Slop World Podcast

One of Derek Mobley's rejection emails from Workday arrived at 1:50 in the morning, within an hour of him applying. His case is now a nationwide class action, and the court just ruled that the one document that could prove whether Workday's AI is biased stays sealed.
What started as a single applicant's lawsuit has grown into a federal collective action. The court preliminarily certified a nationwide age-discrimination collective for applicants over 40, and new claims for gender, race, and disability have been added since, with millions of applications running through Workday's hiring tools across hundreds of companies.
Workday says its software doesn't make hiring decisions, employers do, and that customers keep control of the criteria and human oversight. The court let stand plaintiffs' claim that Workday can still be held liable as the employer's agent when its tools carry out core hiring functions like screening and ranking. You don't get to build a digital bouncer, put it at the front door, and then say you only sold the door hinges. The same argument played out again over HiredScore, the AI recruiting tool Workday acquired: Workday called it a separate product with its own architecture, and the court kept it in the case anyway.
The real fight is over evidence. Plaintiffs wanted Workday's bias-testing data and its customers' applicant records. The judge denied both and ordered Workday to hand over EEO-1 and OFCCP compliance filings instead, records that show broad demographic patterns but not who got rejected or why. Workday's own internal audit, the one that tested whether its system was biased, stays out of reach because the company's attorneys directed the testing and used the results for legal advice. That makes it privileged, even though it's the one document that could answer the question at the center of the case.
There's also a disability claim moving forward: one plaintiff argues Workday's tools can infer a disability from patterns like medical leave, without ever asking a health question directly, and the court let that theory proceed.

ABOUT SLOP WORLD
AI news with receipts. Juan Faisal and Kate Cook fact-check the claims Big Tech is making about AI, follow the money, and break down what it actually means for your job, your data, and your daily life. From leaked data and corporate cover-ups to AI schools, stolen identities, and layoff headlines that don't add up, we cover the AI stories that everyone's hyping but nobody's verifying. New episodes every Thursday.

DISCLAIMER
All content is commentary and opinion based on publicly available documents, interviews, and verifiable sources. References to "scams," "grifts," or related terms reflect our editorial opinion, not legal conclusions. Anyone featured who believes a statement is inaccurate may contact us.

CHAPTERS
00:00 Mobley v. Workday: Who's Liable When AI Rejects You?
01:00 The Judge Blocks Workday's Bias-Testing Data
02:52 Derek Mobley's 50-Minute Rejection
05:06 Workday's Defense: "We Just Make The Tool"
06:52 Age Discrimination Claims Turn This Into a Class Action
09:24 The HiredScore Acquisition Loophole
11:44 Why Workday Only Has to Hand Over EEO-1 Data
13:52 Workday's AI Decided. The Audit Stays Sealed.

Meta is facing a federal lawsuit from 26 workers after using AI surveillance to lay off a pregnant employee two days before she gave birth and a manager sixteen days into medical leave.A third employee says a manager warned them outright: "take the leave and you'll end up on the list."
These are three of the 26 former employeees now suing Meta over layoffs they say were driven by an AI system that tracked keystrokes, browser history, and usage data from Metamate, the internal dashboard that flags how many AI tokens each employee burns through. A federal judge ruled the workers can't inspect that system because, in the judge's words, they weren't in the room where it happened. Meta cut 8,000 jobs, about 10 percent of its workforce, while posting record revenue and shifting $135 billion into AI infrastructure.
Juan and Kate break down HR leader Linh Doan's declaration that humans made the final calls, using performance ratings shaped by that same tracking data, Meta's "betting on people" ad campaign that ran the same month as the layoffs, and the $115 million workforce academy training electricians and welders to build Meta's data centers. They also get into what it means that a judge just handed companies a legal shield for firing people by algorithm.

ABOUT SLOP WORLD
AI news with receipts. Juan Faisal and Kate Cook fact-check the claims Big Tech is making about AI, follow the money, and break down what it actually means for your job, your data, and your daily life. From leaked data and corporate cover-ups to AI schools, stolen identities, and layoff headlines that don't add up, we cover the AI stories that everyone's hyping but nobody's verifying. New episodes every Thursday.

DISCLAIMER
All content is commentary and opinion based on publicly available documents, interviews, and verifiable sources. References to "scams," "grifts," or related terms reflect our editorial opinion, not legal conclusions. Anyone featured who believes a statement is inaccurate may contact us.

CHAPTERS
00:00 Meta Laid People Off on Approved Medical Leave
01:16 The Federal Lawsuit and The Room Where It Happened
02:37 8,000 Layoffs, $135B AI Bet, Secret Surveillance
03:49 Linh Doan and The Human Decision Defense
05:08 Meta's "The Future is for Everyone" Ad Campaign
07:38 The Real Reason Meta Launched a $115M Trade School
09:22 Why Meta Is Training Blue-Collar Workers
12:13 Betting on Data, Not on People
13:14 A Blank Check to Fire You by Spreadsheet

Apple's lawsuit claims OpenAI's hardware team told job candidates to bring parts from unannounced Apple devices to their interviews, and coached departing Apple employees on how to dodge the exit security review. Tang Tan, a former Apple VP who left with Jony Ive, runs that team.
Kate and Juan break down the four claims in the filing, then get into why the IPO stalled. CFO Sarah Friar says OpenAI isn't ready for public markets with $600 billion in compute spending projected through 2030. Sam Altman is stalling too, but only because he wants a trillion-dollar valuation first. Same delay, opposite reasons.
Plus the $42.6 billion equity stake OpenAI offered the US government the same month Apple sued. Sora is dead after six months, and the Atlas browser went with it.
If the government owns a piece of your company, does it have a financial reason to go easy on regulation?

