AI doomerism is definitely out of control. I don't see AI "killing us all," but I definitely see many jobs—and maybe even most white-collar jobs—becoming irrelevant within the next five years. Models aren't conscious or exercising free will, but they can enable bad actors to hack critical systems humans rely on, similar to a terrorist gaining access to nuclear material. More concerning to me is how our government, irrespective of party, is failing to address critical concerns regarding imminent labor displacement and potential safety risks. These issues include: - Failure to ban TikTok, which has allowed China to weaponize the issue of AI, turning Americans against what is likely the only way to re-educate and empower the next generation of workers. This comes even after TikTok has allegedly been used to push pro-Chinese messaging to Americans for upwards of 10 years, with little action. - Failure to protect the intellectual property of American AI labs, both open- and closed-source, from Chinese hacks and distillation attacks. - Failure to enforce laws already on the books that protect Americans from companies that release faulty, defective, or harmful products (per @linamkhan). - And now, the intent to violate the Sherman Act by labs that want to collectively "slow the frontier." This seems to have a hidden agenda: shutting down open-source progress by imposing significant cost barriers on anyone who wants to build in AI but doesn't have substantial capital (as @davidsacks wrote). Thankfully, we have leaders like @finkd who have rejected this effort by frontier labs to "cartelize" AI. In many ways, post-pandemic America has parallels to post-WWII America, when a generation tarnished by trauma and years away from the workforce led to a series of policies designed to help soldiers reintegrate into the economy, including the GI Bill. If we want to prepare for AI while American workforce participation is at a 40-plus-year low, we really need to do the following: - Create programs that leverage AI to educate people on how to build businesses or explore new forms of work that AI agents can perform. I fully expect everyone from an Ivy League-educated investment banking analyst to an executive assistant at a small business to be replaced by AI agents. But there are myriad opportunities to create new products, businesses, and forms of work. Even @bot @grok is hosting a conference this week showcasing "how to build a new business with AI." Even if government fails again, the private sector can focus on education and dedicate a significant amount of capital to such programs. - We cannot let fear guide us into the abyss again. Fear contributed to the deaths of thousands of Americans and perhaps $20T+ in national debt after 9/11. We fought prolonged wars because we were afraid, and we were duped into believing they were necessary to keep us safe. In reality, they also became a means for bad actors to extract money from American taxpayers. I believe this effort to "slow the frontier" is yet again a way for frontier labs to absolve themselves of guilt if an AI system causes harm while creating unjust barriers to competition. Instead of fearing AI, we should do what we can to understand and harness it. We already have incredible companies like $NET with the capacity and capability to stop "agent swarms." For every risk AI creates, there will be a business or product designed to counter that risk, just as there has been in every other technological transition in human history. What we cannot fall for is allowing fear to enable more regulatory capture. That's what frontier labs want, because every day that open-source or competitive models improve, the intelligence advantage currently held by the frontier labs erodes. Losing this advantage may be existential to their oligopoly, but it isn't existential for the rest of us.