This wasn't the case 20 years ago. There was actually a good article on HN a few months ago showing how most of the brands are owned by one company that has ruined them all. Still a few good brands. Need to research https://news.ycombinator.com/item?id=48147665
I get what you're saying, but I don't see the issue here. 95% of people don't need latest Claude Opus or Fable for their work. Most people are not software engineers. Having a model that excels at other things and is faster, cheaper, accessible directly via social media, and "good enough" is a viable pathway. AI aside, when was the last time your company provided you with the "best" tool? Microsoft has made being third best in the desktop OS and cloud provider markets a highly profitable art form. I think it's too early to pick winners in AI right now.
And as others said here, xAI is also probably throwing money into AI and hoping for a breakthrough. Except in this case it's a rocket company, social media company, cloud compute provider, and satellite ISP all rolled into one that can not only bankroll the development and perform all kinds of crazy accounting shell games but can potentially benefit from any breakthroughs in other lines of business. If those Google and anthropic compute contracts hold, a lot of investment is recouped.
Maybe I'm desensitized from the launching of the Tesla Roadster into space, "bulletproof" cyber truck, and the boring company flamethrower, but this doesn't seem too wild to me.
The most ridiculous thing about cannabis remaining schedule 1 (high risk of abuse, no known medical use) so long is that there is both prescription THC and CBD. It should have been changed|to a higher (less strict) schedule decades ago with the release of marinol in 1985. Then we could have had a lot more studies being done. Needless to say, not easy to do studies on schedule 1 drugs.