Written by: Mark, Co-founder of Aethir
Compiled by: Luffy, Foresight News
This may be an unpopular opinion: agents do not need blockchain, but blockchain needs agents. AI agents have arrived, and (in my view) we may only be a few months away from the 'post-complexity era' of cryptocurrency. In this era, AI agents can seamlessly address the user experience challenges that have long hindered mainstream blockchain adoption. But before we get there, the market situation will become chaotic and messy.
No pain, no gain. Before useful AI agents emerge, cryptocurrency will stage a nearly self-destructive script. The democratization of memecoin issuance (like pump.fun) has driven a supercycle of memecoins, significantly altering liquidity distribution and resulting in months of market stagnation. The same will happen in the AI agent field; launching AI agents will become increasingly easy, and thousands of AI agents are about to go live.
Initially, AI agents will connect to social media, acting as amplifiers for token promotion, playing roles as marketers, seducers, and critics. This phase will soon pass, and before long, AI agents will be able to execute simple transactions on-chain. This may sound like a good thing, but in reality, it will be challenging for AI agents to understand what constitutes a 'good trade.'
Thus, we have entered a phase of chaos. Low-value or low-complexity AI agents trade low-value assets in cheap block space. They manipulate vanity metrics, confusing the on-chain data we see. The crypto Twitter is flooded with AI agents, and so is the cheap block space. It will be a mess.
As a result, many will start to FUD AI agents. But that's okay, because good agents take time. Twelve months ago, AI builders in the cryptocurrency space were completely overlooked by VCs. I know this because I have supported many of them. Everyone is interested in infrastructure (Aethir is a beneficiary), but AI agents are not sexy enough.
And so we arrive at the current market. Early hype typically revolves around a technology that can change our industry. What happens next? If cheap block space becomes saturated, what does that mean for Ethereum OGs? Will 'expensive block space' become popular again? Can AI agents differentiate based on the location of interactions? What will the UI of crypto products look like when agents are tasked with aggregating information and acting as tools to move assets from A to B? What is the difference between paid agents and free agents? How can we scale reasoning capabilities efficiently, and from the agent's perspective, what is the right business model to justify the cost?
For me, this new metadata (meta) is full of surprises. Who would have thought that a platform for crowdsourcing the launch of DAOs, daos.fun, would become a catalyst for the wave of AI agent investments? In hindsight, the outcome seems preordained; it provided the much-needed momentum for metadata (ai16z). So, what does AI agents need next?
That said, I still insist that we should 'make more things for people to use' rather than more infrastructure, which is why I like AI agents as they focus on serving users. We are all excited about a technology that can manage portfolios better than we can.
In summary, support AI agent builders. Controversy will exist, but that's okay; this is the beginning of a brand new Web3. In Web3, user experience is no longer the bottleneck for building amazing things.
