In this episode of Spilling the Tea on HR Tech, Stacey Harris and Cliff Stevenson dig into a news week in which the same question keeps surfacing: how much should you trust the data in front of you? Oracle laid off 30,000 employees, and the financial story behind that decision is more complicated than the AI narrative being used to explain it. Meanwhile, a meta-analysis of six major consulting firms working from the same data produced conclusions so different they barely resemble each other, which is its own argument for never acting on a single source.
They also tackle the real cost of removing humans from decision-making chains, using Laszlo Bock's post on AI-assisted military targeting as a case study for what happens when speed becomes the only goal.
Key points covered include:
↪️ The official narrative on Oracle's layoffs is AI replacing unnecessary roles, but the numbers behind $58 billion in borrowing and $156 billion in data center commitments tell a more complicated story about cash flow needs.
↪️ Six consulting firms analyzing the same AI data produced an array of starkly different conclusions. For instance, one firm concluded that AI had "zero economic impact" while another cited quadrupling productivity rates. According to the hosts, this shows why industry benchmarks and recommendations are not a reliable basis for your organization's AI decisions.
↪️ "Human in the loop" is not enough if that human is only doing a final sign-off without visibility into how the AI reached its recommendation
↪️ AI pricing is moving away from flat subscription models toward usage-based and connector-based charges, and buyers who treat AI as an unlimited resource right now are going to face a reckoning as that infrastructure debt comes due.
Don’t miss this exciting thought leader conversation! Follow the hosts and companies mentioned below:
Sapient Insights Group
Download the 2024-25 HR Systems Survey White Paper
Instagram | Twitter | LinkedIn
Stacey Harris
Instagram | Twitter | LinkedIn
Cliff Stevenson
Powered by the WRKdefined Podcast Network.


