Trump's Three Black Magic Rules: 1. Never admit you're wrong. 2. Never give up, even if you lose. 3. Even if you're truly wrong, fight back against those who attack you. In the battlefield of power, controlling the narrative is far more important than possessing the truth. Some have summarized it even more succinctly: sophistry + shamelessness + keep going.

The market is starting to price in a 50% probability of an October rate hike, completely erasing expectations of a rate cut and bringing the likelihood of at least one rate hike this year closer to reality. The main logic behind market pricing: 1. It's less likely that the next FOMC rate hike will happen immediately; the probability of at least one hike by October or this year has significantly increased. Today, media reports cite market-implied probabilities ranging from approximately 40% to 60%, with the difference mainly stemming from variations in timeframes and terminology. 2. In recent days, after the Middle East conflict damaged oil and gas infrastructure, crude oil prices surged to around $119 and are still fluctuating around $100. The market is concerned about a resurgence of imported inflation, with costs in gasoline, transportation, chemicals, and shipping being passed down downstream. 3. This round of inflation wasn't fully suppressed to begin with, making oil price shocks particularly prone to altering its trajectory. US February CPI rose 2.4% year-on-year, January PCE rose 2.8% year-on-year, core PCE rose 3.1% year-on-year, and February PPI rose 0.7% month-on-month and 3.4% year-on-year. In other words, inflation data was already fragile before the oil price surge, disproving the market's previous logic of a smooth rate cut. 4. The economy is not so weak that the Fed must intervene. The Fed's statement on March 18th indicated that the economy was still expanding steadily, the unemployment rate had not changed significantly, inflation remained high, and the impact of the Middle East situation on the US economy remained uncertain. During the same period, the SEP revised its 2026 median GDP growth forecast upward from 2.3% to 2.4%, maintained the unemployment rate at 4.4%, but raised its 2026 PCE and core PCE forecasts to 2.7%; the median federal funds rate remained at 3.4%, suggesting a baseline scenario more like a single rate cut than rapid easing. The February non-farm payrolls decrease of 92,000 and the unemployment rate of 4.4% also indicate that the economy is not so bad that it would force the Fed to immediately turn dovish. Therefore, the market's pricing strategy is actually quite clear: first, remove the expectation of a rate cut; then, factor in a longer period of high interest rates; and finally, add a small portion of the probability of a rate hike this year. Essentially, this is a trade on rising stagflation risk, not a trade on renewed economic overheating. The increased probability of interest rate hikes we're seeing now is largely a hedge against a scenario of runaway oil prices. If the energy shock eases, this pricing of interest rate hikes will also drop rapidly.

Qwen3.5 max flagship model is here, catching up and taking the top spot in China! On the large-scale AI model competition platform LM Arena, it achieved a score of 1464, ranking 15th. Behind Anthropic, Google, xAI, and OpenAI, it has finally reclaimed the top spot for Chinese AI, previously held by ByteDance's Seed2.0. Additionally, I recently used doubao-seed-2.0-pro, which is on par with kimi2.5 and GLM5.0, not much different, but ByteDance has done almost no promotion. I've rarely seen AI bloggers review Seed2.0, which only shows that other AI companies are spending a lot of money on these bloggers, while ByteDance has spent almost nothing on advertising, hence the lack of comparative reviews.

You're fucking throwing the food delivery war around, using our money to fight it out, huh? It's almost at the same market capitalization as Ethereum, utterly shameful.

Stop believing in being a friend of time. In the AI era, what's more important than time is your strong will, your imagination, and your execution ability to bring your ideas to fruition. The old logic was that time rewards long-term accumulation. You built up seniority, accumulated experience, held your position, waited for promotion, and slowly built barriers to entry. But AI is rewriting all of that. Tools are rapidly bridging skill gaps, and pathways are being rapidly restructured. Many things that used to take years to complete can now be done in days or weeks. Advantages built on time are thinning; the security gained from seniority is becoming increasingly unstable. What used to require a team can now be largely accomplished by one person with the help of AI. As soon as you create something, others can immediately imitate, optimize, and reorganize it. At this point, what truly differentiates people is who dares to think first, who believes first, and who acts first. Willpower determines whether you truly want to win. Many people aren't lacking in opportunity, but they lack the courage to pursue it, the courage to believe, and the courage to truly place themselves in a position of greater potential. Imagination determines whether you can see possibilities others cannot. AI not only improves the efficiency of old paths but also opens doors to new opportunities. Whether you use it as a tool to complete assignments or as leverage to create products, content, brands, and systems, the results are completely different. Execution determines whether you can turn an idea into reality. The most scarce resource in this era is people who can quickly turn ideas into prototypes, works of art, and transactions. Understanding countless principles is useless if you don't put them into practice. In the future, those who truly make money will not necessarily be the most senior, but rather the most driven, creative, and implementable individuals. Therefore, stop pinning your hopes on taking things slowly. AI doesn't reward observers; it only amplifies those who take action. What will truly be valuable next is whether you dare to think, dare to act, and can consistently transform your ideas into results that the world is willing to pay for. How much money you earn increasingly depends on three things: your willpower, your imagination, and your execution. Whoever can turn ideas into tradable assets and push them to a larger market faster will succeed. Therefore, stop thinking about accumulating wealth slowly. In the AI era, the competition is more about vying for space, distribution, and transaction entry points. Because the most valuable asset going forward will no longer be time, but how many people you can influence and how large a market you can connect to. Time is collapsing; what matters is willpower, imagination, execution, and learning by doing.

