Felix Ortiz: Can I just — I want to start with a specific person, because I think she's the whole episode in one quote.
Felix Ortiz: Cristina Estupiñán, software developer, goes on record with Business Insider saying she felt compelled to feign excitement about AI tools in interviews — her word, feign — despite genuine personal skepticism. And that's reported in the same piece documenting that Challenger, Gray and Christmas tracked over 155,000 AI-related layoffs in 2025. Same article. Same moment.
Tess Hollis: So she may have lost work because of AI, and she's now performing enthusiasm for AI to get back in.
Felix Ortiz: To an AI, potentially. Because Experis — under ManpowerGroup — deployed a chatbot called Sophie to run initial screening interviews. Thirty minutes, realistic face, actually asks you questions. A tech professional named Bijo Thomas interviewed with Sophie and got hired as a senior AI solutions architect at Experis. But the point is Sophie is the gatekeeper.
Tess Hollis: Hold on — an AI is screening candidates for how much they like AI.
Felix Ortiz: That's the architecture, yeah.
Tess Hollis: Because Monday.com just filed with the SEC in July 2026 that it's cutting 600 people — 20% of its workforce — for AI-driven growth. Business Insider's CEO Barbara Peng framed her own company's 21% staff cut in 2025 as AI adoption strategy. So the companies generating this displacement are also designing the re-entry test. And the test is: prove you love what displaced you. I don't know what else to call that.
Felix Ortiz: I think the honest answer is — okay, companies would resist the word coercion. They'd say it's signaling fit. But there's something in the gap between what they're claiming they want and what the structure actually produces that I don't think they've reckoned with.
Tess Hollis: And the gap you're naming — that's not accidental, right? That's the closed loop. The structure produces the outcome and then evaluates you against the outcome it produced.
Felix Ortiz: Yeah, and okay — here's the plainest version I can make it. Imagine you got fired from a bakery because they bought a bread machine. Now you're interviewing at the next bakery and they sit you in front of the bread machine for thirty minutes so it can decide if you're sufficiently excited about bread machines. That's Sophie. That's the actual situation.
Tess Hollis: Wait — and Bijo Thomas actually went through that. Thirty minutes with Sophie, gets hired as a senior AI solutions architect.
Felix Ortiz: At Experis, yeah. Which — I mean, maybe that works out great for him, I don't know. But the selection mechanism is what's strange. Sophie is, right, she's described as looking like a real businesswoman. Business Insider reported this in July 2026 as a new phase. Not a pilot. A phase.
Tess Hollis: So what does Sophie actually evaluate? Like — what signal is she reading for?
Felix Ortiz: That's — okay, I don't actually know the rubric. And that's the part that I keep stumbling on. Because if an AI interviewer is optimizing for something — some pattern of enthusiasm, word choice, whatever — then the candidates who get through are the ones who model that pattern, not necessarily the ones who can do the job. Entry-level postings requiring AI skills have nearly tripled since Fall 2025, so there's already a real pressure. You add Sophie on top of that and you've got — wait, no, here's what you've got. You've got a system that can only select for the performance of the thing.
Tess Hollis: That's the recursion. The AI can't actually evaluate genuine AI competency — it can only detect whether you sound like someone who loves AI. Which is a completely learnable performance.
Felix Ortiz: And Cristina Estupiñán already figured that out. She's performing it. She told Business Insider she feigns the excitement. So she passed the test. Which means the test — it's not measuring what companies think it's measuring.
Tess Hollis: So what exactly has the company selected for? Not competency. The ability to perform what they want to believe about themselves.
Felix Ortiz: Right — but the part that breaks the clean version of that argument is Business Insider itself. Because Barbara Peng cut 21% of staff in 2025 and said it's AI adoption strategy. But more than 70% of their staff were already using Enterprise ChatGPT before the cuts. The adoption had already happened. So what exactly did the layoffs accomplish?
Tess Hollis: That's the number I can't get past.
Felix Ortiz: Yeah, and then look at Microsoft — Amy Coleman, their Chief People Officer, says the 4,800 roles they cut are not being replaced by AI. But also credits AI with changing how work gets done. I mean — those are two sentences that can't both be load-bearing.
