He Yi analyzes user thinking, intelligent interaction and proxy metaverse through Cookies. I have never considered myself a trader, and I may never be one.
Therefore, I don't mess with complex "tools", no price alerts, don't spend time on products like "Bloomberg Terminal", and absolutely don't touch leverage or perpetual contracts.
Everyone probably knows by now that my style is to pick a few projects or trends that interest me, and then spend a lot of time delving into the details. This is also why I can consistently produce two in-depth analyses of over 3,000 words per week. However, I have to admit that I do feel overwhelmed in keeping up with all the developments in the crypto x AI space, which is part of my job.
Every morning when I wake up, I always find some new things that surprise me that I didn't know about. I'm not saying I need to understand every project and drive myself crazy (most of them are just nonsense anyway). But I do feel that I'm still missing out on some very important developments.
Let me ask you a question: if I spend more than 10 hours a day researching these topics and closely follow crypto Twitter, why do I still feel like I'm not keeping up and missing so many new narratives? I think the answer is that my information input efficiency is not high enough. I work hard, but I don't have a smart enough strategy to stay on top of things.
Doing research and writing is one thing (I think I'm still pretty good at that). But the other part of the job is knowing what topics to research and when to research them. Based on this, let's talk about how I can improve my way of acquiring projects, trends, and narratives. The key is that this is not something I want to spend a lot of time on, but rather to establish a daily 10-minute routine in the morning over a cup of coffee, so that I can know what the crypto proxy Twitter is focusing on.
First, I will focus on using BlockTempo. I need to clarify that this is not an advertisement, I just feel that by seriously integrating Cookies into my workflow, I can better stay on top of things.
This morning, I decided to spend about $7,500 to buy 10,000 Cookie tokens to unlock the advanced features.
Seize the Opportunity
Over the past few months, the BlockTempo team has left a deep impression on me.
When Goat launched in October, the team immediately realized that the proxy space would be the main narrative of this bull market.
So they decided to go all out and build the best analysis and data platform for AI proxies.
While there are other tools like aiagenttoolkit, kaito (not limited to proxies), etc., I think for the needs of most people, BlockTempo is undoubtedly the best choice right now.
Last December, I had the opportunity to exchange with the founders of BlockTempo, and I was impressed by their discussion of their plans for the next few months and how they are trying to keep the product as crypto-native as possible (such as token gating features, in-depth crypto Twitter analysis, etc.).
Brief overview of the tool:
Tracking over 1,000 proxies: including their heat on crypto Twitter, social media interactions, market value and influence comparisons, etc.
The team is indexing on-chain data and social media data, providing a comprehensive analysis of the dynamics in this field from both qualitative and quantitative perspectives.
The basic features are open to everyone, and anyone can explore on the website. You need to hold 10,000 COOKIE tokens to unlock the advanced features. The market cap of COOKIE tokens is currently between $150-200 million, and it just went live on Binance perpetual contracts today.
In addition, it also provides rich APIs for integration by other projects. For example, Virtuals has integrated the "Sentiment" feature into its interface.
Free Features
This section outlines the situation in the proxy space: the number of proxies and the market capitalization.
I'm not particularly concerned about the number of agents, many of which are projects of average quality. But the market capitalization indicator is indeed very useful, as it shows how much capital has flowed into the crypto agent field.
I'm not saying we should compare the market cap to other fields, but I do find the chart provided by Messari quite interesting. I'm surprised to see that the trend of the agent field is very similar to the DeFi market cap in 2020/2021.
The peak market cap of DeFi was around $65 billion, and then the entire crypto market crashed in May 2021 (who else remembers that terrible period?).
We are currently about 5 times away from that digital number... Maybe this can be used as a potential reference point. However, as I mentioned in the previous article, don't be overly influenced by these cyclical comparisons.
How are the key agent ecosystems performing
Currently, it's mainly Virtuals and ai16z, but I'm curious what changes will happen in the next 6 months.
Earlier today, I read an article about the new agent framework Pippin, and I found the first two paragraphs quite meaningful to share. Although I don't have much comment on the framework itself, the following quote reminds us that we should not rely too heavily on any one ecosystem at the moment. After all, Goat has only been online for 3 months!
A common phenomenon in the crypto field, especially in emerging hot areas, is that many people find a "good project" and form a tunnel effect. This approach may be effective in the short term, but what happens when variables change and you don't adapt in time? In a field with only 4 months of history, it is very naive to think that the current leaders will always be ahead, especially as more advanced developers and technologies continue to emerge.
