Is "AI" This Year's "Metaverse"?
In which I add my voice to the many others scrambling to get their two cents in about this hot new AI chatbot thing.
What happened to the metaverse?
In late 2021 Facebook rebranded as Meta during a recorded event in which they tried convincing us that the metaverse was the future of our lives and that they were going to be at the center of it. Now, nearly a year and a half later, the word “metaverse” is little more than a meme. Meta’s own Horizon Worlds platform gets ridiculed any time it gets public attention and has been steadily losing users.
WSJ: Company Documents Show Meta’s Flagship Metaverse Falling Short (via archive.ph)
What happened? There are lots of problems that could be dug into, but I think a critical one is that they simply overpromised and underdelivered. Now I love my Quest VR headset and I use it several times a week, but I wouldn’t hesitate for a second to tell you that it falls a long way short of everything Meta has hyped. It might be comfortable enough and clear enough for gaming, but the standards for a productivity device are a lot higher, and a productivity device it is not. There’s plenty of fun games and experiences, but Meta’s much hyped metaverse platform isn’t one of them. It’s bland, soulless, and empty. Anyone who’s gotten into VR because of metaverse hype has every right to be disappointed.
Will AI follow?
So here we are today. ChatGPT has had an explosion of interest unlike anything before and companies are racing each other to be the first to capitalize on it. Notably, both Microsoft and Google are rolling AI chat bots into their search engines. Microsoft already has a public early access version available. Google so far has promises and mockups. But both companies’ AI language models are already showing warning signs of being rushed, underdeveloped, and terribly suited for their given tasks.
simonwillison.net: Bing: “I will not harm you unless you harm me first”
Reuters: Alphabet shares dive after Google AI chatbot Bard flubs answer in ad (via archive.ph)
Maybe it’s the circles I’m in, but I’m already seeing far more criticism than praise of these AI implementations. There’s even been criticism from within Google.
If the tech crowd and early adopters are already this critical, what’s going to happen when these features are rolled out to the wider public? Microsoft and Google are so far overpromising and underdelivering on the benefits of AI chat in search. They have yet to prove that their language models are even fit for this use case in the first place.
I’ve played around with ChatGPT and have gotten some good mileage with it as a sort of idea generator. This tech can already be very useful in a lot of cases, which is why it’s frustrating to see Microsoft and Google so far botch it. I wonder if they risk souring people on all things AI for years because of their mad dash for first. But who knows, depending on how much longer this tech needs in the oven maybe that’ll be a good thing.