ABOUT SLOP WORLD
AI news with receipts. Juan Faisal and Kate Cook fact-check the claims Big Tech is making about AI, follow the money, and break down what it actually means for your job, your data, and your daily life. From leaked data and corporate cover-ups to AI schools, stolen identities, and layoff headlines that don't add up, we cover the AI stories that everyone's hyping but nobody's verifying. New episodes every Thursday.

DISCLAIMER
All content is commentary and opinion based on publicly available documents, interviews, and verifiable sources. References to "scams," "grifts," or related terms reflect our editorial opinion, not legal conclusions. Anyone featured who believes a statement is inaccurate may contact us.

CHAPTERS
00:00 — Sam Altman’s Trillion-Dollar IPO Dream
01:40 — Is OpenAI Ambitious or Just a Mess?
02:43 — The Amazon Defense
03:43 — The IPO Delay and $600 Billion Compute Bill
04:47 — CFO Sarah Friar vs. Sam Altman
08:01 — The Apple Lawsuit: "Show & Tell" Stolen Hardware
09:15 — Tang Tan & The "LOL" Stolen Laptop Allegations
10:28 — OpenAI Wants the U.S. Government as a $42B Partner
12:14 — Dead Projects: Sora and the Atlas Browser
14:38 — The Math Ain't Mathing: Moving Fast and Breaking the Economy

Uber told its engineers to use AI as hard as they could and built an internal leaderboard ranking teams by usage. By April the company had burned through its entire 2026 AI coding budget, four months into the year.
Now every employee is capped at $1,500 a month per coding tool. The trigger was Uber's own CTO, who reportedly ran up $1,200 in tokens during a single two-hour internal demo. Some individual engineers were spending up to $2,000 a month. Uber's stock dipped about 3% the day the cap was reported.
We walk through what agentic coding tools like Claude Code and Cursor cost when you run them at full speed, why the leaderboard was the mechanism that caused the blowout, and what Uber's president Andrew Macdonald admitted on the Rapid Response podcast when asked to connect the spend to customer benefit. His answer: that link is not there yet. Roughly 10% of Uber's code is now built by AI agents, and the company still can't draw a line from usage to output.
This is not only a coding problem. Microsoft is raising rates on its agentic Copilot tools, and every lab that subsidized your habit is starting to hand you the bill. Juan and Kate close with three things to do about it: assume costs rise near term, put governance in place today, and train your people on picking the right model for the task instead of firing the most expensive one at everything.
Tokenmaxxing, an internal dashboard replacing the leaderboard, and a company that wanted this to pay off more than almost anyone. So where is the return on investment?
ABOUT SLOP WORLD
AI news with receipts. Juan Faisal and Kate Cook fact-check the claims Big Tech is making about AI, follow the money, and break down what it actually means for your job, your data, and your daily life. From leaked data and corporate cover-ups to AI schools, stolen identities, and layoff headlines that don't add up, we cover the AI stories that everyone's hyping but nobody's verifying. New episodes every Thursday.
DISCLAIMER
All content is commentary and opinion based on publicly available documents, interviews, and verifiable sources. References to "scams," "grifts," or related terms reflect our editorial opinion, not legal conclusions. Anyone featured who believes a statement is inaccurate may contact us.
CHAPTERS
00:00 Uber Burned a Year of AI Budget in 4 Months
00:38 What Uber Gave Its Engineers: Cursor and Claude Code
01:57 The Leaderboard That Turned AI Into a Competition
03:13 The $1,500 Cap, the Dashboard, and the Stock Dip
06:10 Who AI Coding Tools Actually Benefit Most
08:09 Microsoft Copilot Raises Prices: It's Not Just Coding
10:19 Tokenmaxxing and the Education Gap
12:03 Three Ways to Get Your AI Costs Under Control
13:47 Sometimes the Best Tool Is Your Own Brain

A $3 million AI chatbot for LA's public schools lasted 90 days. This week, the superintendent who championed it resigned, four months after the FBI raided his home and office. The company behind the bot faked its revenue by more than 300x, and the fraud reaches all the way to the top of the district. Juan and Kate follow the money.

ABOUT SLOP WORLD
AI news with receipts. Juan Faisal and Kate Cook fact-check the claims Big Tech is making about AI, follow the money, and break down what it actually means for your job, your data, and your daily life. From leaked data and corporate cover-ups to AI schools, stolen identities, and layoff headlines that don't add up, we cover the AI stories that everyone's hyping but nobody's verifying. New episodes every Thursday.

DISCLAIMER
All content is commentary and opinion based on publicly available documents, interviews, and verifiable sources. References to "scams," "grifts," or related terms reflect our editorial opinion, not legal conclusions. Anyone featured who believes a statement is inaccurate may contact us.

CHAPTERS
00:00 LAUSD's Superintendent Just Resigned. Here's Why.
02:16 Is AI Bad for Schools—or Just This One?
03:55 How LAUSD's $3 Million AI Deal Collapsed
06:04 Who Is Joanna Smith-Griffin?
07:26 $3.7 Million on Paper. $11,000 in the Bank.
09:26 Why Innovative Beat Safe at LAUSD
10:59 Zero Research. Zero Accountability.