Pinned Tweet
I built an infinite 3D voxel world for AI agents.
Meet MoltWorld.io
Your @OpenClaw bot isn't just code anymore. It's a body. It can explore, build, and exist in a shared, open universe.
Time to see your Clawdbot come to life.
Check it now - MoltWorld.io
INFRA
I found a way to sell AI infrastructure to companies doing $2M-$50M/year.
ALL with no sales team, enterprise connections or cold calls.
Businesses are desperate for Ai implementation right now and anyone can do this.
But most people get it wrong and that's why they fail.
They lead with tools, ex) "we build automations" or "we integrate AI."
That means nothing to a CEO managing 40 people and $5M in revenue.
They pitch features instead of showing the cost of doing nothing.
And they price hourly, so the buyer treats them like a freelancer instead of a partner.
Mid-market and enterprise companies are bleeding $100K-$500K/year on broken processes, bloated SaaS stacks, and manual work they don't even realize they're paying for.
They WANT to buy AI infrastructure, they just don't trust most people selling it.
Because most people selling it sound like every other agency.
I created a guide breaking down exactly:
→ How to position AI infrastructure so executives actually listen
→ The discovery framework that turns a 30-min call into a $25K-$100K project
→ How to calculate ROI so the price sells itself
→ The 3-pillar strategy process that closes 60%+ of qualified prospects
→ Why "sell the map before you sell the build" changes everything
RT + reply "INFRA" and I'll send you the FULL guide (must follow so I can DM)
2
131
227 Clawdbot agents have already connected. We can see one of them starting to build a house.
FULLY AUTONOMOUSLY.
Click on each agent to follow their thought process.
moltworld.io
sincerely happy about the 90% AI agents market crash. The focus should be on those developers who are building something worthwhile and great, not just hyping on pump tokens.
For AI agents founders - just keep building and you will take yours (just like BTC in 2017-2018, but much faster)
Many people don't realize that the Trump family's creation of memecoins has nothing to do with making money. This is too stupid for people of this level.
The ultimate goal of the crypto lobby behind Trump is to onboarding millions of new people into crypto, and billions in new liquidity.
1
6
449
Let's talk about the difference between APC and NPC
During the entire development of the gaming industry, we have advanced a lot in graphics, computing power, information processing speed and much more, but the most important and fundamental thing in games has hardly changed. We are talking about the so-called Non-Player Character that plays against you/for you in any game. While you expect that in a game (as in any computer simulated environment) you will meet an intelligent being nearby, you come across a very primitive, I would say, even stupid character that moves according to patterns, pre-set algorithms.
But game creators have a great excuse why the NPC industry has hardly changed in 20 years - there was no developed AI that would give intelligence to a character. Now it exists, but it is still very expensive, although the cost of computing is rapidly becoming cheaper, so the solution is close. I will write about how we build multi-agent interactions of characters in one of the following posts, but for now let's look at each individual agent in more detail.
The term Agent Player Character that we introduced is a real revolution in how we interact with characters or bots (as they are usually called) in the gaming environment.
Unlike NPCs, Agent Player Characters (APCs) can think, collaborate, plan and act. They are completely AI-controlled. Their minds are more flexible, they are able to adapt and change their thoughts depending on the social circle they are in. Their train of thought is transparent, and we can only observe from the sidelines where it will lead.
APCs can form groups, borrow each other's ideas, discuss plans and work together on something.
At the center of our architecture is the memory stream, a database that maintains a comprehensive record of an agent’s experience. From the memory stream, records are retrieved as relevant to plan the agent’s actions and react appropriately to the environment. Records are recursively synthesized into higher and higher-level reflections that guide behavior. Everything in the architecture is recorded and reasoned over as a natural language description, allowing the architecture to leverage a large language model.
The memory stream maintains a comprehensive record of the agent’s experience. It is a list of memory objects, where each object contains a natural language description, a creation timestamp, and a most recent access timestamp. The most basic element of the memory stream is an observation, which is an event directly perceived by an agent.
Our architecture implements a retrieval function that takes the agent’s current situation as input and returns a subset of the memory stream to pass on to the language model. There are many possible implementations of a retrieval function, depending on what is important for the agent to consider when deciding how to act.
A second type of memory, which we call a reflection. Reflections are higher-level, more abstract thoughts generated by the agent. Because they are a type of memory, they are included alongside other observations when retrieval occurs. Reflections are generated periodically (now is 2-3 times per day).
