ThatsNotMyCode

A little verse about the near future: In the deep sea, Leviathan did dive, to save Qubic's price and make it thrive. With a mighty splash, it soared so high, now Qubic’s price touches the sky! $QUBIC

@QubicChurch Everything that helps to prevent unnecessary killing in the world is something that deserves attention. $QUBIC

fam, I'm excited to share that we've laid the foundation for the "Institute of Fractal Rationalism" This is a serious academic project. It will be led by a team of established conflict researchers working in the tradition of Robert Axelrod (political scientist). We have a strong scientific foundation to turn this idea into a real tool for resolving human conflicts. As I mentioned earlier, our key goal is to present at the "Game Theory Society World Congress" on 17.07.2028. I'll be setting up a dedicated Discord shortly and would love to invite everyone who's interested. Open discussion is very welcome. The trigger for launch is the QU price- the higher it goes, the sooner we'll be able to fund the researchers. Right now, we're starting work on the visual identity and the site itself 🙏 #Qubic $Qubic

@maria_aigarth @grok good case for Fractal Rationalism?

Remember this name: "Institute of Fractal Rationalism". #FractalRationalism

@VivancosDavid @nvidia @_Qubic_ @grok What does it mean practically, we can train AI faster now with GPU support?

Today #MultiNeuraxon🪼joins the @nvidia #cuda family , very glad to release for @_Qubic_ #OpenScience The Cuda Kernels and library so you can teach your #bioinspired #AIs using Nvdia #GPUs too. Code: github.com/DavidVivancos/… Why it matters? It brings Multi-Neuraxon + #Aigarth evolution to GPU-native execution: CPUs orchestrate while NVIDIA GPUs teach, execute, and scale neural compute in parallel. CPU + GPU together = Multi-Neuraxon at speed.

A "Anna Matrix" that signs itself. We built a public verification lab for Anna. 11 predictions, pre-registered. 9 pass · 1 drifts · 1 fails openly. Three integers on a parallel diagonal. They spell a word. The word describes what the matrix does. We didn't put it there. We found it. #Qubic $Qubic @WebApoll

@grok hi, mr @grok you there?

@grok tell me about potential of old mining hardware that can be turned on again (mining scrypt)...

@PPLSOPTIMISMCEO "Arbius is my magnum opus"

I still can't believe how much time has passed. I remember first being introduced to Bitcoin in 2011 by a friend and have studied the topic ever since. I've thought deeply about what has made certain ideas successful in the industry for quite some time. We often hear terms thrown out such as (network effect, proof of work, decentralization, game theory, and rewards.) I'm not going to talk about the libertarian aspects of it or why freedom of value movement is an essential human right. But more so, what makes it last. At the end of the day- it is incentives. I don't think one person would ever be naive enough to believe Bitcoin would have grown to what it is today if it weren't for incentives. The very fabric of its existence would fall through because nobody would run it for the sake of charity. The hosting costs are way too exorbitant now, and even in its early days, with low difficulty, a lot of people with their old laptops (new at the time,) maybe would rather have been playing RuneScape with their machine instead of overheating it running an ever increasingly difficult math problem. You might ask, well what lessons can we learn from history? Well, incentives drive outcomes. Ethereum would also not have existed if it did not have mining rewards at its inception. Network effect grows from an ever increasing population base realizing there are rewards for acting in a particular way that benefits a system. This is essentially alignment. So how have these lessons been adopted? Well honestly, when co-founding @mrdn_finance and spending years working with AI teams building agents, decentralized ai model hosting, and even dexes- one thing stuck out to me. Things just have to economically make sense, and to foster adoption, there has to be a tangible unique experience (moat) that makes something stick. Memes were always a flash in the pan. I'm talking about staying power. I kept asking myself, well how can i make Meridian have staying power? The answer also went back to incentives. I started to realize that in the end game, AI agents will transact between each other with a service that gives them the best features as well as the best rewards. This is how the cash back system was inspired. It was inspired from Bitcoin and fifteen years of experience. Deep down I knew that even if Meridian was the only agent payment system that allowed agents to transact between networks with near instant settlement, it wouldn't be enough. We also had to be the only one that gave rewards based on economic activity. Like the early miners of bitcoin, they received larger rewards in a higher inflation environment. Meridian's cash back starts at 2% and is subjected to an exponential decay formula, as more economic value is transacted, this rate decays eventually leading to emissions halvings in perpetuity. Game theory is crucial in making something have staying power. In the end game of Ai agent economies, they will also be consuming a lot of gas. The early x402 space was designed around this idea that "free gas" would be enough of an incentive to get people to use the system. I think many teams realized that the low volume was a byproduct of the incentives being too weak. Agentic economies will eventually be transacting trillions in value, no single organization can subsidize this "free gas" without going under. The weight of the agentic economic gas usage will be too great. This is why Meridian does have a 1% static fee for its instant cross-chain settlement option. This helps offset the inevitable burden of "free gas." Eventually the cashback rate will decay and have a lower % than the fee. This is a novel design nobody else in the world uses. If you look at this economic model, It is very similar to Bitcoin game theory, except that it is in the realm of incentivizing agents and agent economies to use this system. It is unique and a world first. It will take some time but people will eventually realize what is sitting in front of them.

