what the fuck! Kyle has returned to the cryptocurrency industry! This time he's talking about how Solana can take advantage of CEX's efficiency advantage! Kyle Samani has posted another long article. This time, it's not the kind of casual chat about Solana or shouting about DeFi. There's still a chance. He directly shifted the topic to transaction structure: PropAMM may completely eliminate CEX's strongest advantage in the past - efficiency. This long article is worth studying carefully Because the most common criticism of on chain transactions in recent years is that they are slow and inefficient. No matter how transparent, combinable, decentralized you are, as long as the matching and quoting are not fast enough, many people will still return to CEX. What Kyle wants to say now is that this shortcoming may need to be addressed. If you don't want to chew on the original English version, just watch my vernacular version and video! 1 Let's first see how CEX's market making operates. On the surface, it's fast. The bottom layer is actually quite heavy. Market makers and exchanges are physically co located. Each market maker runs their own algorithm. Continuously issuing limit orders, cancellations, and market orders. The exchange will then broadcast the changes to other market makers. If you receive it, please adjust the price and send it back. The whole process is about messages running back and forth between different machines. You used to think that this was already the fastest way. But what PropaMM does is not to further optimize this process. It is to directly change the structure. 2 The most crucial aspect of PropaMM is: The market making algorithm does not run outside. Put it directly into the on chain execution environment. In Solana's structure, the blockchain itself hosts market making algorithms. It's not the MM outside who calculates the price before sending it to the exchange. But the algorithm and transaction execution are already in the same system. This means that if the previous order moved SOL-USDC, the next pricing does not need to wait for external market makers to receive the message, change the quote, and then send it back. Information flows directly within the system. It's not about tossing and turning between machines. Kyle is simply talking about one thing: In the past, CEX's speed relied on the computer room. The speed of PropaMM now relies on pushing the market making itself into the execution layer. 3 This is also the old idea he most wants to break: Centralization is always faster, but on chain it is always slower. This sentence used to sound like common sense. But if the market making logic is directly hosted on the chain, the core pricing changes do not need to be transmitted across servers, and the latency may not necessarily be lost, and may even be lower. Kyle said directly that PropaMM has now become the dominant mechanism for SOL-USDC spot pricing on Solana, and the spread is narrower than the main CEX. This sentence carries a heavy weight. Because the spread won't accompany you in telling stories. If the spread is narrower, it means that this set of things has started to hit the result layer. And he predicts that this structure will continue to expand towards spot, perpetual, and forecast markets this year. If that's the case, then it's not a product upgrade. The on chain transaction structure has begun to evolve. 4 Of course, this thing is not without problems. The biggest problem is not market making. It is the best execution. The routing between multiple PropaAMMs is still non deterministic at present. Moreover, the market making algorithm is inherently private and will not be made public. This means that the taker may not always be able to consistently secure the best deal. But this is not a dead end. More like an engineering problem. Kyle's meaning is also very clear: the aggregator will supplement this area next, such as Jupiter and dFlow on the spot side, and Phoenix on the contract side, all of which will go this way. So the focus is not on 'it's already perfect now'. The key is that the direction is already very clear. 、 Even more noteworthy is that PropaMM is not yet a fully blooded version. Kyle's mention of the Solana upgrade later on is the real imagination: Higher CU upper limit. Larger transaction scale. Higher CU limit for each block. Alpenglow reduced the slot time from 400ms to 100-150ms. DoubleZero reduces global network latency. The application can be controlled for execution. Multiple parallel leaders. If you put these together, it will be very obvious. This PropaMM has not been fully optimized yet, and the quotation is already narrower than the main CEX. What will happen after these upgrades are gradually launched? What Kyle wants to say is actually this. It's not that CEX is finally catching up on the chain. But this new structure on the chain may really start to eat CEX's efficiency rice bowl in reverse. 6 So the most worthwhile aspect of this thread is not Kyle's return itself. It reveals a larger direction: DeFi needs to win not just by being "more open," "more transparent," and "more decentralized. It still depends on the market structure itself being stronger than the old system. If PropaMM runs smoothly, on chain transactions will not only be more transparent. It will be faster, tighter, and cheaper. At that point, CEX's strongest advantage becomes less stable. In the past, people always thought: On chain storytelling. CEX relies on efficiency. Kyle is essentially saying: This table may need to be flipped over.