Tess Hollis: So what's the defensible read? That AI changed the shape of the work, not the headcount — and the headcount cut is just regular restructuring with better PR language?
Felix Ortiz: That's actually — wait, that's probably the most honest version of it. TechCrunch tracked 20-plus major tech companies citing AI in workforce cuts through July 2026, and the framing they kept flagging was companies posting record revenues while cutting. So it's not cost pressure. It's something else getting labeled as AI.
Tess Hollis: Which means the AI hiring mandates — the hackathons, the performance review tracking, the tripling of AI skill requirements in job postings — those aren't measuring whether AI is working. They're measuring whether you'll perform the narrative that justified the cuts.
Felix Ortiz: And what the sourcing doesn't resolve — I'll say this straight — is whether candidates who perform that enthusiasm are actually getting better outcomes post-hire. We don't know if they're getting through Sophie and then finding the role's already half-ghost.
Tess Hollis: That part gets worse when you factor in who's actually pushing back — and there's a piece of this we haven't hit yet that I think reframes the whole thing.
Felix Ortiz: Oh, yeah. Yeah — that's the part I want to get to.
Tess Hollis: The pushback isn't fringe, though. CNBC reported on an actual coalition of AI skeptics coalescing — like, organizing — in 2025 and into 2026. So the private doubt that Cristina Estupiñán is suppressing in interviews? It's organized enough to have a coalition.
Felix Ortiz: That's — okay, that's the thing I actually think holds up even after every push you've made today. Because if it's a coalition — a reported, named coalition — then this isn't isolated candidates being personally anxious. The suppression is structural. There's enough of it to coalesce around.
Tess Hollis: And Hollywood.
Felix Ortiz: Yeah — okay, Hollywood is the one that actually snapped it for me. LA Times reported they were fighting AI publicly — in negotiations, in the press — while quietly posting AI-related job listings. Same public-versus-private gap. Different industry. Which means it's not a tech sector quirk, it's — wait, no, it's actually the operating logic of the whole moment. You perform opposition out front and recruit for capability in the back.
Tess Hollis: So the candidates performing enthusiasm and the studios performing resistance are doing the exact same thing from opposite sides.
Felix Ortiz: Exactly. And the 155,000 AI-attributed layoffs from Challenger, Gray and Christmas — those didn't stop in 2025. Significant cuts kept going through the first half of 2026. So the system displacing workers is the same system running the re-entry evaluation. That's not incidental pressure on candidates. That's structural.
Tess Hollis: Huh. Okay. That part holds. The gap between what people say in interviews and what they actually think isn't a personal failure — it's a rational response to a system that punishes honesty. I'll give you that.
Felix Ortiz: And what that means for hiring — I mean, what it actually selects for — is performative flexibility. Not capability. Which is a way weirder outcome than anyone naming a skills gap is willing to say out loud.
Tess Hollis: And that's where I can't follow you to a clean answer. Because Sophie — the actual rubric Sophie is running — there's no validated method for distinguishing someone who genuinely knows how to use AI from someone who's rehearsed the correct emotional register. That's not speculation. The research is honest about it. Nobody has confirmed whether candidates who pass those screens are succeeding post-hire, or whether the gap shows up six months in when the role gets restructured anyway.
Felix Ortiz: Yeah. I — honestly, I don't have an answer for that. And I think that's the part I've been circling around without landing on. Like, we know the displacement is real — 155,000 in 2025, cuts still running into 2026. We know the enthusiasm theater is real. What we don't know is what happens after you get through the door.
Tess Hollis: Which means either companies run real competency assessments after hiring and discover the interview selected for theater — or employees find out mid-role that the AI fluency they performed was never tested and the job itself was already half-gone. Which one happens first probably tells us whether any of this was ever actually about skills. I don't know. Nobody does yet.
Felix Ortiz: No. That's — yeah, that's an open question I'm genuinely going to be sitting with.
Tess Hollis: Same. Thanks for thinking through it.
