In my opinion, this is a red flag for anyone building applications that rely on GPT-4.
Building something that completely relies on something that you have zero control over, and needs that something to stay good or improve, has always been a shaky proposition at best.
I really don't understand how this is not obvious to everyone. Yet folks keep doing it, make themselves utterly reliant on whatever, and then act surprised when it inevitably goes to shit.
Learned that lesson... I work developing e-learning, and all of our stuff was built in Flash. Our development and delivery systems also relied heavily on Flash components cooperating with HTML and Javascript. It was a monumental undertaking when we had to convert everything to HTML5. When our system was first developed and implemented, we couldn't foresee the death of Flash, and as mobile devices became more ubiquitous, we never imagined anyone would want to take our training on those little bitty phone screens. Boy were we wrong. There was a time when I really wanted to tell Steve Jobs he could take his IOS and cram it up his cram-hole...
By this logic, no businesses should rely on the internet, roads, electricity, running water, GPS, or phones. It is short sighted building stuff on top of brand new untested tech, but everything was untested at one point. No one wants to get left behind in case it turns out to be the next internet where early adoption was crucial for your entire business to survive. It shouldn't be necessary for like, Costco to have to spin up their own LLM and become an AI company just to try out a better virtual support chat system, you know? But ya, they should be more diligent and get an SLA in place before widespread adoption of new tech for sure.
By this logic, no businesses should rely on the internet, roads, electricity, running water, GPS, or phones. It is short sighted building stuff on top of brand new untested tech, but everything was untested at one point.
Where's any logic here? You're directly comparing untested technology to reliable public utilities.
Many of those things you mentioned are open standards or have multiple providers that you can seamlessly substitute if the one you're currently depending on goes blooey.
To be fair, there's a difference between a tax-funded service or a common utility, and software built by a new company that's getting shoved into production way quicker than it probably shohld
This is for businesses of scale that have the ability to have multiple fallback vendors. AI will be the same eventually, we didn't have lots of utility alternatives to start with.