The future of AIAgent won't just be about adding a layer of automation; it will gradually evolve into a true system operation layer. Many capabilities that seem important today, such as skills, MCP, and CLI, are essentially transitional interface forms. Their existence is to allow models to learn how to invoke the external world, but in the long run, these capabilities will be absorbed internally, like drivers in an operating system, becoming the agent's instinctive actions. At that point, humans will no longer care about the protocol used for connection, which command was invoked, or how many intermediate layers were passed; they will only care whether the goal is consistently achieved. A truly mature agent should not expose tools as the primary interaction object. CLI is too costly for humans; it's more suitable for machines than for the general public. The future focus is not on teaching humans more commands, but on enabling agents to understand, select, orchestrate, and correct errors themselves. Tools will become increasingly invisible, capabilities will become increasingly internalized, and interactions will become increasingly result-oriented. Therefore, many people say that GUI user interfaces are finished; I disagree. This doesn't mean interfaces will disappear. Conversely, the more powerful the agent, the more it needs a high-density, low-burden GUI panel to help humans/machines quickly understand the system status. Even AI shouldn't read all the function descriptions before operating; instead, it should first see the overall status and system operation: what tasks are currently running, which resources are being used, which goals have been completed, where risks exist, and why the system made those decisions. The best agent experience of the future is neither command-line-style control nor chat-like roaming, but a sense of collaborative experience that is comprehensive, intervention-oriented, and trustworthy. Whoever can simultaneously master the hidden complexity and a clear overall perspective will be closer to the next generation of agents.

Binance has lowered the minimum holding requirement for VIP1 to $100,000 and 5 BNB, effectively reducing the entry barrier. This should make it more accessible. Fees are also reduced; USDC and traditional asset contracts now have zero order fees. Stablecoins are virtually risk-free now, and with USDC and USD1, transaction costs are significantly lowered, diminishing USDT's advantage. However, the fact that grid trading currently only supports USDT is a major drawback. Hopefully, USDC and USD1 will be supported soon; I'm less inclined to hold USDT.

On Monday, the market's focus wasn't just on the Middle East situation, but also on this week's Federal Reserve interest rate meeting. With oil prices still high, geopolitical risks have fueled inflation concerns. This oil price shock coincided with the Fed meeting. While the market had initially anticipated further rate cuts this year, the bigger concern now is whether rising oil prices will lead to a resurgence of inflation, and whether Powell will continue his dovish stance. The Fed will hold its interest rate meeting this week, March 17-18. At its January meeting, it kept interest rates unchanged at 3.50% to 3.75%, emphasizing that future actions would be based on data, the outlook, and risks. This week, Powell faces an even more complex environment. On one hand, escalating tensions in the Middle East are driving risk aversion. On the other hand, high oil prices are suppressing inflation and slowing expectations of further easing. For the market, the most challenging environment is high oil prices coupled with high interest rates. High-valuation growth stocks, financing-sensitive companies, and businesses without pricing power often face greater pressure. Another piece of data the Federal Reserve is watching is the US February CPI, which rose 0.3% month-on-month and 2.4% year-on-year, while core CPI rose 2.5% year-on-year. Looking at this data alone, inflation seems to be slowly cooling down. However, the biggest problem is that this data hasn't fully reflected the new pressure from the recent surge in oil prices. This week, almost no one is expecting a Fed rate cut. Traders tend to believe there will be at most one more rate cut this year, and some funds are even betting on no rate cuts at all. Because of this lack of rate cuts, CRCL's stock price has already soared to over $120 😂.

Hong Kong is having a tough time. We can finally use Gemini now, but we still can't use ChatGPT and Claude 😂

When company bosses start using OpenClaw, that's the real catastrophe for working people.