer funds to a blockchain address? The answer is obvious. In the future, will there be a bunch of AI assistants integrated with AutoGPT-like technology that automatically conduct C2C, B2C, and even B2B payments and settlements through blockchain and smart contracts in various task scenarios? At that time, the boundary between Web2 and Web3 will become very blurred.