🎙️ Voice is AI-generated. Inconsistencies may occur.
news article
Based on facts, either observed and verified firsthand by the reporter, or reported and verified from knowledgeable sources.
AI having the ability to complete programming tasks suggests new ways for thinking about how our economy will work.
Photo description | The capabilities of AI-powered bots and agents may soon include engineering and coding.
Credit | Photo Illustration by Newsweek
OpenAI's o1 large language model (LLM) is performing well in the coding tests that are typically administered to candidates for engineering and product jobs at tech companies, according to recent research released by the company. This development portends a new paradigm for the labor economy.
"We're currently living in a transition phase right now with AI," Christopher DiCarlo, philosophy professor at the University of Toronto and author of the recently published book Building a God: The Ethics of Artificial Intelligence and the Race to Control It, told Newsweek.
The o1 LLM competed in the 2024 International Olympiad in informatics, placing in the 49th percentile. It also completed coding tests from HackerRank, a technical hiring software company. These advancements suggest AI tools and agents can support programmers or even provide nontechnical workers with programming capabilities. DiCarlo wondered what might happen if programmers, a highly paid and highly sought-after labor segment, become obsolete.
In HackerRank's Astra coding challenge, the probability of successfully completing a task on the first attempt by o1 was between 63 and 76 percent, demonstrating an ability to "address complex, industry-relevant software development tasks effectively." OpenAI concluded that it has reached a performance level that nearly matches "the best humans in the world."
DiCarlo added that we are rapidly approaching the point where, in many work situations, "humans no longer have anything to add. No human can help improve upon what AI does. We're not quite there yet."
During this transitional phase, AI can make people more productive, but it is certainly building toward a capability where it could replace programmers or empower all employees with programming ability, depending on how executives want to use this emerging technology.
"We're seeing how AI tech-facilitated development is going to help current systems and probably going to replace them, but right now it's helping us," DiCarlo said. Those making policy decisions will determine what the guardrails look like and what the cost to humanity would be.
From DiCarlo's perspective, people working in trades and in fields that require human interaction or physical manipulation or movement are safest. "We'll always need plumbers," he said.
Others see the loss of traditional corporate or knowledge-work jobs as an opportunity for people to find new ways of working, either via freelancing, new industries or the creator economy. Many new industries and job types exist today because of the expansion of leisure time and access to markets driven by technology advancements such as spreadsheets and the internet, like yoga instructor or social media manager.
"The pessimistic outlook on AI is 'They're going to take all our jobs,'" Steven Schwartz, co-founder and CEO of Whop, a creator-economy platform, told Newsweek. "I would say that they're taking the jobs that people don't actually like."
Schwartz, a 26-year-old CEO and serial entrepreneur, is optimistic that AI makes boring office jobs a thing of the past. His platform is working to enable that pathway and has already minted over 150 millionaires since its founding in 2021, he said.
In the progress of AI's capabilities, OpenAI's findings represent the arrival of a new frontier, in line with the earlier technologies that competed in checkers, chess and later the game show Jeopardy! to prove their ability to handle complex problems, DiCarlo said.
What's interesting is that this level of coding proficiency by AI has arrived quicker than most expected, he added. This means policymakers and business leaders may want to accelerate their projections for the arrival of a post-automation economy, powered by LLMs.
"Not only is it going to surpass all of our abilities soon, but how much more powerful is it going to get?" he said. "They are going to make it better and better, keep making it better, and nobody has a clue what's going to happen."
Businesses will also have to determine if their goals are purely to make profit, or if the impact of being a job creator and working with others to serve customers is meaningful to them. The idea of the one-person company, boosted by a full team of AI bots operating finance, marketing and sales, is one that DiCarlo said he sees as possible.
"The bottom line is when machine algorithms can do your work for you, if you only care about making money, then that's the path. If you care about community, if you care about what humans can bring to the table that AI tech currently cannot, then you keep humans in the loop, and you use a hybrid setting," he said.
With rapid change approaching, business leaders and policymakers must take heed of their responsibility to adjust and set guardrails to prevent disarray.
"The economic landscape is going to change big time in a very short period," DiCarlo said. "So what are we doing to prepare for that? How can we help people transition as the ability of AI goes up and the layoffs continue? How do we balance people's lives? Because we're talking about human beings here."
