OpenAI CEO accuses Anthropic of using "fear marketing" to promote Claude Mythos

CoinFeed reported on April 23 that, according to Decrypt, OpenAI CEO Sam Altman stated on a podcast that Anthropic is promoting its new model, Claude Mythos, through "fear-driven marketing" to argue that a small number of "trustworthy" institutions should control AI. Claude Mythos, due to its ability to automatically identify software vulnerabilities and perform complex cyberattack simulations, has been restricted to testing by a limited number of institutions. Some government departments and security research institutions have warned that while it possesses defensive potential, it also carries the risk of misuse. Altman stated that packaging models as "bombs" and selling them as "bunkers" can be an efficient business strategy, but it should not be used to monopolize the allocation of AI capabilities.