Unpopular opinion: AI Agents do not need Block, but Block cannot do without AI Agents.
Today, Agents (intelligences) have emerged, and in my view, we may only need a few months to enter an "era of decomplexification" in the crypto domain. In this era, AI Agents will seamlessly solve the long-standing user experience (UX) challenges that have hindered Block's mainstream adoption. However, before reaching this stage, the entire industry will go through a period of chaos.
Teaching AI Agents to Identify "Good Trades" is Difficult
No pain, no gain - to make intelligences truly useful in the crypto domain, the entire industry may experience a "self-collapse" before that. Just as the introduction of meme coins (such as pump.fun) democratized and significantly changed the distribution of market liquidity, leading to months of sideways market, a similar phenomenon will also occur in the intelligence domain.
As the cost and threshold for developing intelligences become lower and lower, thousands upon thousands of intelligences will be deployed one after another.
Initially, these intelligences will mainly connect to social media, becoming unidirectional tools for Token dissemination, sharing, promotion, inducement, or mockery. However, this model will soon become boring. Subsequently, intelligences will begin to execute simple transactions on-chain. Although this sounds like a good thing, in reality, it is very difficult to teach intelligences how to identify a "good trade" (you can ask anand iyer for his views on this).
We Are Entering a Chaotic Phase
Therefore, we are entering a "chaotic phase". In this phase, low-value or low-complexity intelligences (Agents) will frequently trade low-value assets on low-cost Block spaces. These intelligences will manipulate some surface indicators, disrupting our ability to discern real on-chain data.
Crypto Twitter will be flooded with content generated by intelligences, and cheap Block space will be heavily occupied. The entire industry may descend into chaos.
Many people will start to experience fear, uncertainty, and doubt (FUD) towards intelligences. But this is not a bad thing, as it takes time to develop truly excellent intelligences. 12 months ago, crypto intelligence developers were even struggling to get the attention of venture capitalists - I know this firsthand, as I was supporting many teams at the time. Back then, the focus was on infrastructure (such as AethirCloud), and the potential of intelligences was not yet widely recognized.
The Potential of AI Agents
Now, we are at the early hype of a new technology that has transformative implications for the entire industry. What will happen next?
- If low-cost Block space is heavily occupied by intelligences, what does that mean for established Blocks like ETH?
- Will "premium Block space" become a trend again?
- Will intelligences be differentiated by the Blocks they use?
- As intelligences can both aggregate information and serve as tools to transfer assets from point A to point B, how will the user interface (UI) of the crypto domain evolve?
- What are the differences between paid intelligences and free intelligences?
- How can we expand the reasoning capabilities of intelligences in a cost-effective way?
- From the perspective of intelligent applications, how can we design appropriate business models to cover these costs?
Unexpected Surprises
For me, the most interesting part of these new trends is the unexpected surprises. Who would have thought that a platform for crowdsourcing the launch of DAOs - Daos.fun (daos.fun, baoskee) - would become an opportunity to drive the development of asset management intelligences. In retrospect, this seems obvious, although perhaps a bit ahead of its time, but it has indeed injected much-needed momentum into this trend (ai16z.ai and Shaw have also contributed to this). So, what do intelligences need next?
Nevertheless, I remain firmly in support of "building more things that people will actually use", rather than continuing to pile up more infrastructure. This is also why I like the intelligence trend - it is centered on the ordinary user. We are all excited about this technology because it can manage our assets more efficiently than we can ourselves.
Finally, please support the developers of intelligences - although they may face FUD, it's okay, because this is the starting point of a new Web3. In this Web3, user experience (UX) is no longer an obstacle to creating exceptional things.
Do your own research (DYOR), but stay optimistic - the future will be very exciting.
