AI’s Slowdown Is Everyone Else’s Opportunity

The multi-trillion-dollar artificial intelligence boom was built on certainty that generative models would keep getting exponentially better. Spoiler alert: they aren’t. 

In simple terms, “scaling laws” said that if you threw more data and computing power at an AI model, its capabilities would continuously grow. But a recent flurry of press reports suggests that’s no longer the case, and AI’s leading developers are finding their models aren’t improving as dramatically as they used to.

OpenAI’s Orion isn’t that much better at coding than the company’s last flagship model, GPT-4, according to Bloomberg News, while Alphabet Inc.’s Google is seeing only incremental improvements to its Gemini software. Anthropic, a major rival to both companies, has fallen behind on the release of its long-awaited Claude model.

Executives at OpenAI, Anthropic and Google all told me without hesitation in recent months that AI development was not plateauing. But they would say that. The truth is that long-held fears of diminishing returns for generative AI, predicted even by Bill Gates, are becoming real. Ilya Sutskever, an AI icon who popularized the bigger-is-better approach to building large language models, recently told Reuters that it had leveled off. “The 2010s were the age of scaling,” he said. “Now we're back in the age of wonder and discovery once again.” 

“Wonder and discovery” puts quite the positive spin on “we have no idea what to do next.” It could also, understandably, spark anxiety attacks for investors and businesses, who are expected to spend $1 trillion on the infrastructure needed to deliver on AI’s promise to transform everything. Wall Street banks, hedge funds and private equity firms are spending billions on funding the buildout of vast data centers, according to a recent Bloomberg News investigation.

Does this all add up to a terrible gamble? Not exactly.

There’s no question that the main beneficiaries of the AI boom have been the world’s largest tech companies. Quarterly cloud-storage revenue for Microsoft Corp., Google and Amazon Inc. has been growing at a steady clip and their market capitalizations, along with those of Nvidia Corp., Apple Inc. and Meta Platforms Inc., have soared by $8 trillion in aggregate over the last two years. Returns on investment for everyone else — their customers — are taking longer to show up.

Yet a break in the market hype around AI could be useful, just as it’s been for previous innovations. That’s because technology typically doesn’t hit a brick wall and die, but goes through an S-curve. The idea of the S-curve is that initial progress takes years before rapidly accelerating, as we’ve seen over the last two years with generative AI, before it starts to slow again and, crucially, evolve.

Critics over the years, for instance, regularly declared Moore’s Law dead just before a manufacturing breakthrough for chips pushed it forward again. The development of airplanes progressed at a glacial pace until the transition from propellers to jets in the late 1950’s led to a leap forward — before the technology seemed to plateau. But just like chip manufacturing, aviation’s development didn’t stall, but transform; passenger planes have become far more fuel efficient, safer and cheaper to operate, even if they’re only nominally faster than they were in the 1960s. 

A similar plateau for AI and its scaling laws might also mean a new approach to development and measuring success, which until now has focused too much on capability, and not enough on other areas such as safety. Some of the most advanced generative AI models fall short on critical areas like security and fairness, according to a recent academic study that measured how well they followed Europe’s upcoming AI law.

For much of this year already, AI researchers have been looking at new paths for improving their models that don’t just involve throwing more data and computing power at them. One approach is to focus on enhancing a model after it has been trained, in the so-called inference phase. This can involve giving a model extra time to process multiple possibilities before settling on an answer, and it’s why OpenAI described its most recent model, o1, as being better at “reasoning.”

The beauty of the S-curve is that it can give everyone else some breathing room, instead of clamoring for the latest tech that will give them an edge over their competitors. Companies that have been experimenting with generative AI, and grappling with ways to boost their productivity, now have some time to redesign their workflows and business processes to better capitalize on current AI models, which are already powerful. (Remember it took years for businesses to reorganize themselves around computers in the 1980s.)

Stanford University professor Erik Brynjolfsson’s writing on the “productivity paradox” points out that output often appears to stall or drop when major new technologies arrive, before surging. A pause for AI gives businesses more space in that all-important investment phase.

It also gives regulators time to design more effective guardrails. The European Union’s AI Act, which companies will be subject to from 2026, needs to be more specific in how it defines harms. As standards bodies do that work, it helps that newfangled models leading to a batch of unexpected problems aren’t about to flood the market.

Generative AI has been on a bullet train during the past two years, and the momentum has clearly been lucrative for tech giants. A slowdown at the station offers a much-needed break for everyone else.