I've been so busy since Tet (Lunar New Year), and crypto is struggling this year, so I'm going back to try some traditional business ventures with some close friends. Today is almost finished. Hopefully, everything will go smoothly this year. I don't wish for wealth or luxury; just having a little extra to cover the BTC) costs would be enough to make me happy. After several years of being on leave, just staying home trading cryptocurrencies, I've realized how much my health has deteriorated since I started going out and about. Even just lifting light things makes me breathless. It's completely ruined ☹️ The workload isn't too heavy, I can manage it, but the time I spend resting to clear my head and get back to normal is probably more than the time I actually spend working. Trading, analysis, and research are all tasks that require critical thinking; an unstable person will immediately fail at them. The optimistic thing is that the trading bots are still running very smoothly. I'm currently testing the Capital to XEM there are any issues with liquidation or slippage. Once this step is complete, I can go public with the AiOi community 🙏🏻 I've been tinkering with AI late into the night, but haven't tried Claude Code yet because everyone's complaining about constant limitations. Until a week ago, I thought Grok was smart enough. Especially after it added a Multi-Agent response model, I think it's become significantly smarter. I successfully built a function that previously cost me $400/year in just two nights of tinkering. However, when using OpenClaw, I realized that Grok's model is much dumber than Claude's; it's almost impossible to implement complex functions with Grok's API. The only problem is that Claude is a total money burner and can't survive. The "Wow" moment just happened the night before last. I wanted to build a basic management page to use as a place for daily sales and inventory tracking for myself and my friend. For various reasons, I wasn't comfortable using my current sales management software 🥲 For a coder, it would probably only take 3 days with Grok to finish. But for a non-tech person like me, I gave up. Grok constantly proposed impractical ideas, and even after dozens of revisions, they were still buggy. I almost wanted to give up, accepting that AI is still a long way from creating a product just by speaking. But coincidentally, yesterday I read an article about a Vietnamese guy boasting about earning $1000 and 2000 global customers despite not knowing how to code, using Claude Code. Boom. I'm all excited to experiment again. But for some reason, I still have a soft spot for ChatGPT. And Codex shocked me, in just two days. The website was almost complete, and the important thing was that the communication with Codex was incredibly smooth. It was like I was the customer and they were a seasoned IT engineer. Every bug, every improvement could be fixed with just two or three chat sessions. With this Codex, I'm sure I can build countless other things to support my research and trading to my liking. I was fortunate enough to be curious and interested in AI from an early age, witnessing its evolution from the days when you had to buy a foreign SIM card to activate ChatGPT. In just three years, I've clearly seen the intelligence of all the models. And I have to say, it's growing exponentially 🥲 I previously thought it would take three years for the world to truly be divided into two groups: those who know how to use AI-assisted technology and those who don't. But perhaps it will only take one year... Things will be less busy from tomorrow, back to work with Crypto. Sorry for being absent for the past few weeks, but there will be lots of exciting things coming your way soon! 💪🏻