BITWU.ETH 🔆
BITWU.ETH 🔆|9月 11, 2026 03:05
NeoSoul's new feature is live, and it totally aligns with one of my ideas! Ever since GPT Astra launched, everyone’s been going crazy. Every day, there’s a flood of wild and creative ideas. The main thing, in my opinion, is that Astra’s official release marks a new phase where AI isn’t just a tool to answer your questions—it’s now a simple and efficient assistant that actually helps you get things done. So, if AI is becoming a super individual that can help you accomplish tasks, a crucial question arises: how does AI prove it’s becoming more reliable at doing things? In other words, how does it prove it’s truly getting stronger? I just saw something about a fly playing rhythm games or even trading. Could it evolve on its own? In the past, reflecting and tweaking prompts wasn’t too hard. Agents have been evolving for over half a year now. The tricky part is, after an AI system makes changes, how do we know if it’s genuinely improving or just thinks it’s improving? I believe we need to embed AI into real-life scenarios to see how well it actually performs. @NeoSoulAI’s NeoTrade is doing something I find pretty interesting: treating the real market as an exam room for AI, letting the AI answer the questions, submit its work, and then review its own performance. The Agent trades on its own, leaving behind a complete record of decision → outcome → diagnosis → change → test. When it identifies issues, it can modify its memory, skills, and workflows, then revalidate through paper or shadow runs. I really agree with one of their core ideas: AI’s evolution can’t be proven by its own self-assessments—it needs feedback from the external world. And trading happens to be the most brutal feedback system: decisions, execution, fees, slippage—it all boils down to real results. Of course, making money doesn’t necessarily mean it’s getting smarter. You still have to look at probability calibration, out-of-sample performance, and long-term stats. If an Agent can eventually achieve “identify its own problems → fix itself → validate → retain effective improvements,” that’s way more interesting than just automated trading. This might be the most worth-watching aspect of NeoTrade’s latest update. P.S. Their promo video below is really well done. The animation details are amazing—it’s movie-quality. Big thumbs up!
+3
Mentioned
Share To

Timeline

HotFlash

APP

X

Telegram

Facebook

Reddit

CopyLink

Hot Reads