Because there is a lot of money tied up in AI now, in a way that doesn't just reek like a bubble waiting to implode but even more stinks like a bunch of what used to be called "wash trading" [1]. And that's just the money side.
The "social kool-aid" side is even worse. A lot of very rich and very influential people have bet their career on AI - especially large companies who just outright fired staff to be replaced both by actual AI and "Actually Indians" [2] and are now putting insane pressure on their underlings and vendors to make something that at least looks on the surface like the promised AI dreams of getting rid of humans.
Both in combination explains why there is so much half-baked barely tested garbage (or to use the term du jour: slop) being pushed out and force fed to end users, despite clearly not being ready for prime time. And on top of that, the Pareto principle also works for AI - most of what's being pushed is now "good enough" for 80%, and everyone is trying to claim and sell that the missing 20% (that would require a lot of work and probably a fundamentally new architecture other than RNG-based LLMs) don't matter.
The "social kool-aid" side is even worse. A lot of very rich and very influential people have bet their career on AI - especially large companies who just outright fired staff to be replaced both by actual AI and "Actually Indians" [2] and are now putting insane pressure on their underlings and vendors to make something that at least looks on the surface like the promised AI dreams of getting rid of humans.
Both in combination explains why there is so much half-baked barely tested garbage (or to use the term du jour: slop) being pushed out and force fed to end users, despite clearly not being ready for prime time. And on top of that, the Pareto principle also works for AI - most of what's being pushed is now "good enough" for 80%, and everyone is trying to claim and sell that the missing 20% (that would require a lot of work and probably a fundamentally new architecture other than RNG-based LLMs) don't matter.
[1] https://www.bbc.com/news/articles/cz69qy760weo
[2] https://www.osnews.com/story/142488/ai-coding-chatbot-funded...