Robert Ta argues that Anthropic and OpenAI are using AI safety advocacy to boost their IPO valuations by lobbying the government to mandate their proprietary models. Open-source alternatives like Kimmy K3 and GLM can handle 80-95% of enterprise workloads at a fraction of the cost, making forced reliance on proprietary AI a threat to startups and American competitiveness.
If we hold the thought that Anthropic and OpenAI, they're going to be the biggest IPOs, hopefully in history from their view, right? Following SpaceX at their $2 trillion or so IPO recently. They're thinking, how do we maximize this IPO? And that means growth. And that means a certain amount of market dominance that they have to show and a narrative that makes a lot of sense. So what would their posture be? Their posture would be, hey, US government, make it so that people have to buy our stuff. That would be the easiest way to posture towards their short-term goal of their IPO doing well. I'd say at the detriment of like everybody's long-term and short-term goals to succeed as an enterprise in America. And then for America winning the AI race, I'd say that wouldn't work well, but from their perspective, I think that would be really great for their IPO. The other part of the narrative is, hey, this would be bad for American business, which is forcing everybody to use proprietary intelligence, especially when Kimmy K3 just came out and proved that you can get near frontier level intelligence for a fraction of the cost, then why wouldn't people move most 80, 95% of the workloads at their enterprise as an American business to that sort of open weight, open source model that some of these ones are proving like GLM and KIMI, and then save the super intelligence expensive tasks at the frontier, at the cutting edge for frontier intelligence like Anthropic and OpenAI. And then there's also the news that a bunch of startups and stuff signed a letter that's similar to this about how basically all startups are going to die if they were forced this way. Yeah, my take is this is a good thing. For all of us in the world, any AI builder, this is a good thing. The only bad that comes out of this is short-term Anthropic and OpenAI's IPO potential.
