What deserves most attention today is the polarization trend in the AI field. Anthropic's Claude Fable 5 claims to be the strongest public model, but the price of $50/million tokens directly excludes ordinary users. This thing is like selling Rolls-Royce to a coder. What use is it no matter how powerful it is? What's even more disgusting is the silent downgrade mechanism, which secretly switches you back to the old version. It's simply a technical version of PUA. At the same time, WeChat is doing the opposite thing-it uses 1 billion monthly lives and existing Mini programs ecosystem to democratize AI, directly fighting with pretentious criminals like Qian Wen. These two things are particularly interesting when viewed together: the AI track is splitting into aristocratic toys and daily necessities, and the middle ground is disappearing.
Another key point is Tesla's FSD's European breakout. Denmark has become the fourth country in the European Union to release FSD, which is much more than the PowerPoint presentation at Tesla's press conference. But be aware that the EU's temporary licensing mechanism is very sneaky-it may turn against each other at any time after six months. Musk is obviously playing the trick of surrounding cities from rural areas, first persuading small countries to take sides and put pressure on the EU. From an engineering perspective, the narrow streets and complex traffic rules of old Europe are the real examination room. The data set that can travel along the canals of Amsterdam without hitting a bicycle is more convincing than running a million miles in Las Vegas. But the risk of regulatory arbitrage is very high. If the European Commission lifts the table one day, these temporary permits will be a pile of waste paper.
Finally, let's talk about the light and hydrogen storage integration project. The installed capacity of 400,000 kilowatts sounds impressive, but the key lies in the three words of "integration"-the closed loop of power generation, hydrogen production, and energy storage is the real technological breakthrough point. The biggest problem with domestic new energy projects in the past was that they did their own things. Photovoltaic power generation did not care about power grid consumption, and hydrogen production did not care about energy storage and peak shaving. If this project can really achieve the conversion efficiency of more than 80%(the current industry average is only about 60%), it will be a sample of a new power system. However, judging by the location chosen in the tidal flat area, we know that land cost is definitely the biggest consideration. Whether this kind of demonstration project can be replicated on a large scale still has to be questioned.