On-chain AI agent big pullback, should we buy the dip?

Original Title: What's next for AI agents: The Third Wave
Original Author: 0xkyle, Crypto Kol
Original Compiler: zhouzhou, BlockBeats

Editor's Note: This article discusses the future development of AI agents in the crypto market. The author believes that current AI agents mainly perform simple response tasks, and in the future, they will shift towards more interactive and practical applications and avatars. The author is optimistic about AI projects on Solana and recommends two investment strategies: buying large-cap AI projects like AI16Z and investing in small-cap, developing interactive and application-focused AI projects, believing these will yield higher returns.

The following is the original content (reorganized for readability):

Let me briefly say that if you follow me, you will see me posting a lot of profound nonsense, such as "the place to skate is where the puck will be" or "buy low, sell high attention," which are all pretentious statements.

But there’s a reason I keep saying these things— I believe the crypto market, even after so many years, is still very inefficient. People continuously chase short-term gains, wanting to feel the thrill of FOMO, experiencing the immediate dopamine rush from rising numbers, only to find that it often signals a peak after 20 minutes, and their positions quickly become trapped.

Therefore, the investment skill to make money is— buy low, sell high. Buy when no one is talking about it, sell when everyone is discussing it. Buy when you feel everything is over, sell when you want to take a screenshot. Buy cheap, sell expensive.

So the purpose of this article is to discuss this; AI agents on Solana are currently in a significant correction phase. Meanwhile, altcoins on centralized exchanges are rising, and Base is siphoning off all the liquidity. There’s much discussion about whether SOLETH and SOLBTC have already peaked, and the overall sentiment is quite pessimistic. Therefore, I am betting on two things:

1. Solana will rebound, and we will see a second wave on-chain.

2. AI agents will bounce back.

Now that the background is set, I will start my monologue about how I see the future trend if AI agents on Solana rebound. Here’s my framework:

Macro: Hype -> Fundamentals -> Practicality

Micro: Responders -> Fundamental-ish -> Applications/Avatars

Macro = the defining narrative of each cycle; the first wave included Render, Akash, Tao, etc., and many general junk projects also emerged. Render and Akash are almost unusable, I don’t know what FET and OCEAN are for (or their evolution, AGI, etc., I genuinely don’t understand), not to mention NEAR (the first memecoin launched on it was completely crushed, and I was there at the time).

The second wave includes Goat, Zerebro, AI16Z— more foundational applications.

Next is the micro cycle = types of applications that will emerge, Goat brought various chaotic projects— memesAI, Shoggoth, Project89, etc. The applications following Goat are deeper— Zerebro is now releasing music tracks, Lola can conduct on-chain transactions, AVB has collective behavior, etc.

Now, the macro and micro are colliding in the third wave— I believe we will see the practicality of applications and avatar forms. You are already starting to see its prototypes; Alch now allows users to build their own applications using AI.

Griffain allows you to control your wallet with AI agents, AIXBT is everyone's favorite analyst, helping researchers increase their analytical capabilities tenfold.

"Responders" will become a thing of the past; if your AI agent can only respond on Twitter now, it doesn’t deserve to be at the top. We are already accustomed to this situation; people always want more.

So, the next wave will no longer be just "responders," but actual interactivity. Avatars are great because they allow you to converse and interact with them. Applications enable you to build things, truly using AI, rather than just having a bot that replies to you in the comments.

Therefore, there are several ways to participate in this trend. In the past few weeks, the AI space on Solana has experienced significant downturns, but many are still working hard.

Undoubtedly, if you follow my tweets, you will know that I started to be optimistic about AI16Z since the 180mm FUD. I believe AI16Z is the easiest way to access the entire AI ecosystem because Eliza has been widely forked, just like in the last cycle when people didn’t want to pick individual games to invest in but rather bet on gaming chains— I believe AI16Z is the "L1" of AI agents.

Strategy 1: Buy large-cap AI projects that perform well; for example, AI16Z, and possibly TAO.

Strategy 2: Buy small-cap AI projects that are doing the types of projects I just mentioned.

I choose Strategy 2 because it can bring asymmetric returns in this field. Many AI agents are currently not favored by the market, as I mentioned in the previous paragraphs— low attention, rotation games, etc.

These AI projects are actually building the applications, interactivity, avatars, etc., that I am optimistic about in the upcoming wave.