In the AI era, where will we go from here? Everyone is anxious because AI is developing too fast, increasing efficiency exponentially, and many jobs and companies may be disrupted. The most crucial question here is: what will be the final form of AI? This question is important because historical experience tells us that in the course of a revolutionary development, intermediate forms become cannon fodder. In fact, even the newest AI technologies have become obsolete in the past two years. For individuals, this easily leads to a grueling cycle of frantically learning new skills every day. I've been pondering this question for a long time. Recently, an answer has become increasingly clear: personalized, private AI. 1. Complete Personalization Powered by AI Everyone is an independent individual with different preferences. When it comes to food, some like spicy food, some prefer mild flavors; there are hot pot, barbecue, spicy hot pot, and so on. When traveling, some like hiking, some prefer a relaxed and leisurely pace. The core value of business lies in solving problems. Currently, the basic business model is that large companies address the common needs of most people, while mid-sized companies address the concentrated needs of a small group. In the future, AI will be used to address the diverse needs of each individual. From the perspective of actual social development, the Industrial Revolution increased productivity, leading to a shift from standardized goods to diverse styles. From portal websites where everyone consumes the same information to personalized recommendations like those on TikTok, all these developments reflect a growing demand for greater personalization. In the future, with the support of AI, completely personalized needs can be met, satisfying the unique preferences of each individual. 2. The Relationship Between Humans and AI There's a concern that AI, with its current capabilities, will eventually eliminate the need for humans. This is an extremely pessimistic view. The fact is, every technological advancement changes many things. For example, after the advent of the internet, newspapers and magazines began to disappear, but self-media bloggers and livestreaming e-commerce hosts emerged. The core issue is that as long as there are people, there will be demand. And human desires are never truly satisfied. Once something good exists, if something better appears, that's what we want too. As for the possibility of AI developing self-awareness and the fear of AI ruling the world as depicted in movies, I think that's unfounded worry. Moreover, AI is a human invention; even if AI becomes powerful enough, why not become a friend? 3. Many Intermediate-Form Companies Will Die Many current products are intermediate forms and will inevitably be phased out. For example, Vibe Coding's IDE mode is a fork of the VS Code interface, which is designed for programmers. For ordinary people, there's no need to look at so many functional interfaces. Just clearly describe what you want, and let AI code to implement the desired functionality. Currently, society has a large number of intermediate-form products because AI replacement will take time. Most companies are simply adopting AI by adding it to their products or having employees use AI to improve efficiency. However, they haven't realized that they are merely intermediate forms. At the same time, entrepreneurs should constantly remind themselves not to be bound by fixed models/knowledge of the existing world, and to break out of established thinking patterns; otherwise, it's easy to create intermediate-form products. Instead, they should think more about the future capabilities of AI and the new social forms that AI will build. 4. The Future Form In the future world of AI + humanity: 1) An important form will definitely be the personalization brought about by AI. The popularity of Openclaw aligns with this underlying logic. 2) Large-scale models and related infrastructure – this is self-explanatory. 3) Professional services + AI-oriented interaction methods – this will also be prevalent in the future. While AI capabilities are very strong, they cannot replace all knowledge, experience, professional skills, and unique data in specific fields. These will become even scarcer in the future. However, the best future form requires these to be interactive with AI. In other words, these professional services will not take their current form, but will be accessible and usable by AI. Okay, that's all for now. I'll share any further thoughts later.

It's time to buy some, family!

@riyuexiaochu This is great! I previously made an agent-to-agent call, but I think this is better and more reasonable. Mine is lighter. I'll try Superset today and see how it goes.

After a breakout with increased volume, the $FOLKS 1H chart is currently in a healthy pullback confirmation phase. The price is consolidating around 1.538, and the 1H EMA20 (1.4566) has become strong dynamic support. The 4H chart shows three consecutive bullish candles, breaking through the previous consolidation range and indicating a bullish trend. Market logic suggests further price increases, and combined with stable open interest, this appears more like major players entering the market rather than a simple sell-off, building momentum for a subsequent rally. 🎯 Direction: Long ⚡ Entry/Pending Orders: 1.530 - 1.535 (in batches near the current price) 🛑 Stop Loss: 1.485 🚀 Target 1: 1.620 🚀 Target 2: 1.700 🛡️ Trade Management: - Position Sizing Recommendation: Standard Position - Execution Strategy: After the price reaches 1.620, reduce the position by 50% to lock in profits, and move the stop loss of the remaining position up to the entry price of 1.535. If the price fails to hold above the 1H EMA20 and falls below 1.520, consider exiting the position early. Depth Logic: The order book depth imbalance reaches 16.47%, with the buy (Bid) depth significantly exceeding the sell (Ask) depth, indicating strong support below. The 1H RSI (64.45) is in a healthy strong zone, not yet overbought, suggesting further upside potential. The 4-hour chart shows the price has stabilized above all EMA moving averages, indicating a bullish structure. With a funding rate of only 0.005% and no overheated selling pressure, the bullish momentum is likely to continue. View real-time quotes 👇$FOLKS {future}(FOLKSUSDT) --- Follow me for more real-time crypto market analysis and insights! #Lobster #meme #robo @BinanceSquareCN $ETH {future}(ETHUSDT) {future}(BTCUSDT)

