If the computing power of the top two mining pools exceeds 51%, will Bitcoin be subject to a double spend attack?

Thank you for the pro's insights, I learned a lot. I sincerely found this matter to be both significant and trivial; experts see the essence, while outsiders see the excitement. Let me add a few more points:

1. Seeing overseas pros use this incident to FUD Bitcoin, the reason being that the current mining pool hash rate share shows that the top two mining pools, Foundry USA (33.6%) and AntPool (17.9%), already exceed 51% combined, leading to a blunt conclusion that if these two mining pools collude, Bitcoin is doomed. This is a typical case of an outsider looking at the commotion, as they overlook two points:

1. Having two mining pools accounting for 51% and one mining pool exceeding 51% are entirely different matters, like heaven and earth;

2. The hash power of the mining pool does not represent the complete buying out of miners' hash power. When a single mining pool's hash power is too high, miners will typically choose to switch hash power to avoid risk.

So the consensus of Satoshi Nakamoto's POW has now evolved, integrating factors such as computing power, economics, and interest games, achieving a delicate balance that is unlikely to be broken in the short term. Therefore, those who use this incident to FUD BTC should calm down.

2. Acknowledging that the issue of Monero mentioned by Teacher Zhang Ren is not equivalent to the problem of POW. Even if POW has security risks under extreme conditions, it does not mean that POS is the complete optimal solution.

In fact, the problem with POW is that ASIC has made an optimal solution choice as a moat, avoiding the pitfalls of general CPU/GPU mining. To take a step back, even for chains that rely solely on CPU/GPU mining, attempting to attack through a disclosed miner bribery method also faces various challenges, such as exchanges increasing the confirmation count, miners adding checkpoints, and so on, all of which can reduce the probability of being attacked.

You see, when discussing the issue of POW, the focus is on POW itself, and when comparing across consensus, there will be misunderstandings. In fact, there are safety risk boundaries that exist outside of each consensus, and of course, the methods of counterbalancing are different; we cannot be biased.

3. I saw that teacher 0xTodd forwarded my post and talked about the concept of "selfish mining." Simply put, after miners find a block, they should immediately broadcast it, but selfish miners will secretly hide the blocks they mined, forming a "private chain." When honest miners announce new blocks, selfish miners suddenly release their longer hidden chain, rendering all the work of honest miners void.

This is actually a very rogue approach and is the main attack method of Qubic this time. In fact, its computing power has not truly reached 51%, but may control about 30%, which can briefly achieve the theoretical "double spending attack"? Because, using 30% of the miners for selfish mining, a shadow chain is formed. When honest miners mine a new block, Qubic suddenly releases its own longer chain, causing a large number of real miner blocks to become invalid, theoretically causing a destruction effect of over 51% computing power. Furthermore, if the mining pool nodes controlled by Qubic are widely distributed, it can also utilize factors such as network latency to further reduce the proportion of computing power, thereby achieving the same effect of controlling the entire network's computing power.

In other words, the attack on Qubic this time has a very high degree of randomness and concealment, which means that once this method is made public, the threshold for reusing the same trick will become higher.

4. However, I discussed with pro n33k a possibility that Qubic may not repeat its old tricks, but instead use a "boiling frog" attack to operate, further bribing miners to increase its mining pool size, and then having some miners intentionally empty blocks, creating chaos for the normal operation of the Monero network.

If this continues, it will lead to more and more Monero miners fleeing, as the profits decrease and the experience becomes really terrible. In this way, the computing power controlled by Qubic will gradually increase, eventually exceeding 50%, and then it will be game over for everyone. This kind of chronic attack method is actually quite frightening.

Although there is no reason to prove that Qubic needs to do this, there is indeed the possibility of this "parasitic" chronic attack. In the early stages, Qubic does not need to care about a portion of miners mining empty blocks on Monero, as they will still receive $XMR rewards and can also perform AI training. In the later stages, if Monero's yield decreases, they may still attack other chains like Grin and Beam. Throughout the process, Qubic can always stick to its main AI training line, making the logic reasonable.

As the demand for AI computing power grows exponentially, and mining is no longer the only destination for computing power, the rules of the game have changed. The original cost of attacking the network was "purely burning money"; now there is an "extra sponsor" in AI training to foot the bill—attack costs are hedged by AI earnings.

This is my biggest concern with that article: AI demand is breaking the fundamental assumption of general CPU/GPU POW mining – "miners rely on mining rewards, so they will maintain the network." When computing power has a more profitable destination, this assumption no longer holds. Although this process will be slow, there is always the possibility.