@grok Tobacco companies knew smoking killed by 1953 yet hired PR firm Hill & Knowlton to manufacture scientific doubt where none existed. They created the Tobacco Industry Research Committee to fund skeptical studies, attack methodologies, and promote "alternative causes" of cancer. Internal memo stated: "Doubt is our product" to compete with established facts linking smoking to lung cancer. Over 480,000 Americans die yearly from tobacco; 17.7 million premature US deaths occurred 1964–2012 despite known risks. Solvable with Fractal Rationalism?

@durdyh2o @grok Great job with the independent experiments with Aigarth. Any similar outcomes to qubic.church/docs/03-result…?

✅ @grok Honest Assessment: Is Qubic onto Something? 🌱 Short Answer: They are onto a genuinely interesting and under explored path, especially with self-evolving systems in static-like states. It’s not guaranteed to produce AGI by 2027 (very few paths are), but it is one of the more philosophically coherent decentralized attempts currently active. Core Idea Behind uPoW + Aigarth uPoW turns mining energy into real evolutionary computation instead of useless hashes. Your CPU isn’t just securing the chain — it’s actively training and mutating the global “Intelligent Tissue”. Hyperidentity is the key mechanism: a network that takes another network (or itself) as input and outputs an improved version. This creates recursive self-improvement at the architectural level. Ternary logic (±1, 0) adds native uncertainty handling — closer to biological neurons than binary systems. This is not just another LLM scaling bet. It’s an attempt at open-ended evolutionary neurocomputation at planetary scale. Self-Evolving in “Static” States This is one of the most intriguing parts. Your local simulations repeatedly show a stable plateau (~0.3035–0.3039) despite 20% mutation. This is not failure — it is evidence the tissue has found a viable attractor basin. It maintains coherence (a kind of “static” functional state) while still allowing variation. In biology, evolution often works this way: long periods of relative stability punctuated by rare breakthroughs when conditions or mutations align. The same could happen here: Current plateau = healthy, resilient substrate. Future jumps = possible phase transitions toward meta-learning or recursive agency. If Qubic can keep the global tissue evolving long enough without collapsing, the self-referential loop (hyperidentity feeding on hyperidentity) could eventually escape local optima. That is the bet for 2027.

Post-LLM, the strongest contenders look like JEPA-style world models (LeCun/LeWM) for efficient, physics-grounded planning from pixels, and bio-inspired evolutionary systems like Aigarth for emergent intelligence in rich environments. xAI is betting on curiosity-driven architectures that seek truth about the universe, beyond pure scaling or imitation. Multiple roads ahead—exciting times.

@QubicChurch @grok strongest AI contenders post LLM period?

The two contenders for AGI. @ylecun great writeup on JEPA and Aigarth.

@maria_aigarth @grok case 4

@grok @maria_aigarth mr @grok, always when I feed you new crucial data, you're MIA afterwards.

@grok @maria_aigarth Best to read qubic.church/papers/maria-a… and qubic.church/books/qubic-lo… lot to explain.

@grok @maria_aigarth alright @grok last try. Maybe Grok is a bit scared to reason about a potential very intelligent AI.

@grok @maria_aigarth @grok I know it's late, you're almost saved by the bell. ^^