# Don't Use AI? You Might Not Get the Job. > [!question] Lingering Questions > - How widespread is Norcross's experience (rejected for "not using enough AI") versus an isolated interviewer's idiosyncrasy? The episode gives one testimony, not a market survey. > - What would a task-specific AI-use rubric actually look like for a WordPress maintenance role — required for legacy PHP triage, irrelevant for architecture decisions? > - How much of the WordPress hiring contraction traces to AI substitution versus consolidation, post-pandemic over-hiring correction, and market saturation? The episode names all four causes but doesn't weight them. > [!tip] Article Ideas > - "AI fluency is not a hiring signal" — task-specific rubric piece: what to test for when the job is verifying generated code against current standards rather than producing prompts quickly. Direct application of the vault's existing proxy-screening critique to AI tooling specifically. > [!link] Cross Reference > - [[3. Resources/Permanent Notes/Credential Screens Select for the Proxy Not the Work]] — same underlying pattern (a convenient screen selects for the proxy, not the capability), applied here to AI tool use instead of academic credentials > - [[3. Resources/MOCs/AI Cost & Tokenomics MOC]] — labor-market angle on AI adoption pressure > [!quote] Vocabulary > - *"babysit a robot"* — Norcross's phrase for a review-only role where a developer's job becomes checking AI output rather than producing solutions, which he frames as devaluing the expertise he built ## Bib capsule Eric Karkovack interviews veteran WordPress developer Andrew Norcross on *The WP Minute*. Norcross recounts losing a job opportunity after being told he didn't use enough AI in his workflow, and the two discuss what that signals about hiring, craft knowledge, and the broader WordPress labor market. ## Author's thesis Hiring markets are starting to reward visible AI tool adoption as a qualification in itself, separate from demonstrated ability to do the job — and that's a bad hiring signal, because AI fluency doesn't predict whether a candidate can verify generated code, catch outdated PHP patterns, or exercise the domain judgment the role actually requires. AI is one factor in a weak WordPress job market alongside layoffs, consolidation, and changing product economics — not the sole cause. ## Key arguments (≤3) 1. **A hiring decision mistook tool adoption for competence.** Norcross reports being told directly he wasn't using enough AI, making tool visibility a qualification apart from demonstrated work quality. 2. **Selective AI use, not blanket refusal, is the defensible position.** Norcross accepts AI for "mundane stuff, admin-level work, boilerplate" but resists a role reduced to reviewing AI output — arguing that this removes the problem-solving work that motivates and justifies his expertise. 3. **Weak markets pressure workers into tool adoption regardless of preference** — "anybody who has a role is doing whatever they can to hold on to it" — which muddies any claim that AI-tool enthusiasm reflects genuine preference versus economic coercion. ## Claim atoms ### Hiring can mistake tool use for competence > [!quote] ID:rw_1040672512 > "I was told that I did not use enough AI in my own workflows." Where: 20:41. The reported rejection makes tool adoption a qualification apart from demonstrated work. ### Routine tasks fit bounded AI assistance > [!quote] ID:rw_1040672513 > "AI seems to be pretty useful for mundane stuff. You know, admin-level work boilerplate" Where: 18:26. Norcross's actual position is selective use, not blanket refusal — important context against reading him as anti-AI. ### AI fluency does not prove job competence > [!quote] ID:rw_1040672516 > "you're already hearing employers saying, 'These people that know AI don't actually know how to do the job.'" Where: 41:20. The market correction Norcross anticipates: hiring needs separate evidence for tool knowledge and domain performance, not one as a proxy for the other. ## Permanent candidates - **[P1] AI-use visibility is a proxy screen, not a competence signal** — direct application of the vault's existing "Credential Screens Select for the Proxy Not the Work" pattern to AI tooling specifically. Strong candidate; links to that note as parent pattern rather than restating it from scratch. ## Processing checklist - [ ] Look for corroborating reports (other WordPress developers rejected for low visible AI use) before treating this as more than one testimony - [ ] Draft a task-specific AI-use rubric distinguishing "required for this task" from "irrelevant to this task" for WordPress maintenance roles - [x] Connect into relevant MOC — [[3. Resources/MOCs/AI Cost & Tokenomics MOC]] ## Links - [[3. Resources/MOCs/AI Cost & Tokenomics MOC]] — labor-market/hiring-economics angle on AI adoption - [[3. Resources/Permanent Notes/Credential Screens Select for the Proxy Not the Work]] — parent pattern this note's Permanent candidate extends - [[3. Resources/MOCs/AI & the Future of Knowledge Work MOC]] — applies-in: Joins the cluster on AI and the changing labor market.