Good morning. I’m Joe Castaldo, a Globe business reporter who focuses on artificial intelligence. I’ll be subbing in for Chris this morning, and taking a look at the world’s wealthiest company as it reports earnings after the close today. Nvidia Corp.’s stock has played a large part in propelling the year’s broader market gains – but its future might not be so straightforward. Hope you aren’t afraid of heights. More on that, but first:

The federal government is aligning itself with the incoming Trump administration as it leaves open the possibility of a new deal that would exclude Mexico. This coincides with the announcement that Howard Lutnick would serve as U.S. commerce secretary, in charge of his “tariff and trade agenda.”

Canada Post’s growing part-time, temporary workforce has become a stumbling block in negotiations for a new collective agreement.

TC Energy Corp. says a $1-billion deal touted as Canada’s largest equity purchase by an Indigenous partnership may not be completed as planned.

  • Earnings include Metro Inc., Nvidia, and cloud-based data storage giant Snowflake Inc.
  • Be prepared: Alert Ready is testing out its emergency alerting system across Canada today. Here are the times in each province.

Nvidia chief executive Jensen Huang at an October event in Denmark. Mads Claus Rasmussen/Reuters

Ilya Sutskever, the co-founder of OpenAI, has touched off a rollicking debate in the artificial intelligence world.

And this time he didn’t even fire Sam Altman.

Instead, he said this: “The 2010s were the age of scaling, now we’re back in the age of wonder and discovery once again. Everyone is looking for the next big thing.”

Why is that noteworthy?

Sutskever, who recently co-founded a company called Safe Superintelligence, is a machine-learning guru who has played a pivotal role in the generative AI era in which we now live, for better or worse. One of his chief contributions has been the notion of scaling: if you throw lots of data into an AI model and fire up lots of graphics processing units (GPUs), you’ll get better results.

And he’s been right. Scale is a big reason why OpenAI’s GPT-4 appears to be so much smarter and more capable than previous releases.

But now Sutskever is saying the gains from scaling are starting to plateau, raising questions about the future of AI development and everyone’s favourite GPU supplier, Nvidia, which reports earnings today.

His comments come as Reuters, Bloomberg and The Information have reported that developers including OpenAI, Google and Anthropic are having trouble building more advanced AI models. One question they’re grappling with: is it worth it to spend millions of dollars on more GPUs just to eke out some meagre performance improvements?

Nvidia Corp’s. earnings marks potentially the most important day remaining on the market calendar for 2024. The Globe and Mail

Scale is king

The drive for scale has pushed Nvidia’s stock price into the stratosphere and prompted other companies to hoard GPUs in the belief the world will need endless computing power for these huge AI models.

David Cahn, a partner at Sequoia Capital, has called this idea a “delusion,” and questioned where the revenue is going to come from to justify the expense. Similarly, Goldman Sachs said in a report this month that everyone is searching for AI applications that will actually make money, noting investor concern about “capital incineration for early investments.”

Off the wall

Worries that scale has diminishing returns have been prevalent for a while, in truth, but it could still be too early to draw conclusions. Some in the AI field have been arguing that any concerns about progress hitting a wall are very premature. Altman himself tweeted cryptically in his all-lowercase-style: “there is no wall.”

Others, such as Meta’s chief AI scientist Yann LeCun, have taken a victory lap. “I don’t wanna say, ‘I told you so,’ but I told you so,” he wrote on Threads in response to Sutskever’s comments. He’s long argued that entirely new approaches are crucial to building more advanced AI systems.

So what should you take away?

In your humble correspondent’s opinion, this is a reminder of just how new and uncertain the entire generative AI landscape is, and that technological progress rarely moves in a straight line.

If Nvidia’s stock price continues to reach new heights, it’s something investors may want to keep in mind as well.