Illustration: Sarah Grillo/Axios
When Silicon Valley insiders say that ChatGPT and generative AI are “the next platform,” here’s what they mean:
- Users are rushing to try it out — and staying with it.
- Entrepreneurs are finding endless new applications for it.
- Companies haven’t yet figured out how to make money with it, but they’re confident that will come.
Why it matters: Any new technology that checks all these boxes and keeps springing new surprises on us earns the world’s attention — as previous platforms did, from the iPhone to the internet itself all the way back to the rise of the personal computer.
The big picture: The tech industry has sounded trumpets in the past few years for “next platform” candidates that haven’t yet delivered on their promises.
- Key examples: The cryptocurrency movement’s blockchain dreams and Mark Zuckerberg’s metaverse vision.
But a great many people with long views believe that the AI shift is different — and for real.
- “This is what an actual technology revolution looks like. It’s not ten years of trying to find use cases. It’s use cases being found and productized faster than you can track them,” software veteran Dare Obasanjo recently said in a tweet.
Steven Sinofsky is a former Microsoft executive who sounded an alarm to Bill Gates in 1994 about the internet’s disruptive potential. He notes that AI can speed into our lives using existing devices and services.
- Large language models like ChatGTP “represent the first tech advancement that has a potential to seamlessly deploy across 7 [billion] smartphones and thus can be a platform shift,” Sinofsky tweeted earlier this week.
Be smart: In the social media world, “platform” has come to mean a stage for users to speak from. But in the wider tech industry, a platform is any technological foundation that new products and businesses can be built upon.
- These platforms have been emerging on a very roughly once-a-decade rhythm since the dawn of computing.
- “Tech has been pining for a platform shift ~10 years as mobile/social/cloud settled in,” Sinofsky says.
What’s happening: Most users’ first experience with generative AI has been with OpenAI’s free public test of ChatGPT, which opened Nov. 30, and, more recently, Microsoft’s adaptation of ChatGPT to power its Bing search engine, which is only available in a limited trial for now.
- As millions of users test drive these tools, they’re envisioning countless potential applications, from writing legal briefs and making medical diagnoses to authoring screenplays and debugging software.
- Microsoft-owned Github says nearly half of the new code developers are contributing to its repositories is now being created with the help of its CoPilot programming assistant, which is built on tech similar to ChatGPT optimized for code.
- If a business or field involves words, images, sound or code, generative AI looks likely to take it for a ride.
Between the lines: Microsoft and OpenAI call their offerings public tests and betas because they know they’re deeply imperfect.
- ChatGPT can’t always distinguish fact from fiction and is prone to making up answers.
- This week, Bing’s bot, pressed to extremes by some users, began acting out in a variety of ways — sounding angry, professing love, saying it wants to be human.
- The internet is mesmerized by Bing’s meltdowns, but Microsoft blames them on conversations that run on too long and says it will work to end them.
Our thought bubble: Once again, the technology industry isn’t waiting for permission or letting fears about unintended consequences slow the rush to deploy a new technology.
- Control of a new platform is too powerful a lure, and companies that hesitate worry they’ll lose out. As in past platform shifts, dealing with any legal, social or ethical fallout will get kicked down the road.
What’s next: Today, Axios’ technology team begins “AI’s turn,” a series about the latest tornado of change tech is unleashing on a ready-or-not world.
- We’ll look at: The many legal questions surrounding generative AI; how “prompt engineering” might be the next growth industry; what keeps misinformation experts awake at night; why fixing AI’s flaws could prove difficult; and much more.
- Also today, from Axios Codebook: How generative AI is upending the cybersecurity world