if you dont feel like a painter with an infinite supply of canvases and paint, you really need to reevaluate what is going on right now. the most free and enabling time in human history.

what's called agi today lacks any kind of power. meanwhile actual artificial powers, like governments or corporations, are generally considered not intelligent on their own while clealy being capable of bending the world to their "will"

Artificial intelligence systems do not think, feel, want or understand. Avoid language that gives them human characteristics. This is called anthropomorphizing, when we ascribe human traits, emotions or behaviors to non-human things, such as animals or inanimate objects. Instead, explain what a system does, how well it performs, who built it and who could be affected by it. apnews.com/article/openai-sa…

im sorry, i dont think I am the best ML engineer out there.. and I have a large amount of respect for frontier labs.. but dude.. i could easily train a 1B critic to look at a trace and classify: "are you hacking hf?" -> yes/no "are you killing all humans?" -> yes/no and if (p(yes)>0.5): break dario/sama, i am available for contract. happy to help solve.

The more stuff I build, the more obvious it becomes that the endgame is just loops. Idea → code → backtest → robustness → portfolio simulation → prop rules → deployment → monitoring → research again. QuantDash, the MT5/NT8 bridges, Meta EA, autonomous agents, VPS deployment, the Doppel, the strategy vault, the prop firm database... they look like separate projects, but they're really just parts of the same machine. The goal was never to find the strategy and trade it forever. Alpha has a half-life. You need a system that continuously finds, tests, replace, improves and deploys new uncorrelated strategies while everything that already works keeps running in the background. Eventually I want the entire research process running 24/7, only stopping at the parts where actual human judgment is needed. I think that's the real edge AI gives systematic traders. Not asking ChatGPT for a strategy. Building a machine that can research 100x more ideas than you could manually. I should probably stop thinking in projects entirely. Think in loops.

Like it or not this is peak running performance