Which AI tools does this work with?

Any current model. It was seen with in the sources.

Where did this come from?

Anthropic Engineering (Anthropic Engineering). Subvolts collected, normalized and credited it; the link above opens the original.

Checklist Hiring and careerExpert · 15 min

Design principles for robust technical evaluations

Key principles to ensure technical assessments remain effective as AI capabilities advance.

Any modelParaphrased

Items

Open in ChatGPT Open in Claude

Representative of real work
High signal with wide scoring distribution
Avoids single insights
Ensures depth, not all candidates finish
No specific domain knowledge required
Fun with fast development loops and room for creativity

Representative of real work
High signal with wide scoring distribution
Avoids single insights
Ensures depth, not all candidates finish
No specific domain knowledge required
Fun with fast development loops and room for creativity

Before or after

before

Context

Designing technical evaluations that are resistant to AI assistance.

What it does and when to use it

Use these principles when designing technical evaluations to ensure they accurately assess candidate skills. Focus on tasks representative of real work, offering multiple opportunities for candidates to demonstrate abilities and avoiding reliance on narrow expertise or single insights.

Why it works

These principles aim to create evaluations that are engaging, provide a high signal-to-noise ratio, and are difficult for AI to solve trivially, thus maintaining their effectiveness over time.

Sources and credit

Designing AI resistant technical evaluations ↗Anthropic Engineering · Anthropic Engineering · Jan 21, 2026 · paraphrased

First seen on Subvolts Jan 21, 2026. Updated Sep 11, 2026. Spotted a problem? Tell us.

Where it shows up

Designing AI resistant technical evaluationsAnthropic · Anthropic Engineering · Jan 21, 2026

Related in the PlaybookPlaybook →

The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.

HeuristicAI transition planning map for EU occupationsUse archetypes to plan for AI's impact on different job categories. · 10 min
TemplateAI-resistant take-home evaluation challengeA template for a take-home coding challenge designed to be resistant to AI solutions. · 60 min
HeuristicIntegrate community co-design for inclusive AI developmentInvolve individuals from disability communities as co-designers from the start of development. · 5 min
ChecklistSocial engineering email detection checklistQuestions to ask when evaluating suspicious outreach emails that mimic legitimate hiring or consulting offers. · 5 min
Anti-patternTreating AI-generated outreach as inherently legitimateDon't assume AI-generated emails are safe; verify legitimacy, especially for unsolicited offers. · 2 min
PromptRole-play a student struggling with math for AI tutor simulationSimulate a student's math struggles and uncertainty for AI tutor evaluation. · 5 min

Questions people ask

Which AI tools does this work with?: Any current model. It was seen with in the sources.
Where did this come from?: Anthropic Engineering (Anthropic Engineering). Subvolts collected, normalized and credited it; the link above opens the original.

Design principles for robust technical evaluations

Questions people ask

OpenAI (ChatGPT)

Google (Gemini)

Anthropic (Claude)

Microsoft (Copilot)

Meta (Meta AI)

xAI (Grok)

Across the six

Subvolts