Yesterday at the Base press conference, it was announced that the use of Flashblocks technology has improved performance by 10 times, and transactions can be completed in just 200 milliseconds. A friend asked me how this is achieved. Just now, I looked at the documentation and found that the core is to use a pre confirmation scheme similar to Base Rollup, which divides the original block into ten sub blocks and places them one by one according to the order of transaction initiation. Every time a sub block is filled, the transactions in that sub block are internally pre confirmed as valid. After saving enough 10 sub blocks, it is packaged into a normal block for broadcast. Because it is divided into 10 sub blocks, the performance is also improved by 10 times. You will definitely ask a question, isn't this the difference in block size that existed a long time ago? Yes, logically it is like this, but the biggest difference is that these sub blocks are not actually true blocks, they are just confirmed internally without real consensus and block generation, so it is called pre confirmation. So this process also brings a risk and a benefit. The risk is that it is an internal pre confirmation and has not completed the normal block production, so there may be a risk of invalid transactions. This completely depends on the reliability of the nodes. However, to be honest, these L2 are all centralized with only a few nodes, but the degree of centralization has been increased from 99% to 99.9%, so the problem is not significant. Another advantage is that this approach is equivalent to ending MEV, as there is no longer a process of block sorting. Instead, transactions are pushed to sub blocks for confirmation in chronological order of their initiation, and then packaged into a block for broadcasting. Therefore, all transactions on the chain are in chronological order, and there are no tips or clips involved. So currently, adopting pre confirmation for the entire L2 is a big trend, and it is estimated that other chains will follow suit in the future.