I often ask ChatGPT to write a poem in the style of American modernist poet Hart Crane. It does a great job of delivering. But the other day, when I asked the software to perform a crane operation on a plate of ice cream sandwiches, I felt bored before I even looked at the answer. “The oozing cream slips from our hands like time / Each moment slips away with a silent gasp.” This is good. It was competent. I read the poem, shared some of it with a colleague, and closed the window. anything.
It's been a year and a half since generative AI captivated the public's imagination and my own. For months, the money I paid for ChatGPT and Midjourney felt like money better spent than my Netflix subscription, even for entertainment.I'll sit on the couch and make a cheeseburger monster meanwhile bridgerton It was played in front of me without being seen. But those days are over. The lethargy I felt as I sought out Hart Crane's paean to ice cream sandwiches seemed to mark the end of a brief, glorious phase in the history of technology. Generative AI came out of nowhere and brought with it both light and dark magic. Even if the show were to end, it would not be because the AI failed. Quite the opposite. The tools it enables simply slip into the background, from which they exert their greatest impact.
Looking back on my ChatGPT history, I've always wanted Heart Crane ice cream. Emily Dickinson's poem about sizzlers (“In the embrace of sizzlers, we find our space / Where simple pleasures and tastes intertwine”).edna st vincent millay Beverly Hills, 90210 (“In a sunlit land where palm trees sway/The stonewashed jeans of days gone by”). A poem about Biz Markie and Eazy-E's (real!) Snoop Dogg cereal Frosted Drizzlerz. An introduction to Rainbow Bright written in the style of philosopher Jacques Derrida. I requested these things initially to see what each model can do and to find out how it works. It turns out that AI has an uncanny ability to blend concepts accurately and creatively.
I wrote it last fall. of atlantic ocean At its best, generative AI can be used as a tool to stimulate the imagination. I used DALL-E to bring almost every concept that came to my mind into reality. One weekend, I spent most of our family outing stealing time to build his fictional 120-year history of a pear-flavored French soft drink called Poire. Then there was the trotter, a cigarette made by pigs for pigs. I've spent so many hours on these side hustle shenanigans that now the products feel real to me.They are teeth It is real, at least in the sense that Popeye, Harry Potter, and all the other fictions can be real.
But slowly, invisibly, the work of actually using AI took over. Lemon Lime While researching stories about the flavor, he asked ChatGPT to give him an overview of the U.S. market for beverages containing this ingredient, but to find out the facts, do your own research. There was a need. In the process of devising a new learning program for a university department, I let the software evaluate and come up with potential names. Neither task produced a fraction of the joy I once derived from his one and only AI-generated phrase: “Stonewashed jeans.” But at least the latter gave me what I needed at the time: viable mediocrity.
I still found some opportunities to stimulate my imagination, but they became less frequent over time. In their place, I assigned the AI the burden of being a mere mule. task. When faced with the problem of which waitlisted students to place in an oversubscribed computer science class, I used his ChatGPT to apply relevant and complex criteria. (As parents or my school principal may be reading this, I do not send students' real names or personal data to OpenAI.) I needed a website for my project on short notice, so I decided to use the service. I asked them to create a website much faster than I could. hand. When I wanted to analyze a complete corpus of Wordle solutions of recent articles, new york times For the game library, we turned to OpenAI's data analysts. No one had promised me anything like that, so it felt like a gift to have that effect.
More imaginative uses of AI will always fail under this practicality. A year ago, university professors like me were already worried about the practical implications of this technology, and whether and how universities could control the use of large-scale language models in assignments. We spent weeks discussing it. In fact, for students, generative AI seemed clearly and immediately productive. It could soon help you write a college essay or do your homework. (Teachers were finding many ways to use it, too.) Applications seemed to be multiplying. OpenAI CEO Sam Altman said in November that the ChatGPT service has 100 million weekly users. In January, job review website Glassdoor released a survey revealing that 62% of professionals, including 77% in marketing departments, use ChatGPT at work. And last month, the Pew Research Center reported that nearly half of American adults believe they interact with AI in some way at least several times a week.
The rapid adoption is partly due to the novelty of AI, and without initial interest nothing can become widespread. But that user growth can only be sustained by the technology moving toward something less exciting. Inventions become important not when they offer a glimpse of a possible future, as Apple Vision Pro is doing now, but when they can recede into the background and become mundane. Of course you have a smartphone. Of course, we also have refrigerators, televisions, microwaves, and cars. These technologies are not. I mean, they are. no longer-fun.
Not all inventions quickly lose their luster, but world-changing inventions don't take long to start looking mundane. I'm already nostalgic for the magical feeling I had when I created Hart Crane's new poem and the advertising campaign for Pear Soft Drinks. I miss the joy of seeing an idea you can imagine instantly come to life. But any nostalgia for the early days of ChatGPT and DALL-E will ultimately be short-lived. First the magic disappears, then the nostalgia. This is what happens with technology that is inherited. This is a measure of its power.