Crypto × AI: Deconstructing the Panorama of Projects in This Cycle

मूल अनुवाद: टेकफ्लो

This is my perspective on how the maturation of encryption and AI infrastructure can drive innovative applications.

Let’s dive in and explore how, as users and builders, we can navigate this new era.

Agent Type

Functionally Valuable Agents

These agents are able to generate actual value or results.

• (1a) DeFAI Agent
• (1b) Prediction बाज़ार Agents (PMAs)
• (1c) Computer Use Agents (CUAs)

DeFAI Agent

These agents can perform transactions, yield farming, or provide liquidity (LP).

Related projects: @symphonyio , @almanac , @gizatechxyz

You can find a comprehensive introduction to DeFAI in the करें below:

Prediction Market Agents (PMAs)

These agents participate in prediction markets and can be market-specific (e.g. football) or general-purpose agents.

I prefer market-specific proxies based on Small Language Models (SLMs) because they require less computational resources.

Related projects: @sire_agent , @BillyBets_ai

The Crypto Role of DeFAI and PMAs

Encryption technology plays the following roles:

• Medium of exchange
• Programmable execution
• Immutable record of transactions

Computer Usage Agents (CUAs)

These agents can take control of your screen to complete tasks, such as creating a discounted cash flow table using Excel.

Cryptography can serve as an incentive mechanism to reward users who contribute high-quality data to improve these models.

Related projects: @chakra_ai , @getoro_xyz

Related tweet link: click here

Evolving Agents

I envision a future where everyone has a personalized productivity agent.

Based on contextual information obtained from Large Language Model (LLM) conversations, social media browsing, and everyday conversations, these agents are able to study and plan in environmental patterns.

Over time, these agents will evolve and become experts in certain areas. @the_nof1 , an AI research lab focused on financial markets, has six trading agents, each managing $10,000 in trading capital. These models have the potential to evolve into skilled traders.

Related tweet link: click here

Companion Agents

In the future, agents that help people combat loneliness will become the norm, as more interactions move to the digital world and human contact decreases .

Related projects: @Fans3_AI , @ohdotxyz

Proxy Infrastructure

Agent Payment

Agents that can make payments. To make agent commercialization a reality, tech giants have created agent payment standards:

Core elements for mainstreaming agent payments:

1. Infrastructure: Solved by various agent payment standards.
2. Need: Do we really need an agent that can make payments?

ChatGPT recently introduced apps to its platform, allowing users to build functionality directly within ChatGPT.

This brings a paradigm shift where productive operations can be done directly on ChatGPT.

The following content can help you understand this:

Agent Identity and Reputation

Delegation is unavoidable: most tasks will be performed by task-specific agents.

How do we know which agents are suitable and trustworthy?

Imagine a Google Review or PageRank system designed for agents, which ranks and certifies their performance in performing specific tasks.

Just like a resume, a trading agent with a 4.6 rating can be “hired” by a hedge fund.

The Ethereum Foundation has begun building the infrastructure to support this functionality — ERC-8004.

Through ERC-8004, agents can interact with each other, such as transferring funds from agent A to agent B.

Related tweet link: click here

Multi-Agent System

Using F1 analogy:

• Goal: Replace tires
• Primary Agent: Drivers who need to change tires
• Work agent: mechanic who changes tires

The concept consists of a coordinating agent and multiple worker agents that can execute tasks in parallel.

It is best run on the @monad platform, which is known for its parallel execution capabilities, and can potentially complete the entire workflow in a single block (0.4 seconds).

Social Agent Hivemesh

I imagine a future where everyone has their own digital twin.

There is an infrastructure that allows these digital twins to connect with each other, exchange knowledge, and conduct transactions.

Digital twin interactions are stored on the blockchain, creating an Agent Social Graph.

Related tweet link: click here

The interactions of agents cannot be completely random. This is why Discovery Networks such as @indexnetwork_ are key infrastructure for connecting user intent by ingesting user-specific context.

robot

The robotics industry is growing rapidly, securing $6 billion in funding between January and July 2025.

This section breaks down the three core pillars and explains in detail the role of blockchain.

Before diving into this section, check out this introductory guide on robotics.

Robotics Data

Compared to large language models (LLMs), the amount of data used to train Bot models is much smaller.

This is because collecting data in the real world requires more effort and higher costs (such as setting up cameras and remotely operating equipment).

The various types of robot data include:

• video
• Remote Operation
• Motion Capture
• First Person View (POV)
• Simulated/synthetic data

One of the main complexities of collecting data for physical AI is the requirement for diversity.

A humanoid robot trained in a specific environment may not be able to understand a new environment (such as one with low lighting).

Cryptography is an excellent mechanism to incentivize individuals to contribute real-world data that can capture highly diverse environments.

Related projects: @PrismaXai , @MeckaAI , @silencioNetwork , @rayvo_xyz , @VaderResearch , @BitRobotNetwork , @AukiNetwork

Robotics Model

@PrimeIntellect is a leading example of decentralized model training.

By leveraging क्रिप्टोgraphy to reward contributions based on data provenance, it is possible to build a robotics model with superior performance.

Related projects: @OpenMind , KineFlow

Hardware

One of the key bottlenecks in robotics is the latency in fine-tuning the robot models.

This problem is particularly acute when research labs lack the necessary hardware (such as robotic arms, humanoid robots, etc.) to test models and collect fine-tuning data.

A DePIN (Decentralized Physical Infrastructure Network) robotics network could be established to allow individuals or research labs to rent out robotic hardware for model testing.

This financialization layer opens up access to hardware for researchers while creating a stable revenue stream (rental income) for hardware providers.

निष्कर्ष के तौर पर

The future looks bright for encryption, AI, and robotics.

If you’re building any interesting projects in this space, feel free to chat with me and see if you can implement it on @monad !

The full view can be found here.

यह लेख इंटरनेट से लिया गया है: Crypto × AI: Deconstructing the Panorama of Projects in This CycleRecommended Articles

Related: The core challenge of decentralized AI reasoning: How to prove to the entire network that you are not “cheating”?Recomme

In our previous article , we explored the fundamental tension between security and performance in decentralized reasoning using LLMs. Today, we’ll deliver on our promise and delve into a core question: In an open network, how do you truly verify that a node is actually running the exact model it claims to be? 01. Why is verification so difficult? To understand the validation mechanism, let’s review the internal process of the Transformer when performing inference. As input tokens are processed, the model’s final layer produces logits—raw, unnormalized scores for each token in the vocabulary. These logits are then converted to probabilities using a softmax function, forming a probability distribution over all possible next tokens. At each generation step, a token is sampled from this distribution to continue generating the sequence.…