in spring In 2007, I was one of four journalists appointed by Steve Jobs to review the iPhone. This was probably the most anticipated product in the history of technology. What is it like? Was it a tipping point for the device? Looking back at my review today, I am relieved to say that this is nothing to be ashamed of. I recognized the generational significance of this device. But while I gave him this much admiration for his iPhone, it’s the volcanic amalgamation of hardware, operating system, and apps, and its surprising side effects, such as its hypnotic effect on our attention. I couldn’t have expected it. (I urged Apple to “encourage outside developers to create new uses” for the device.) And the rise of services like Uber and TikTok, and the rise of family dinners on the display. Nor did I say that we should anticipate a transition to a central communal trance. . My main job, of course, was to help people decide whether to buy this damn thing for $500, which was very expensive for a cell phone back then. But reading this review now, some may wonder why I spent hours frustrated by AT&T’s network and his web browser’s inability to handle his Flash content. It’s just like whining about which sandals to wear when a three-story tsunami is about to collapse.
Reading about what people are experiencing with modern AI apps, such as large language model chatbots and AI image generators, reminds me of my lack of foresight. Quite naturally, people are preoccupied with the effects of a sudden swarm of shockingly capable AI systems, but scientists believe these seemingly rapid advances will take decades. I often point out that it took. But, like when I got my first iPhone in 2007, if you focus too much on the current versions of products like Microsoft’s Bing Chat, OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Bard, you’ll find AI built in. There is a danger that future potential trajectories will become unpredictable.
This fallacy is clearly observable in what has become a popular new media genre best described as immediacy. The modus operandi is to attempt some task previously confined to humans and then take it to extremes, often ignoring the warnings provided by the inventors. The great sports journalist Redd Smith once said that writing a column is easy, just open your veins and let the blood flow. But would-be critics are now pushing the bloodless version, where all you have to do is open your browser and get a prompt. (Note: This newsletter was created the old-fashioned way of opening a vein.)
Prompt-and-pronunciation columns typically involve using one of these fairly early systems to see how well it replaces what was previously confined to the human realm. . A typical example is new york times The reporter used ChatGPT to answer all of her work contacts for the week. wall street journalproduct reviewer used AI to test her voice (hey, we first I did it!) and decided to replicate the look. There are dozens of similar examples.
In general, those who direct such stunts come to two conclusions. One is that while these models are great, they fall well short of what humans are best at. Email doesn’t convey the nuances of the workplace. The clone is dragging one leg into the uncanny valley. Worst of all, these texts and his generators will hoax when asked for factual information, a phenomenon known as “hallucinations” that currently plagues AI. And it is an indisputable fact that the output of today’s models often has a soulless quality.
In a way it’s scary. Will our future world be run by flawed “mind children,” as roboticist Hans Moravec calls them digital successors? But in another sense, its shortcomings are comforting. Indeed, the AI can now perform many low-level tasks, and is unmatched in suggesting a plausible trip to Disneyland or a gluten-free dinner menu for his party. . But conceivably, bots will always need us to spice up the event. prose.
