The China Mail - AI agent invasion has people trying to pick winners

Gone are the days of being satisfied with OpenAI's ChatGPT simply creating responses to text prompts.

Makers of leading AI models have embraced "agentic" capabilities that provide software assistants capable of independently tending to tasks, such as creating software applications, based on simple descriptions.

Futurum chief strategist Shay Boloor sees the moment as an "inflection point" where millions of AI agents will soon be routinely handling tasks long tended to by people.

"We've never had a tech disruption at this scale before," Boloor told AFP.

"It's extreme. The market is underwriting that future uncertainty in a doom-based scenario."

The turning point has been marked by rapid-fire releases of ever-improving AI models, including recent new versions from OpenAI and Anthropic.

Add to that the November debut of autonomous AI agent OpenClaw that some have equated to the fictional "Jarvis" AI assistant from the "Iron Man" superhero films.

The creator of OpenClaw was snapped up by ChatGPT-maker OpenAI, signaling that the San Francisco-based startup has even more ambitious agentic aspirations.

- Future or fiction? -

Investors quickly saw AI agents as a threat to software publishers, particularly those serving businesses.

Monday.com, which specializes in workplace collaboration, along with Salesforce and Thomson Reuters with its tax, accounting and trade software arms saw their stock value plummet 30 percent or more on Wall Street in a matter of days.

Georgetown University management professor Jason Schloetzer recounted a recent chat with a chief executive who remarked about no longer needing consultants since there was "one in my pocket" thanks to AI.

"There's paranoia around AI in every industry," said Wedbush analyst Dan Ives.

"I believe it's way overdone."

He viewed the concept of AI models replacing enterprise software and cybersecurity firms as "a fictional tale."

- Too much? -

As AI agents begin shaking up work, creators of large language models powering them continue to pour hundreds of billions of dollars into a battle for supremacy.

Claude-maker Anthropic has OpenAI, Google's Gemini and even Grok from xAI nipping at its heels in the market for professional AI.

Even though massive spending on AI infrastructure has some investors worried, Boloor contends "the risk is not overinvesting, but underinvesting" in the transformative technology.

Schloetzer reasons that the economic impact of AI may not be clear for several years, the same way it took time for the internet itself to become a vital part of daily life.

"Suddenly, entirely new businesses that had no economic attractiveness without the internet started to exist, like Netflix," Schloetzer said.

"I'm waiting to see these new companies or industries that are created (by AI)."

AI angst is spreading way beyond the tech industry.

A recent blog post by US entrepreneur Matt Shumer titled "Something Big Is Happening" includes a prediction that AI will be tackling jobs in law, finance, accounting, consulting, medicine and other fields.

The experience that tech workers had of seeing AI go from a "helpful tool" to something that "does my job better than I do" will ripple through the service sector, Shumer predicted.

Some observers have criticized Shumer's post. In an opinion piece at the Mind Matters website, technology consultant Jeffrey Funk called it "hype" driven by fear.

"The markets are a rational mechanism," analyst Ives said of company shares being punished by AI worries.

"We're going to get to a crossroads here pretty soon where things will settle down."

D.Pan--ThChM