DeepSeek V4.1 Flash， Helpless choice Since GPT forcibly reduced the intelligence of Chinese users, the previously purchased Pro 20x has been relegated to obscurity, purely a waste of money and adding to its own burden. So, copy DeepSeek V4.1 Flash and see if it can work. In actual use of the official DSH framework and standard mode (related to DeFi projects), my feeling is: The ability to write code is basically sufficient, but the intelligence may not be very online. Give an intuitive example. The issue of Jupiter's quotation sliding point failure. Set a 10% sliding point, but the price ranges from: 0.025 SOL can buy 32081 → 20850 MEME Token. Can you tell me that this indicates a 35% drop in token prices?! That's a 35% surge, okay? I'll be good. In addition to this, I spent several hours communicating with it about a strategy today, and I found that the understanding of V4.1 Flash at many strategic levels was wrong or seriously deviated from my original expectations. I can only repeatedly correct its erroneous views. This makes Dim sum tired. I originally thought that AI should not only have a hundred times stronger coding ability than me, but also be able to cover all aspects at the strategic level, with the ability to both pick up high and let go lightly. But this V4.1 Flash often makes mistakes, so you definitely wouldn't dare to fully entrust it with strategy, let alone when it comes to funding issues. A few days ago, using Fable 5.1, the whole process was basically in a state of: it said xxx, I said yes, yes, go ahead according to your plan. Now it's my turn to constantly correct errors in V4.1 Flash, but my brain is not enough. Isn't I looking for AI just to replenish my brain? As for why I said the code ability was okay earlier, it was because I analyzed the code written by GPT 6 Astra and found N problems with V4.1 Flash, which were fixed one by one. This indirectly proves how severe the intelligence reduction of GPT 6 is and how it does not treat Chinese users as users. Speaking of this, who can be blamed? It's only the model companies of the domestically produced Light that are crazily distilling that are to blame. so-called: The predecessors planted trees, and the descendants enjoyed the shade; Previous generations cut down trees, and future generations suffer. We have become pure Chinese GPT retail investors who have suffered.