Cloud Hosting

But now the gravy train may be about to end in a way that not only shocks budding software engineers but also threatens the hefty salaries of mid-career professionals.

In February, Nvidia CEO Jensen Huang turned heads when he told the World Government Summit that over the course of the last 10, 15 years, almost everybody who sits on a stage like this would tell you it is vital that children learn computer science. Now, he said, its almost exactly the opposite. Because artificial intelligence systems can now help create software and theyll only get better.

Nvidia has benefited tremendously from demand for the hardware that fuels AI, sending its stock surging from about $130 a share in 2021 to nearly $900 today. And Huang has become Wall Streets go-to AI guru.

But is he right? With all the talk about AIs promise and peril, is there enormous upheaval coming to software jobs and the people and places that rely on them?

Yes, says Matt Welsh, a former professor of computer science at Harvard. A massive shift is on the horizon though he doesnt believe it will happen overnight.

Ive seen trends come and go, notes Welsh, who worked at Apple and Google before cofounding Fixie.ai. Ive seen technologies emerge over time. Ive never seen anything like AI, in terms of the pace of innovation and the dramatic effect its having on the whole industry. Not the introduction of personal computers. Not the iPhone. Not the internet. None of these things have had quite a dramatic effect in this [short] period of time.

He says that programming has long been a kind of priesthood, an exclusive club for the lucky few. But thats not going to last.

Soon, much as Huang has predicted, programming will become something that anyone can do. Because the language that youll need to program wont be Java or C++. Itll be English.

Late last year, Tesla CEO Elon Musk argued that AI could be the most disruptive force in history and that there will come a point where no job is needed.

Welsh believes that the disruption to the tech industry will come in two waves. First, an initial phase in which software engineers using ChatGPT or Microsofts Copilot become considerably more productive. And, for many, that phase is already underway.

Increased productivity may allow tech companies to reduce the number of engineers on the payroll while maintaining the same level of output. If you just look across the software industry, the vast majority of it is people doing things that are extremely mundane and easy to automate at some point in the not-too-distant future, he says.

But the second wave will likely bring much more disruption, Welsh argues. That will happen when we stop needing programmers to write these programs in the first place ... and that, I think, is not something that anyone in the industry is prepared for. This change, he believes, will start to unfold over the next several years.

In November, the job site ResumeBuilder polled 750 leaders at businesses that currently or plan to use AI in 2024 and discovered that more than a third say the technology replaced workers this year, and close to half believe that AI will lead to layoffs in 2024. Forty-two percent of workers, meanwhile, worry about how AI might affect their job, according to a survey by CNBC and SurveyMonkey.

When I challenged Welsh on the notion that an entire industry and millions of careers could be turned upside down, he countered with history. Disruption has come for industry after industry, he said. Highly skilled people once wove cloth. Now, for the most part, sophisticated machines weave cloth. And across the manufacturing sector, millions of jobs have been lost or drastically altered over the past century.

But what about the low quality of AI-generated code? A recent research paper argued that code written with an AI assistant tended to resemble the not-so-hot work of an itinerant contributor.

Welsh likens this to a story of burritos and lasagna. The invention of the microwave in American homes enabled people to plop a frozen burrito into the microwave and have a meal in two minutes, he notes. Was this meal as nutritious and good as the homemade lasagna that mom would have made in the oven for four hours? No. But was it sustenance and calories? Yes.

Still, its a huge topic of debate, and not everyone is convinced that the rise of AI will lead to a massive restructuring of the industry.

Jonathan Bell, an assistant professor of computer science at Northeastern, says hes certainly seen an AI-powered surge in engineers productivity. But he believes that if the stakes are high on the software youre building, youre going to want a human defining and refining the problem and, crucially, running quality control.

Bell, who studies how automation affects coders, is currently working on software to help facilitate Northeasterns doctoral admissions process: Could I ask some large language model thing to write that entire piece of software for me? Possibly. Would I want to trust that it was correct? Definitely not.

He argues that increased productivity may simply allow companies to get more ambitious and work on more projects. What weve generally seen in the past, he says, is that companies say, Wait! With the same amount of money, I could have software that does more stuff? And its more complicated, and delivers more value to our users? Lets do that ... Thats the optimistic view, and the cynical view is that you can lay off 30 percent of your staff.

Welsh maintains that we should gird ourselves for a major shift in the tech industry one that big companies know is coming: I just think that were actually now on the cusp of this, versus something that seemed completely fantastical even five or ten years ago. I think it may become reality much sooner than we anticipate.

Follow Kara Miller @karaemiller.

Follow this link:

Could AI coding threaten software jobs? - The Boston Globe