Quote · This Week in Tech (Audio)
TWiT 1091: But You Didn't Move the Bodies - Surprising Supreme Court Move on Geofence Warrants
Where this was said
Enterprise AI Job Reality Check: Where Agentic AI Is Working and Where It Isn't
At 1:56:39 · chapter starts 1:48:10
Drawing on her coverage of contact center platforms at NoJitter, Lisa offers a grounded counterpoint to both AI hype and doom. AI is genuinely good at structured, routine customer queries: forgotten passwords, statement copies, balance lookups. Where it excels is not in replacing agents but in briefing them instantly — by the time a frustrated human reaches a live agent, the AI has already assembled their history, the query context, and suggested resolutions. The risk, she cautions, is burnout: when AI handles all the simple calls, human agents face an unrelenting stream of the hardest, angriest interactions. The panel also debates the ethics of companies not disclosing when a customer is talking to an AI, with broad consensus that transparency — and the right to reach a human — will eventually become a social expectation if not a legal requirement.
A Swedish court ordered Google to pay $1.97 billion (with interest) to Klarna's PriceRunner for favoring Google Shopping in search results over competing price comparison services.
Google lost its final appeal against a $4.7 billion EU antitrust fine related to its Android operating system, meaning it must now pay the full amount.
A year ago companies ran token leaderboards with bonuses. Today CFOs are installing monitoring tools and setting hard limits. One CTO said they went from 'letting every flower bloom' to 'CFOs with a lawnmower.' The Uber example is stark: they burned through their entire AI inference budget in four months.
Companies that had encouraged employees to maximize AI token usage reversed course in Q2 after CFOs discovered inference costs were skyrocketing, with one executive reporting Uber burned through its entire AI inference budget in four months.
