BEREA, Ky. — If you want a snapshot of how fast “agentic AI” is moving from demo to disruption, look at what happened this week: Anthropic shipped new Claude “Cowork” plug-ins aimed at legal, sales, marketing, and data work—and markets reacted immediately.
In U.S. trading, Thomson Reuters fell nearly 18% in a single session, while the selloff spread globally into software and services names.
In India, where much of the IT sector is built around large, staffing-intensive delivery models, the reaction was even more direct: the NIFTY IT index dropped 6.3%, its steepest fall since early 2020, with major firms like Infosys and TCS sliding sharply.
🔄 What Changed: “Assistants” Are Becoming “Doers”
A year ago, most workplace AI was pitched as copilots: help you write, summarize, or brainstorm. This week’s anxiety is about something else—agents that complete tasks end-to-end (pull documents, run workflows, produce deliverables) and can substitute for chunks of human-in-the-loop professional work.
Reuters reports that Anthropic’s new plug-ins specifically raised fears about automation pressure in legal, financial, sales, and related information-services businesses.
🗣️ Jensen Huang’s Counterpoint: “Illogical” to Think AI Replaces Software
At a Cisco-hosted AI event in San Francisco on Wednesday, Nvidia CEO Jensen Huang pushed back on the narrative that AI “replaces software.” He called the fear “illogical,” arguing that AI systems still depend on software tools and infrastructure rather than making them obsolete.
That tension is the heart of the moment: markets are trading the near-term fear (services revenue gets compressed) while platform leaders argue the longer-term reality (more AI means more tooling, more infrastructure, more demand for software).
🍏 Apple’s Timing Didn’t Help the “Everything is Fine” Vibe
One day before Huang’s comments hit wires, Apple published its own signal that the tooling layer is changing fast: Xcode 26.3 adds “agentic coding” support.
This update lets developers use coding agents—including Anthropic’s Claude Agent and OpenAI’s Codex—directly inside Xcode to tackle tasks autonomously (explore a project, run tests, fix bugs, and more).
This isn’t just “autocomplete for code.” It’s Apple saying: the mainstream IDE can now host agents that operate like a junior developer who never gets tired.
🌏 And Yes, Pricing Is Getting More “Global”
On the consumer side, OpenAI is continuing to push a lower-cost subscription tier. The ChatGPT Go plan, which initially launched in India at ₹399/month (about $4.50), is now part of a broader push to localize pricing. This aggressive pricing strategy suggests the “agent economy” will be volume-based, not just luxury-priced.
📝 My Read: This Isn’t “The End of Software” — It’s a Brutal Repricing of Labor-Heavy Models
The hottest take is that “AI just deleted software.” The more accurate take is messier:
- Software isn’t going away. It’s becoming the substrate agents run on.
- Labor-heavy workflows are being repriced. If a tool can do the first 60–80% of routine work quickly, the value shifts to oversight, domain expertise, and the last-mile decisions.
- Markets move faster than reality. Stock selloffs are often an early, emotional proxy for slower operational change.
In plain English: the band isn’t playing on the Titanic—but the ship is absolutely turning, and entire business models are trying to keep their footing while it does.
About the Author
Chad Hembree is a certified network engineer with 30 years of experience in IT and networking. He hosted the nationally syndicated radio show “Tech Talk with Chad Hembree” throughout the 1990s and into the early 2000s, and previously served as CEO of DataStar. Today, he’s based in Berea as the Executive Director of The Spotlight Playhouse—proof that some careers don’t pivot, they evolve.