You have to wonder whether Sam Altman ever sleeps. In the past four weeks, the billionaire chief executive of OpenAI — and father to a newborn baby — has signed chips and data-centre deals worth more than $500 billion with Oracle, Nvidia and Advanced Micro Devices (AMD).
He also rolled out a new social media app, Sora 2, that allows people (in America and Canada) to make AIgenerated videos of themselves, a shot across the bow of TikTok. Meanwhile, a new “Instant Checkout” feature, launched at the end of September, allows users to buy products from five million shops managed by ecommerce giant Shopify — without ever leaving OpenAI’s ChatGPT app.
Days after this, it unveiled a capability to integrate other apps, allowing users to create a Spotify playlist or search for homes with Zillow (the US equivalent of Rightmove) from inside ChatGPT. It even rivals YouTube by serving up videos in its search results.
While some of the features, announced at OpenAI’s demo day in San Francisco last week, roll out gradually, the plan is coming into focus: to collapse the chaos of the web into a single interface.
Marc Andreessen, the billionaire tech investor, wrote in 2011 that “software is eating the world”. Altman appears to be trying to eat the internet, by turning ChatGPT into the “everything app” to rule them all — a tool that people can turn to as their personal assistant, travel agent, confidant, legal adviser, doctor and personal shopper. “Most people will want to have one AI service, and that needs to be useful to them across their whole life,” he said last week to Ben Thompson of the tech site Stratechery. “I do feel like this is a once-in-a-lifetime opportunity for all of us and we’ll take the run at it.”
Thompson said Altman is apeing Bill Gates by trying to turn ChatGPT into the dominant “operating system” of this new age. “OpenAI is making a play to be the Windows of AI,” he said.
Under that vision, “agents” — autonomous applications — would spring up by the billions, doing in seconds what would take humans hours or days, and for a fraction of the cost. Everyone would be endowed, in effect, with a personal AI workforce marshalled by their preferred chatbot, diverting money from the humans whom you might once have paid for that work.
“If you have a couple of entities that actually crack artificial general intelligence, these will make more profit than anything we’ve ever seen,” said Alex Blania, chief executive of World, the digital identity start-up, in 2023. “You’re talking about a significant percentage of global GDP.”
That belief is why the AI bubble inflated so quickly. It is why Thinking Machines Lab — a revenue-free start-up, launched by OpenAI cofounder Mira Murati, that launched its first product just days ago — can be valued at $12 billion.
Ruchir Sharma, at the equity research firm Rockefeller International, estimated that 40 per cent of America’s GDP growth this year was due to AI spending.
For Altman, the tip of the spear is forging the dominant AI super app. “We’ve gone from people’s entire world [being] social media, which was pretty awful but at least there was a chance of diversity,” said Hany Farid, a professor at UC Berkeley School of Information.
“The thing about these large language models is that there are going to be a relatively small number of winners, because you need massive computing, data and infrastructure. Your entire online existence is going to be funnelled through four or five billionaires. I worry about that consolidation of power. I think it’s dangerous.”
What makes this moment different is the nature of AI itself. It is a conversational engine inviting intimacy and trust; that one of the uses of these systems is therapy shows how different the technology is from what has come before.
There is, of course, a cost. Disturbing anecdotal evidence is mounting that chatbots are assuming outsized roles in people’s lives, with sometimes devastating results. The parents of 16-year-old Adam Raine sued OpenAI in August, alleging that the chatbot pushed him toward suicide. Meanwhile graduates are finding it harder to land jobs because companies are using bots to handle rote tasks.
Yet the prize is so vast that Altman and his ilk are ploughing ahead regardless. “We’re basically doing the same thing we did with social media,” Farid said. “Early on, there were signs that something was not right, but we kept moving fast and breaking things.”
OpenAI has argued that safety is central to everything it does, and Altman appears undeterred. He is eyeing something more fundamental than crafting the AI age’s operating system. “My favourite historical analogy [for AI] is the transistor,” he said. “I think it will just kind of seep everywhere into every consumer product and every enterprise product.”
No comments:
Post a Comment