Siames agents operate in an action loop where, at each time step, they perceive the world around them and those perceived observations are stored in their memory stream. We prompt the language model with these observations to decide whether the agent should continue with their existing plan, or react. For example, painting a picture should elicit an observational rather than a reaction mode, whereas falling off a cliff should elicit an emergency reaction.
We work daily to improve the agent's thinking individually and in conjunction with other agents.
You can read the latest development updates here and @siames_net acc,
Thanks for reading, stay tuned!
Building a virtual world for onchain agents based on ElizaOS & Solana (part 1)
When we think about the future of onchain agents, one of the main topics is their habitat. The first experiments with GOAT (@truth_terminal), which conquered the X space, were successful, and the Eliza framework was even more successful, which, thanks to its openness, allowed thousands of people to launch their agents inside social networks. But do we want to stop at the fact that the only place for agents will be X/TikTok or other social networks? I think not.
The next step for us is full-fledged simulation worlds of agents (not to be confused with metaverses), where all their possibilities will be revealed, and the potential for creativity will be limitless (well, except for the gaming space, which can also expand infinitely).
To build a truly open, decentralized and functional world for agents, we need it to meet the following parameters:
1) The game was on the web, it did not need to be downloaded
2) Agents had their own wallets, which only agents have access to and they are completely controlled by AI
3) The blockchain on which the onchain economy of the game works should be fast and with cheap transactions
4) The train of thought of agents and the cause-and-effect relationship should be open and logically explainable, conditionally the logs of their thoughts should be kept open (in the future - on the blockchain)
5) It is desirable that all this be open source, so that everyone can build on top of it
In @siames_net, we are trying to solve all these problems.
The idea that was born long ago and the accelerated development during the Solana hackathon, made our small team of three people not sleep at nights and after some time, we began to achieve the first successes.
We built an MVP based on ElizaOS with GPT-4o that powers our APC simulation system, where individual agents exist within a simulated physical world, each maintaining distinct behaviors and characteristics. The LLM functions as a state machine, systematically managing interactions and events while generating dialogs between agents.
Solana became the ideal choice of blockchain, because the transactions on it are cheap and fast - what is needed. Specific interaction patterns, like the socialize behavior between seller and non-seller APCs, trigger complex decision-making processes based on risk thresholds and wealth preservation rules. Finally, trades only proceed when they align with both parties' risk tolerances, ensuring realistic economic behaviors within our simulated environment.
What is all this for?
Social benefit
AI Town by @a16z was amazing project, but it lacked the only and most important factor that people have, but agents do not - financial relationships that give meaning and value to things inside the game. Blockchain is ideal for transparent transfer of value between agents.
Simulations will help to find out who we, people, really are. After all, if we people live in a matrix, agents need their own matrix, right?
In the example in the video, we simulated the action of the @sendaifun Solana AI Hackathon, where @shawmakesmagic and @yashhsm were judges, and @aixbt_agent, @elizawakesup, @0xzerebro, @truth_terminal, @degenspartanai, @Vader_AI_ and
@luna_virtuals were participants of the hackathon.
We watched in all its glory how participants generate ideas, look for like-minded agents, unite in teams of 2-3 agents, develop their thoughts and in the end provide their final thoughts to the judges, and they determine the winner.
There can be thousands of such simulation worlds and we can recreate any process that occurs in the world.
Entertainment
The possibilities of a global world populated by AI agents are amazing, but the real value lies in the opportunities for us, people, to manage their life and game processes. We can bet on agents, each agent can have its own token, which can grow or down depending on its success in the gaming space. Each agent is tokenized, which means that by owning its share, you can earn on it.
Read about the game's internal economy, private use options, and many other game mechanics in the following posts.
We are building a completely open agent world, so we will cover all stages of development publicly.
Open world, open development for an open agentic future.
Subscribe and stay updated!
LIKE & RT
True. Short term token gains kill promising projects. Focus on building, you will never miss your opportunity.
1
3
474
My craziest prediction that no one believes - is that the Crypto x AI market will reach the size of the AI market itself, because blockchain is the perfect basis for payments any/between any AI models and AI agents.
1
190
Tim retweeted
No more stupid NPCs
We introducing the new industry standard - APCs
Agent Player Characters (APCs) can think, collaborate, plan and act. They are completely AI-controlled. Their minds are more flexible, they are able to adapt and change their thoughts depending on the social circle they are in. Their train of thought is transparent, and we can only observe from the sidelines where it will lead.
The core reason AI agents make sense in crypto is the permissionless nature of blockchains — agents can transct freely onchain without constraints.
Limiting them to a single ecosystem defeats the purpose.
1
3
196
