OpenAI’s recent debut of its GPT-4o large language model technology provided an inadvertent blueprint for many of us to follow in terms of how we conduct and present ourselves.
The company’s setup for their demo of GPT-4o appeared unassuming at first glance. The design of the stage was reminiscent of something from the Dick Cavett show in the 1970s. It also took place inside, where it seemed no more than 30 people or so sat in an audience, occasionally applauding. Mira Murati, Open AI’s chief technology officer, seemed uncertain as she addressed the crowd, and wasn’t quite sure where to look when addressing the audience. That being said, it all felt very human, something that you’re not necessarily expecting when talking about an issue like Artificial Intelligence.
Why did it work? For starters, it helps to have a paradigm-shifting AI technology that can affect almost every aspect of life, but I think that only partially explains why the company’s recent demonstration made such an impression.
OpenAI’s GPT-4o demonstration was streamed live, which isn’t a feat in and of itself, but the company decided to have its employees demonstrate GPT-4o’s features live as well, instead of just throwing to a pre-produced video demonstration, which would have been very easy to do.
Instead, Ms Murati introduced OpenAI’s Mark Chen and Barret Zoph, who were equipped with smartphones synced to a larger monitor, and they proceeded to showcase ChatGPT’s newest voice recognition and conversation technology. It wasn’t perfect, and the audio was patchy at times, but make no mistake, the impromptu conversation and back and forth interaction between man and machine was impressive. That portion of the live demonstration, along with dozens of other parts of the demo, were later clipped, shared and reshared by tech enthusiasts and media outlets, racking up millions of views along the way.
To execute this well, I think OpenAI studied carefully from another technology giant, Apple, but not necessarily the current incarnation of Apple. I am, of course, talking about the 1997 to 2011 version of the California-based consumer technology company. Those years provided countless examples of Steve Jobs, who returned to the company after being ousted in 1985, take to the stage in front of various packed auditoriums. The visual of one person under the spotlight, showcasing new operating systems, iMacs, iPods and ultimately the iPhone, stays burned in the minds of many.
The technology, however, I would argue, was only part of what made those product demonstrations work. In hindsight, watching Steve Jobs on stage, those appearances look less like keynote addresses and more like tightrope walks.
During the January 9, 2007 event, where Apple introduced the iPhone for the first time, Jobs even went as far as making a crank-call to Starbucks, where he proceeded to joke and order 4,000 lattes for the audience that was watching him. Mind you, this was on an unproven and untested iPhone device. It worked, there was ample laughter from the audience, and billions of iPhones later, the memory of that moment lives on.
Oddly enough, Apple in recent years has opted for a slightly different approach. Most of the company’s big announcements have been carefully executed with edited videos featuring a fleet of Apple employees demonstrating products and services. The pandemic kicked off the trend of these highly produced product demonstration videos, and I should point out that Apple is not alone in this approach, but I do think that technology companies grew very attached to this level of control even after the pandemic subsided.
Even Alphabet-owned Google, which had its own impressive debut of new AI products at its annual developer conference, opted for showing videos of the products. While chief executive Sundar Pichai was indeed live in front of an audience, right behind him was a screen where the pre-produced videos played.
Simply put, when the product demonstrations are executed with pre-produced videos, you don’t have to worry about potentially disastrous results.
What technology companies seem to be forgetting, however, is that we’re all human, and technical glitches happen. Within reason, such glitches are rarely fatal.
Back in 1998, when Microsoft’s then chief executive Bill Gates was trying to demonstrate the new plug and play features of Windows 98, the much-anticipated operating system crashed, eliciting laughter and some applause from the audience. Mr Gates and his Microsoft colleague on stage both laughed, and fast-forward to 2024, and Microsoft is still around, very relevant, and incredibly innovative.
In 2010, when the iPhone was beginning to crescendo and Apple was eager to showcase the latest features, the model that Jobs was demonstrating was having trouble maintaining an internet connection. Obviously, that’s really bad optics for a device that’s supposed to be connected all the time and make calls. Jobs, however, had a solution for the moment: he asked the audience to turn off their devices in the audience, so his device could connect. Again, there was laughter, and the moment was not cataclysmic, but rather, a curious footnote.
These examples, combined with OpenAI’s recent successful demonstration, teach a lesson from which all of us can learn something. While we often seek perfection, sometimes it’s the imperfections that really resonate. It’s even more important to embrace those human imperfections and uncertainty amid the onslaught of impersonal AI technology that shows no sign of slowing down.
In such a noisy world where new products, new ideas and new technologies come and go at the crest of a wave, live demonstrations showing real human vulnerability, matter more than ever, and they’re more memorable than pre-produced videos.
