Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Generative AI took the world by storm with its release in November 2022 OpenAIs service ChatGPT. One hundred million people started using it overnight. Sam AltmanThe CEO of OpenAI, the company that created ChatGPT, became a household name. At least half a dozen companies have joined the OpenAI race to build a better system. OpenAI itself was trying to excel GPT-4its flagship model, It was introduced in March 2023with its successor likely to be called GPT-5. Almost every company struggled to find ways to apply ChatGPT (or similar technology developed by other companies) to their business.
There’s just one thing: Generative AI doesn’t really work that well, and maybe it never will.
Basically, the engine of generative AI is filling in the blanks, or what I like to call “auto-completion on steroids”. Such systems are great at predicting what might sound good or believable in a given context, but not at a deeper level of understanding what they’re saying; artificial intelligence is constitutionally incapable of auditing its own work. This has led to major problems with “hallucination,” where the system asserts things that are untrue without any basis, while introducing bone-headed errors into everything from arithmetic to science. As the saying goes in the army: “oft wrong, never doubt”.
Systems that are often flawed and never questionable make incredible demos, but are often bad products themselves. If 2023 was the year of AI hype, 2024 is the year of AI disappointment. Something I argued in August 2023 is, to my initial skepticism, more often felt: generative AI can become stupid. Profit is not there –offers estimates OpenAI’s 2024 operating loss could be $5 billion, and a valuation of more than $80 billion doesn’t match the lack of profits. Meanwhile, many customers seem disappointed with what they can actually do with ChatGPT compared to the unusually high initial expectations that have become commonplace.
Also, essentially every major company works from the same recipe, making bigger and bigger language models, but all tied in more or less the same place, which are about as good models as GPT-4, but no problem. much better. What this means is that no individual company has a “moat” (the ability of a business to defend its product over time), which in turn means reduced profits. OpenAI has already been forced to cut prices; Now Meta is giving away similar technology for free.
As I write this, OpenAI is showing off new products but not actually releasing them. By the end of 2025, if there is no progress worthy of the name GPT-5, which is decidedly better than what its competitors can offer, the bloom will fall from the flower. Enthusiasm supporting OpenAI will wane, and since it’s the poster child for the entire field, it could all soon collapse.
