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).
2. Web3+AI
AI and Web3 are actually two completely different fields. AI requires centralized computing power and massive data for training. It’s very centralized, while Web3 is geared towards decentralization. So it’s not that easy to combine the two. However, the argument that AI changes productivity and blockchain changes production relations is too deeply ingrained in people’s minds, so there will always be people rushing to find that point of integration. I’ve talked with over 10 AI projects in the past couple of months.
Before talking about the new integration track, let’s talk about the old AI + Web3 projects, which are basically platform-type projects represented by FET and AGIX. Well, my friends who specialize in AI in China told me: “Those who used to do AI are basically useless now, whether it’s Web2 or Web3. Many of them are burdens rather than experience. The direction and future are like OpenAI’s large models based on Transformers. Large models save AI.” You can decide for yourself.
So universal platform-type projects are not the favored model for Web3+AI. None of the 10+ projects I talked to were in this area, and currently, I see the following several tracks:
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, I’ll mainly discuss track 1, the assetization of Bot/Agent/Assistant. This is also the track I’ve talked about the most and has the most serious homogenization.
In short, these projects are mostly based on OpenAI, combined with other open source/self-developed technical means, such as TTS (Text to Speech), plus specific data, to FineTune some robots that are better than “ChatGPT” in a certain field.
For example, you can train a beautiful English teacher to teach you English. You can choose whether she has an American accent or a British accent, and her personality and chat style can also be adjusted. Compared to the mechanical and official answers of ChatGPT, your interactive experience will be better.
There’s a DAPP in the circle called HIM, a Web3 female-oriented game that has a virtual boyfriend. It can be considered as a representative of this type.
Starting from this idea, theoretically you can have many Bots/Agents to serve you. For example, if you want to make boiled fish, there may be a Cooking Bot specially designed for this field to teach you. The answers given are more professional than ChatGPT. If you want to go out and travel, there will also be a travel assistant Bot to provide you with various travel suggestions and plans, or if you are a project party, you can create a Discord customer service robot to help you answer community questions.
In addition to making this kind of “GPT-based vertical application” Bot, there are also derivative projects based on this, such as Bot calculating “model assetization”, which has a bit of NFT “small picture assetization” meaning. So is the Prompt in AI now capitalizable, like MidJourney’s different Prompts can generate different pictures, and different Prompts will also have different effects when training Bot, so Prompt itself also has value and can be assetized.
There are also projects such as portal indexing and search based on this type of Bot. When we have thousands of Bots one day, how to find the most suitable Bot for you? At that time, a Web2 world similar to Hao123 or a search engine like Google may be needed to help you “locate”.
In my personal opinion, there are two drawbacks + two directions to Bot (model) assetization at this stage.
Drawback 1 – Homogenization is too serious. Because this is the AI+web3 track that users can most easily understand, it is a bit like an NFT with a Utility attribute. Therefore, the primary market is beginning to show a red sea trend, rolling up, but the underlying is all OpenAI, so everyone actually has no technical barriers, and can only compete in design and operation.
Drawback 2 – Sometimes it’s like the Starbucks membership card NFT on the chain. Although it’s a good out-of-circle attempt, for most users, there may really be no physical or electronic membership card that is convenient. Web3-based Bots also have this problem. If you want to learn English with robots or chat with Musk and Socrates, isn’t it better to use http://Character.AI of Web2 directly?
Two directions – one is the near + medium term, model on-chain may be a way of thinking. Currently, these models have a bit of the meaning of ETH NFT small pictures, 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.
With the rapid decline in storage prices (2TB SSD for 500 RMB), as well as the development of storage-related projects like Filecoin FVM and ETHStorage, it is believed that within the next two to three years, models that are hundreds of megabytes in size can be stored on the blockchain with ease.
You might ask, what are the benefits of putting models on the blockchain? Once the model is on the chain, it can be directly called by other contracts, making it more crypto-native and enabling more possibilities. It gives a sense of being fully on-chain, with all data being chain-native. Currently, some teams are exploring this area, but it is still in the very early stages.
Another direction is for the medium and long term. If you carefully consider the concept of smart contracts, it is actually best suited for “machine-to-machine” interactions rather than human-machine interactions. On the AI side, there is now the concept of AutoGPT, which allows you to create your own “virtual avatar” or “virtual assistant” that can not only chat with you, but also perform tasks based on your requests, such as booking tickets, hotels, buying domains, and building websites.
Do you think it is more convenient for an AI assistant to operate your various bank accounts and Alipay, or to facilitate transfers using a blockchain address? The answer is obvious. So in the future, will there be a bunch of AI assistants integrated with AutoGPT-like features that can 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.
That’s all for the first part, because it’s really taking up too much space. The next part in a few days will briefly discuss parts 2, 3, 4, 5, and 6.
Like what you're reading? Subscribe to our top stories.
We will continue to update Gambling Chain; if you have any questions or suggestions, please contact us!
