Practitioner: Jev is overhyped, GPT-5.6 is the enterprise turning point
The author argues Jev only serves a small fraction of developers and that enterprise adoption is limited by people, not models, while GPT-5.6 can already solve 90% of problems at average US businesses, with Astra going one step further.
This is why the Jev hype never made sense to me. People who know how businesses actually run (mostly chaos), realize that the use cases aren't nearly as large as being sold. Jev is for builders, a fraction of the AI market. I encourage anyone who influences in AI to please, for the love of God, get out of our bubble, and spend a few weeks in a midmarket business trying to actually move the needle. AI adoption and success is a people problem, as much as it is a technology problem. This is the reason Microsoft, Google, Amazon, OpenAI, Anthropic and more, are spending more than $6 billion dollars to try and hire up FDEs. Jane in accounting, and Bob in operations don't give a damn if something is more efficient, they care that it works - and you only have one chance to win them over. They have day jobs, and don't have time to whittle down decisions, they can barely keep their heads above water as it is. GPT-5.6 was the turning point. There simply isn't any business problem I can't solve at the "average" American business (obviously there are edge cases, but this is 90% of business ready) with it. Astra just takes that up a notch further. Removing friction is the single most important aspect of success with AI adoption at the business scale. And I challenge anyone with real AI delivery experience in business to prove otherwise.