It's almost six months since ChatGPT was first released (or perhaps unleashed). On 30 November, 2022, OpenAI announced it would allow people to talk to its AI-powered chatbot.
In the time since, everything and nothing has changed. It has been an astonishing few months not only for the artificial intelligence industry, but also the predicting game: you have been able to say just about anything you want to about the future and what AI will do with it, and have people listen. Twitter has filled up with long threads about how everything was going to change, and snide remarks about how the technology industry had found itself overexcited about something all over again.
Even amid all that change and chatter, Google has stayed largely still and silent. Despite chief executive Sundar Pichai having re-oriented the company in recent years to be focused on AI, it did not immediately have anything of its own to offer in response to ChatGPT. It took months for Google to announce its own Bard, and it is still largely behind ChatGPT's performance.
And so Google has been stuck playing a big game of catch-up. It must be a frustrating one because Google has been working on AI for years, and has arguably integrated it in more responsible and meaningful ways than OpenAI did, but it matters very little.
Now it is facing another opportunity to switch that narrative again, in the form of its I/O developer conference, on Tuesday. (It tried to do this in February, when it announced Bard, in an event that was mostly notable because Bard was shown making a mistake about the James Webb Space Telescope.) The company will once again try and show that it is the AI company to watch, after all.
If you ask Google's Bard what it's going to announce tomorrow, the answers will mostly be about new hardware. There'll be a new phone, a watch, a tablet and updates to Android, it says. (It doesn't tell you where it got the information, so it's presumably searching for rumours, rather than having been given any insider information from its creators. What it takes longer to say is that Google is expected to "talk about its progress on artificial intelligence and machine learning".
That might be because the AI will actually be in the phone, and other hardware. Even now, amid the excitement about ChatGPT, we're still hardly seeing any actual use for it in terms of consumer products. This week, fellow phone maker Apple finally commented on AI, with Tim Cook saying that it was interesting but needed to be used carefully. (Like Google, Apple has faced serious questions about how it is using AI, having got a big headstart with Siri and then seemingly falling behind.)
One of the thrilling things about the recent shake-up in AI – though not if you're a Google shareholder – is just how much it has disrupted the technology industry. Months ago, OpenAI was seen as an interesting if largely academic enterprise; now it is all anyone seems to talk about. Microsoft went from capable and boring to being the future in a matter of months. Meta and Amazon are barely discussed anymore.
That can only be a good thing. I wrote in this newsletter a few months ago about the fact that technology seemed to be offering less and less to normal people, in terms of features or substantial breakthroughs, and the fact that people are now conditioned to react to new technologies like generative AI with skepticism and cynicism is a result of that. That slow progress is surely in part because the technology industry has become so concentrated, with all of the biggest companies trading off ideas and products that were created years ago. It's refreshing to have new things to talk about.
But a leaked report this week showed that the threat might not be from those other AI companies – but from open source researchers. The internal document suggested that public researchers might be getting ahead of not just Google but its competitors, too.
That would, of course, be the true shake-up: not just moving forward with artificial intelligence but moving on from the concentrated and closed tech industry and web that we have lived with for years. AI might not only be changing what you see on the internet and on your devices, but more fundamentally changing how they both actually work.
