I’ve seen a couple write-ups on the potential value of the @_kaitoai airdrop but I haven’t seen anything on why this is one of the few novel, game changing products
Imagine if you had the smartest, most objective folks (more detail in next tweet) in the world that had the ability to consume every tweet related to crypto. They then filtered out all the shitposts, toxicity, and then ranked each tweet based on originality, relevance and impact.
Then you (or in this case the Kaito team) could ask them any series of questions and they would answer them in a quantitative, detailed fashion. This is effectively what Kaito is today.
But what's to come is why I think it will be game-changing, and I think with the new Trump admin what they are trying to build will be entirely possible.
It’s been a while since I’ve been so bullish on a product. I think if you log onto cryptotwitter everyday this is worth digging further into
How can I say that Kaito has the smartest, most objective folks?
They don’t have ppl manually working for them. They have an in-house, proprietary algorithm that effectively reads through every tweet.
I was surprised to hear that they owned their model rather than outsourcing it or using a generic, open source LLM. I dug into this a bit more and noticed that they have a text embedding (text embedding is a fancy way of saying tagging a tweet with the relevant topics e.g. this tweet would get a kaito tag) model on a bittensor subnet -- checkout @openkaito. When they ran a comparison against openAI’s baseline model, they outscored it (image below)
This is definitely step up in caliber relative to the bulk of “AI” related products in crypto today.
Additionally, the more tweets the model consumes, the better it gets – this is how models work, and since the model is reading through all the tweets, the improvement is happening as we speak.
And its evident that other people and protocols are interested in utilizing this.
@eigenlayer did their surprise creator airdrop in collaboration with Kaito -- they used Kaito's dataset and scoring system to enable this. @Punk9277 (kaito's founder) also mentions that there are tons of protocols lining up with interest.
So why am I excited about what's to come?
Watching through @KevinWSHPod pod with Yuhu (Kaito's founder), they talk about this dynamics of consuming information and describe the 3 stages as:
1/ You read through everything, and then use your judgment to determine what is important and what isn’t 2/ You train an algorithm to read through everything, it uses its judgment 3/ You allow anyone (and any model) to read through everything, and use market forces as a way to weigh opinions and use the market’s judgment
The majority of us operate in mode 1. Kaito currently enables 2.
But the 3rd stage is what they are building towards. Yuhu suggests that there’s a possibility to “tokenize” a KOL in a fashion where their earnings are distributed to token holders in a trustless fashion. And this is where market forces come into play -- this effectively allows us to speculate on influencers (and AI agents!).
In a Biden/Gensler regime, I would’ve laughed at this thought, but now that Trump is the POTUS, this actually seems extremely viable from a regulatory POV.
It’ll be cool to see how the Kaito team actually executes on this
altho controversial, kaito is starting to shape into the identity layer
a twitter handle to wallet mapping is invaluable to protocols, esp when you give analytics and metrics attached to how much reach and influence a particular individual has
Asking about the business use case for blockchains (L1s) is backwards looking.
This is the equivalent of asking what is the use case of <insert fav chat app> when messenger, whatsapp, imessage, etc. all exist. Is there really a difference between these chat apps?
Sure, there are nuances around security, UI preferences, and whether the backend architecture is optimizing for consistency vs. availability but at the EOD, for 99.99% of users, there isn’t a difference. The same thing can be said about L1s themselves
So what allows certain blockchains to survive (and thrive)? Why did Solana come out significantly ahead of Avalanche when they came up around similar times? And how can you reason about what is going on with SUI currently?
Some reminders (to self) as things start to get parabolic
1) If you are in “everything” you aren’t going to outperform. You don’t need to catch every coin. FOMO is the mind killer
2) Think through who the marginal buyer is relative to the info source. Are you getting advice to buy a coin from a trusted friend via DMs? Or is a bigger account telling you to buy a 5M market cap coin?