Warning: A Botnet Is Likely to Take Down a Major AI Company, "Only a Matter of When"

One view warns that a botnet is likely to take down a major AI company, arguing it's only a matter of when rather than if.
The argument holds that CEOs won't harden infrastructure against a 1-in-100-year risk since they're incentivized on 3-5 year performance. Under this scenario, a neo-cloud would wake up unable to serve GPUs because a botnet has been silently using them, with the eventual fix being to unplug and plug back in, then call Anthropic or OpenAI for a solution.