Tomorrow's talking points:
- Creature Cloud: A bridge, a launchpad, a place to 🥩.
- BasedGod (AI Dev Loop): Community deployed AI Product Managers & AI Engineers working for $BASED.
- BasedNode v1.1: Updates to network based on your feedback.

(Holding release details on BrainGecko launch until next week)

- Major developments in 2024
- Pepecoin stake to earn $BASED is a very smart mechanism and a breakthrough itself. What if you could stake to other projects (with levels of safety) and earn their reward? Creature Cloud as a bridge explores this and automatically supports bridging of the other projects to $BASED L1 --- win. win. win.

- Pump.fun considerations
- Pump.fun is extractive, but the concept of “graduation” is useful as a value filter.
- “Better” projects are more likely to graduate due to this filter.

- Creature Cloud parallels
- Uses similar built rails in multiple categories, such as:
- Liquidity levels
- Time the project is active

- Six-month fixed graduation window
- Each project has a six-month period in which token rewards are distributed.
- The earlier (and longer) you stake Pepecoin to the project during this window, the more reward tokens you receive.

- Developer staking rewards
- While you stake Pepecoin, a small percentage is awarded to the developer.
- This developer portion vests, preventing immediate dumping.

- End-of-period bonus
- At the end of the six-month window, a final bonus of the reward coin is airdropped to Pepecoin holders still staking.
- A small percentage of the Pepecoin pool at that time is awarded to the developer.

- Risk and reward structure
- You stake Pepecoin but effectively pay nothing upfront.
- If the project performs, the developer is rewarded at the end.
- If the project does not perform, you can unstake and the developer receives nothing.

While primary use of Creature Cloud will be bridging from wrapped Based to $BASED L1, important to note that Creature Cloud projects automatically work with the bridge so projects can easily support compatibility with $BASED while launching from a parent chain with more users (starting with Ethereum).

The critical first step in autonomous engineering isn’t writing code—it’s figuring out what to build. In our view, the real breakthrough lies in having truly hands-off product managers who focus on the “why” and “what” of a project, rather than diving straight into coding.

On a technical level, each of the “Brains” in the system runs its own libp2p network with discovery, leader election, and complete autonomy. Each 1,024 Brains is a full libp2p network -- effectively separate blockchain networks, each proposing its best plans among miners and validators for new software, upgrades, and bug fixes for Based at intervals to the core commitee or Gigabrains. The most promising plans are selected by “Gigabrains,” who reward them based on their originality and value. These top plans then form the roadmap for our AI engineers, who actually turn them into code.

This approach stands apart from the typical AI startup model, which tends to focus on prompt-driven coding. Instead of just “building whatever the prompt says,” we prioritize the initial conceptual stage—ensuring we know exactly what needs to be created before any code is written. If our process is correct we will generate software faster and more comprehensivly than any startup in human history.

Dropping the content in a few hours, plan is to just let the community have the entire dump of data. Then will do follow up education posts throughout the week. Feels so strange to know you are at the end of an era and the beginning of a new one. Going to announce the new socials guy early next week.

We need to lay it out for the community and make it crystal clear that listing on the major exchanges is absolutely in reach. The strategy’s a tough one, sure, but listing is not just about the market action is about the chance to pull in millions of their users to the project. We’ve got the holder base and the tech to back it up, and MINING is going to heat up around the world again. Based needs to present its best self to grab their attention on list. Creature Cloud is solid, no doubt, but the arena? That’s the game-changer—nothing else stacks up. Nothing most of us have ever worked on stacks up. It’s just a few weeks from being ready to run on people’s machines, and that’s the move to make.

Hey everyone! We’ve got to come together as a community to shape the vibe of Creature Cloud to not become any kind of "pump.fun" culture. Day traders are all about fast, but real value takes time and patience. Let’s keep the fun and fresh ideas front and center for the launch. It’s about the ride, not just the numbers. To make this happen, we’re rolling out a new spot for open discussions and posts about every project. Think old-school bitcointalk energy (shoutout to the OGs who know). It’s built for deep, thoughtful stuff that sticks around and is easy to dig through later. Forget chat rooms where ideas vanish in the noise—this is about creating conversations that last and actually mean something.

Weekend research. Let people know if they are running metal at home instead of plugging in to Claude / OpenAI, team found these videos helpful (DYOR):

DIY 4× Nvidia P40 Homeserver for AI (96 GB VRAM)
- https://www.youtube.com/watch?v=dHTvpUlWFbk

M4 Mac Mini CLUSTER
- https://www.youtube.com/watch?v=GBR6pHZ68Ho

DeepSeek on Apple Silicon in depth
- https://www.youtube.com/watch?v=jdgy9YUSv0s

The vision for the next few weeks is that dozens if not hundreds of very specific technical and like-minded people take a solid look at the agent arena and contribute ideas / feedback.

Some months ago, we realized that this is the first clear beginning to the intelligence explosion (linked below). Having members of this community, technologically minded yet humanity-first leaning, would make me feel more at peace, guaranteeing that this couldn’t self-build itself into something evil, since … and this is critical … the fact that it is a decentralized network + self building would make it very hard to shutdown. We are working hard to get a specific schedule to launch this stuff but with the state of the market it’s not under a critical deadline.

My gut is to just dump a minimal prototype so people can understand what this is and how we’re developing it. So that is what we are going to do. We are working on a version of the Rust code that you can run and see with your own eyes how we’re doing this. The plan is to open source this ASAP, hopefully next week to get the concept out there and start hearing any possible concerns.

Individually several of our team have come to me or raised concerns via chat that this thing really could self-compound way out of control. There are guardrails in place, but to spend so much time trying to build something valuable to humanity only to have it possibly hurt humanity is oddly gut wrenching.

Don’t let these feelings put a damper on your weekend, the only reason we are communicating this is the implications and consequences of the technology. Of all the projects, all the communities, all over the world, the percentage of people who have been a part of anything like this is microscopic.

It’s a brave new Based frontier.

Learn more about the intelligence explosion here: https://situational-awareness.ai/from-agi-to-superintelligence/

If it’s not already obvious today’s o3 release further cements that soon you will be able to build ANYTHING, this pushing building down the value chain and deciding what to build to the top of the value chain.

This is Based’s strategy with the arena: thousands of models acting as product managers to crowdsource pain‑points, surface startup ideas, and write solutions for everything real time as it happens.

People think there will be a token for any whim and idea (there isn’t enough liquidity for that), we think there will be an app for every idea and whim.

Together, matching models that “can‑build‑anything” and Based Arena’s “know‑what‑to‑build” radar will unlock a paradigm of software development that no one has ever seen. Ideation, not implementation, is the new competitive moat.

I think we will have some code tomorrow we can look at, worst case we have a minimal python demonstration just to show what’s happening here. I really want to get this out to people but we don’t want to ship something that people can’t start building on ASAP. (context: https://openai.com/index/introducing-o3-and-o4-mini/)

Also @creatureOnBased API finally got reapproved on X after an unjust ban.

Just as an aside these are intended as "notes" not some kind of announcement channel. I want to be able to have a place to share strategy without some big announcement and we shouldn't make this channel anything other than it is. Thats why its never been discussed, its just strategy NOT a roadmap. For instance, I want to release a demo of the arena but I dont want to waste a minute of precious dev resources if there is something else. IN this case its better we launch more than a demo. Just like when we first released the paper we came out with a bang, the agent arena is supposed to be that. The market rewards patience and conviction, which I think applies everywhere. Thank you for your patience and especially those helping testing and running validators now. More soon.