Microsoft Throws a Coming-Out Party for A.I.
It’s been a rough few months for the tech industry. There have been tens of thousands of layoffs, hundreds of billions in value lost on Wall Street and a high-profile scandal at a crypto company that has shaken faith in that young market.
But in a conference center on Microsoft’s sprawling campus, Tuesday was a moment for swagger. Executives and engineers from Microsoft and a small research lab partner called OpenAI unveiled a new internet search engine and web browser that use the next iteration of artificial intelligence technology that many in the industry believe could be a key to its future.
This new artificial intelligence became a fascination for millions of people two months ago when OpenAI released a chatbot called ChatGPT. Capable of answering questions, writing poetry and riffing on almost any topic tossed its way, ChatGPT provided the tech industry with a jolt of excitement in the middle of its biggest job contraction in at least 15 years.
The enthusiasm around OpenAI’s technology — as well as the work of several competitors expected to hit the market soon — reminds tech veterans of other moments that have turned Silicon Valley on its head, from the first iPhone and the Google search engine to the Netscape web browser that set the stage for the commercialization of the internet.
Microsoft played catch up on browsers and badly missed the shift to mobile computing that came with the iPhone; its Bing search engine is a distant second in popularity to Google. But it could be the first big company to tech’s next big thing if the chatbots and their technology, called generative A.I., lives up to its billing.
“This technology will reshape pretty much every software category that we know,” said Satya Nadella, Microsoft’s chief executive. He added that “a race starts today in terms of what you can expect.”
On Tuesday, in a room crowded with nearly a hundred reporters, editors and photographers, Microsoft showed off a new Bing search engine. Yusuf Mehdi, Microsoft’s corporate vice president, used a new conversational interface to look for a 65-inch television suited to video games. As the service listed televisions, he asked it to pare the list to the cheapest models. It quickly did.
The Rise of OpenAI
The San Francisco company is one of the world’s most ambitious artificial intelligence labs. Here’s a look at some recent developments.
- ChatGPT: The popularity of the cutting-edge chatbot that kicked off an A.I. arms race has come as something of a shock — even among employees of the company that created it.
- DALL-E 2: The system lets you create digital images simply by describing what you want to see. But for some, image generators are worrisome.
- GPT-3: With mind-boggling fluency, the natural-language system can write, argue and code. The implications for the future could be profound.
He then used the chatbot to plan a Mexican vacation and research Japanese poets. With a short query, he could ask the system to translate results from Spanish to English or show a particular haiku poem.
“You see, this is just so much better than today’s search,” Mr. Mehdi said.
Mr. Mehdi also unveiled a new version of the company’s Edge web browser that offers its own chatbot service. After loading a news release, he asked the bot to summarize the document. He also asked it to generate a tweet about the new Bing search engine and had it generate a snippet of computer code for a new software program.
Microsoft released its new version of Bing to a limited number of people on Tuesday. Each user will be able to run a limited number of queries and people can join a wait list for access to the full version of the service. The company plans to expand to millions more people by the end of the month.
Other companies are jumping into the chatbot race. On Monday, Google announced that it would soon offer a chatbot called Bard and start adding chatbot technology into its own search engine. Meta, Facebook’s parent company, is fast-tracking efforts to release similar technology in various products. And countless start-ups are building their own generative A.I. products, the name for technologies that generate words, images and other media on their own.
Executives, entrepreneurs and investors hope the chatbots will not turn out to be what the tech industry has seemed to churn out for some time now: a curiosity that falls short of big expectations.
There have been many: Self-driving cars that can’t quite get the self-driving part right. Wearable technologies that still need a smartphone nearby to truly be useful. And crypto currencies that promised to change the world of finance but so far have largely been an asset for speculators.
Microsoft has worked closely with OpenAI, invested $13 billion in the start-up and supplied the billions of dollars in computing power needed to build its A.I. technology. Microsoft declined to discuss the specific technology that underpins its new search engine, but it is likely based on a widely rumored OpenAI creation called GPT-4, the successor to what the San Francisco company released two months ago.
“Satya played his hand beautifully,” said Andrew Ng, a researcher and entrepreneur who previously oversaw the A.I. labs at both Google and Chinese giant Baidu.
Like similar services from start-ups like Perplexity and You.com, Microsoft’s new search engine annotates what the chatbot says, so people can readily review its sources. And it dovetails with Microsoft’s index of all websites, so that it can instantly access the latest information posted to the internet. The company also said that its search engine includes technology that is designed to identify and remove problematic content from the chat service.
Last week, Microsoft released its first A.I. integration into Outlook, its email service, with a tool that helps salespeople write custom emails. In the coming months, Microsoft plans to release features with generative A.I. on average every week, said Charles Lamanna, an executive who oversees the applications Microsoft builds for businesses.
He compared this new wave of A.I. technologies to the rise of the internet or personal computing. “Everybody is in a room with the lights out trying to feel what the heck this market and this opportunity actually looks like,” he said in an interview last week.
But it is unclear how much of an appetite businesses will have for these services because technologies like ChatGPT are far more expensive to operate than traditional software.
“The economics of software probably will have to change,” Mr. Lamanna said. “Software may be a little bit more expensive, but it will do some pretty amazing stuff.”
The new chatbots do come with baggage. They often do not distinguish between fact and fiction. They can generate language that is biased against women and people of color. And experts worry that people will use them to spread lies at a speed they could not in the past.
“Companies often put these technologies out too quickly, disregarding their flaws and then trying to fix them on the fly,” said Chirag Shah, a University of Washington professor who explores the flaws in chatbots. “This can cause real harm.”
Google and Meta had been reluctant to widely deploy generative A.I. because its flaws could damage their reputations. But OpenAI — a new company with no real brand to protect — was willing to push the envelope.
In a 2,000-word blog post published ahead of the press event, Brad Smith, Microsoft’s president, called this a “watershed year” and acknowledged the potential downsides, calling for “wide-ranging and deep conversations” on the issues.
Sam Altman, the chief executive of OpenAI, made the case that it was time to release generative A.I. to a mass audience. “We are eager to continue learning from real-world use,” he said. “You’ve got to do that in the real world, not in the lab.”
This is a developing story. Check back for updates.