In the AI era, a peculiar phenomenon emerges: On one hand, a group of people are dedicated to using AI to collect trending articles from various platforms, automatically summarizing them so their owners can see the top 10 trending events of the day as soon as they wake up. On the other hand, another group is dedicated to using AI to write viral articles daily, capable of writing, adding images, and publishing to various platforms with a single click, ensuring their owners see highly-read results as soon as they wake up. Something feels off here.

Concrete Data Proving AI's Decline in Intelligence: Claude Opus 4.5 This data shows the Marginlab team using Claude Code's Opus 4.5 daily to perform pass/fail tests on 50 questions in the SWE-Bench-Pro test. The data shows that the pass/fail rate has dropped from 60% in early January to 54% currently. The rate of AI decline is 10%.

Here are a few recent observations: 1. For coding ability, CodeX GPT 5.2-CodeX-High is the strongest. It's suitable for writing complex products or finding hard-to-find bugs. It can solve several bugs that others couldn't. The downside is that it's extremely slow; don't use it for minor modifications. Modifying a small part can take a week to read a novel like "A Record of a Mortal's Journey to Immortality." 2. For coding ability, Claude Opus 4.5 is the next best. Its quality and capabilities are second only to the top, but it's much faster. It can also be used in antigravity mode. 3. Gemini 3 Pro. It's the strongest for front-end development. I used it for all my front-end work, and its aesthetics were generally good. Besides coding, it's also highly recommended for other scenarios, such as daily conversations and connecting to the Google ecosystem. Google is incredibly generous; you can use it for free for a year with a US education qualification. 4. Claude Sonnet 4.5 is rather awkward. I haven't used it for a while. Its capabilities aren't as strong as the previous three. Using the API is quite expensive. For a monthly subscription, Opus 4.5 is recommended. 5. A cost-effective option is GLM 4.7. Their coding capabilities are decent, and their models are updated quickly. Most importantly, it's affordable, less than 20 RMB per month. You'll likely never use it all. It can be used for minor modifications, etc. 6. Don't trust AI rankings or those self-media outlets too much. They claim every new AI is the best in the world. I don't know if they've been paid to do so or if they genuinely don't understand. They're just wasting other people's time and energy.

Found a great crypto blogger, the title of Big Cousin's book. Don't miss my crypto life if you missed Binance Life. A new crypto life in 2026, still at the bottom! Worth buying in! Recently, older cryptocurrencies have been experiencing a resurgence, I believe 2026 will also be a banner year for crypto. My crypto life = starting a 100x life! #meme $BinanceLife {web3_wallet_create}(560x5b68e58418f0cf6111922ea8ff0b5e88e8894444)

The most frustrating thing for Vibe coders is that every few days later, another model emerges that dominates the leaderboards, boasting world-leading metrics. You can't possibly try every single one, so you rely on social media bloggers. You'll find that most of them are just hype. Then, after finally settling on a suitable model and product, you're ready to get down to business. But within days, the model seems to have lost its intelligence. It's truly speechless. Recently, I discovered that Gemini3 Pro and Claude 4.5 have both become less intelligent. Yesterday, I used Gemini3 with antigravity to write a task, and it kept encountering problems. Okay, I tested a few commonly used ones, and the results were: 1. Gemini3 has indeed lost its intelligence. 2. I tested Claude Code using Claude 4.5, and it also showed severe intelligence loss. 3. I used GLM 4.7 in Claude Code, which recently appeared claiming to be the world's best. While Claude 4.5 found a few more issues, it didn't find 1/3 of them. Among the four cursors, GPT 5.2 performed the best, missing only one minor issue; the others were all found. Based on this, GPT 5.2 is the most reliable. The only headache is that GPT 5.2 code is too concise; much of it is incomprehensible, making it a headache to read. 😀

A group of high-ranking Yapper accounts on Kaito have been discovered, exhibiting very similar behavior: Followers: Approximately 20,000 to 50,000 Tweet views: Around 1,000 per post Many comments under each tweet Also, they tweet an incredible amount every day, about 20 tweets a day! Truly workaholics!