Regarding the claim that the "Exponential Time Assumption (ETH) is wrong," I've come up with a perfect analogy to help you understand. Many people have heard the story of paracetamol. While this drug can reduce fever, unfortunately, humanity still doesn't know how it works. So people just assumed, "Okay, taking paracetamol = reducing fever," and used it haphazardly, since it worked anyway. And many scholars have conducted numerous experiments and published many papers on this. Then, OpenAI suddenly made a major discovery, stating that "taking paracetamol = reducing fever is wrong." Everyone: Holy crap?! It turns out that the AI found a very rigorous counter-evidence: if the body temperature exceeds 60°C, taking paracetamol won't reduce the fever! Everyone sees that the chemical equation is perfectly written, so what the AI said is probably correct. However, this counter-evidence doesn't change the status quo. You can still take this drug for normal fevers in everyday life, and it doesn't prove that the drug is bad. Furthermore, by the time your temperature reaches 60°C, you're already exhausted, so the effectiveness of the medication is largely irrelevant. This is a very small scenario. However, those most significantly affected are the scholars who previously published various papers. Their foundation was based on the idea that paracetamol equals fever reducer. Now, that foundation has been undermined, and most of them have to redo their work. The exponential time assumption (ETH) applied to the computer science field presents a similar situation. In our everyday lives, for example, the passwords we use are usually only 10 or 20 characters long, so there's absolutely no increased risk. The asymmetric encryption used by Ethereum (ETH) is even less affected; the two ETHs are completely unrelated except for their abbreviations. Unless you're researching highly specialized fields, like being a scholar or researcher—for example, working on graph theory, algorithms, chip circuit design, or DNA research—then previous academic papers might have to be scrapped and republished. For ordinary people, there's really no impact (for now). It might even have a greater financial impact; OpenAI's pre-market price is about to rise again