Joshua Browder, a 26-year-old entrepreneur from the UK, recently supercharged his main product in a way that he could hardly have imagined a few years ago.
His startup DoNotPay had spent several years developing a chatbot that could negotiate erroneous or excessive fines and fees on behalf of consumers — think unwarranted parking tickets — building a database of expertise based on its history of human interactions.
The bot often needed manual intervention, but in December he had a breakthrough. The bot “talked” to Comcast’s online customer service and managed to save someone $120 on their broadband bill. He said it was the first time any such bill had been negotiated purely by AI.
How? Earlier that year Browder got access to GPT-3, a powerful, large language model created by the artificial intelligence company OpenAI. The system understands language better than almost any software before it, and sounds human when it responds. Browder is now planning an AI lawyer that can whisper instructions to people through their earphones when they’re in traffic court.
Browder’s startup, valued last year at $210 million, is one of a flurry of new services being hastily built on top of generative AI tools with names like GPT-3 and DALL-E. Other services promise to draft emails, spur new marketplaces or even replace Google search. They are coming at a time of broader changes to tech, where a combination of market and regulatory squeezes could make the industry more productive than it has been for years.
The business models of Big Tech, which until recently churned out more than $1 trillion in revenue every year, is coming under strain. The Google and Facebook advertising duopoly is shifting to a market where Amazon.com Inc. and potentially Apple are now major threats. A strict new antitrust law on the horizon from Europe is already forcing changes at Amazon and Apple to make life easier for their much-smaller competitors.
This confluence of circumstances might spark a familiar feeling for those who have worked in technology for a long time. The layoffs and share-price drops that marked the last, painful year have happened before, and they are typically followed by an upswing in fortunes. Boom and bust is part of tech’s history, and even amid the turmoil that has come to Elon Musk’s Twitter and the world of crypto, there is good reason to expect 2023 will mark the start of tech’s next boom, driven by AI and a more effective workforce.
For years, tech workers have had the upper hand in the industry’s labor market, commanding high salaries and expensive perks and stodging up big tech’s ranks to the point of bloat. Meta hired an astonishing 30,000 people during the pandemic, leading Mark Zuckerberg to cut 11,000 jobs in November. Stripe, Snap and Amazon made similar cutbacks recently, while Musk winnowed Twitter’s staff down to about 2,000 from 7,500 in less than six weeks. Some 150,000 tech workers lost jobs in 2022, according to Layoffs.fyi, which tracks industry cutbacks. Those left behind are being told to work harder and with “greater urgency,” or to come to the office more frequently than before.
The painful shake-up was necessary. For the last five to 10 years, the tech industry has offered precious few groundbreaking services as it grew fat on old business models. Our most important computing device is still a rectangular metal slab made by Apple or Samsung Electronics Co. Google is so terrified of disrupting its most important revenue source — advertising — that it has barely changed search, and Amazon’s AWS is still printing money as the world’s biggest cloud provider. Meta, at least, has tried a radical venture into virtual reality. But for the most part, the industry and its biggest players haven’t been very innovative.
They have also acted like a giant squids sucking up all the talent in the sector, to the detriment of startups. It was virtually impossible for a new company to compete for senior engineers when a payments giant like Stripe Inc. offered more than $450,000 a year for the position. Want a principal engineer to oversee a new product line? Too bad, because Facebook has paid close to $1 million a year for the role, according to Levels.fyi, a website that tracks engineer salaries in Silicon Valley.
Venture capital funding for low-margin tech startups — think firms that offer food delivery and telemedicine services instead of software — is declining after years of overindulging business ideas that should never have been funded. VC investors say they are now gravitating back to firms that build software and offer higher margins.
The combined effect: Tech startups that are cash-rich enough to survive two years or more without fundraising are positioned to scoop up the best engineers and product managers, as my Bloomberg Opinion colleague Tim Culpan recently pointed out. In other words, instead of talent being squandered on a broad array of businesses that won’t go anywhere, it’s heading to well-structured businesses and being put to good use.
Also read | Revenues dip, tech giants cut jobs
One other factor will help move things along: a massive government bounty. In the early 1990s, when the internet was still called an “information superhighway,” the US passed the High Performance Computing Act to help build out the country’s online infrastructure. It played a key role in stoking the web’s early growth. Some of its $600 million went to the University of Illinois, where a team of developers created the first graphical web browser known as Mosaic.
Now some tech firms are positioned to reap the benefits of a much larger, $52.7 billion investment into US semiconductor research after President Joe Biden signed the CHIPS and Science Act into law in August. That could directly benefit a range of AI services that rely on faster and more sophisticated computation from chips.
For once, tech is having its feet held to the fire. After years of unstoppable growth and cushy perks, it may be the only way for the industry to get back to innovating and create more space for others.