——JW100x
The heat of AI agents
In the past two months, I have never used the term "heat" as frequently as I do now. But the fact is, it is indeed important, as it can help us understand the focus of the market.
Currently, most of the heat is dominated by platforms and major players like aixbt and zerebro. But interestingly, why are those small players with heat below 5% suddenly emerging? Has there been any noteworthy information?
Index table
This function is self-explanatory, basically displaying an overview of all agents, which can be sorted by heat and market cap.
An obvious approach seems to be to discover emerging narratives by looking for projects that rank unusually high in heat but have a relatively low market cap. So if a project with a market cap of less than $10 million has a heat comparable to a project with a $100 million market cap, there is likely something new brewing.
KOL Mentions
This is a really cool feature. You can see which KOLs have mentioned a certain agent on Twitter, and directly jump to their tweets. This is basically tracking the types of influencers involved and how they impact the heat. It's also interesting to see how early some people got involved and check how they came across the agent. This is a great way to refine the sources you use to discover new narratives.
Honestly, this is a very creative feature, and I didn't even think I would need something like this myself.
Advanced Features
Holders and Whales
For each token, you can see detailed holder distribution, understanding the proportion controlled by insiders. Specifically, I think the proportion of the top 10 holders is particularly important for projects that are less than a month old and haven't yet gained attention. This is the simplest way to judge whether a project truly has an execution plan, or is just a "rug pull" project.
The only caveat here is to make sure to check if the developers have locked their tokens in any liquidity pools (if so, the holder proportion may not be displayed, but I want to emphasize this as a strong advantage).
The reason I think this indicator is important in research is that sometimes ambitious, imaginative ideas come with a contract address and a Twitter account. But the reality is, are these visions currently realistic? Or are they just empty talk?
For example, last November, I mentioned a project called CHAOS, thinking it had a cool robot + agent vision, and while I still find the concept interesting, the project ultimately wasted my time, and I'm glad I didn't invest more effort in researching it.
Social Media Interactions
This is a great way to review the performance of proxies in releasing content and what tweets people are most interested in. There's not much to say here, the only thing to mention is that it saves me a lot of time, as I don't have to go through different accounts and play with Twitter lists.
Sentiment Heatmap Dashboard
The next two screenshots are from a special section where you can access the detailed sentiment heatmap dashboard provided by S4mmyEth (if you haven't followed him yet, you really should).
This chart shows the latest momentum changes. Given that I'm already very active in this field, I think this chart is probably the one I use the most. Any situations where there is volatility within 12 to 24 hours are points I need to check, so that I can get a deeper understanding of any new trends.
For example, suddenly (at least for me) it feels like there has been a massive shift towards proxy x metaverse/virtual characters this week, and I'm having a hard time understanding why. I did some research and found that Holoworld was the main driver. This also made me have to look into the relevant documents and form my own views.
The second chart is similar to the first one, but it compares market capitalization and smart engagement on a weekly basis, basically to judge whether a project is a short-term hype or can continue to attract interest. I will pay special attention to projects that continue to see increasing smart engagement even when there is a pullback in the proxy space.
Market Cap vs. Attention Scatter Plot
Finally, a feature worth mentioning is the market cap vs. attention scatter plot. This is a really great chart because it allows you to instantly understand which potential projects are "about to explode". The screenshot above is a zoomed-in version, and you can clearly see aixbt, whose market cap is far below its attention. This is mainly because it is not a "proxy framework", and the current narrative trend is to pour a lot of capital into the "L1 projects" infrastructure we mentioned in our Tuesday article.
But if you zoom out to the bottom left, you'll start to see some interesting alphas. I've blown up the chart four times, focusing on all proxy projects with a market cap below $20 million. To my surprise, there are so many projects I've never heard of before. But at least now I know it's worth taking the time to understand what Seraph, Kudai, Kween, Trust, etc. are. This has greatly simplified my judgment process by preventing me from getting stuck in my own little biased circle.
It's important to note that just because these charts suggest I should pay attention to these projects, it doesn't mean I'll dive deep into each and every one of them. All the projects on the charts may not be worth my time at all, but the key is that this approach is clearly a more effective project screening method than endlessly scrolling through Twitter to find alpha.
That's all I have to share about the Cookies platform today. I know the content above is somewhat biased towards the "trading atmosphere", which is a bit strange for me, as I've always hated looking at charts in projects. My preference is clearly for white papers and calls with founders.
But I think it's time to correct some of my biases and understand that the attention and smart engagement metrics provided by Cookies are different from ordinary trading strategies.
If one of the success indicators in the proxy space is understanding the direction of the metaverse, then my responsibility is to find a way to filter out the most important information.
