Berkeley Haas pilots Rubrica AI to grade handwritten exams, cutting time 40-60%
_akpiper · x · 2026-11-20
UC Berkeley is piloting Rubrica, an AI grading tool for handwritten work, claiming 40-60% time savings as professors revert to written exams to counter student AI use. An MBA student built it after being the sole TA for a 133-student class; it will be trialed this spring with up to 15 Haas faculty, and instructors must review its grades before release. A commenter notes AI grading of student writing has been poor in their experience.
More from Companies & People
- Reid Hoffman recalls leading OpenAI's first commercial round with no product or business plan — reidhoffman · 2026-11-20
- Nat Lambert jokes: how much equity went to Opus 5.5, and do model weights get a say in Anthropic's future? — natolambert · 2026-11-20
- Brian Armstrong on success: no rear-view mirror, just build the next thing — davidsenra · 2026-11-20
- Zapier's 4-level AI fluency rubric: top PMs run eval-first development and ship fine-tuned LLMs — aakashgupta · 2026-11-20
- Peter Yang: Anthropic's biggest Fortune 500 rival is Microsoft Copilot — petergyang · 2026-11-20
- Garry Tan backs Bostrom's 'if nobody builds it, everyone dies', urges taking Deep Utopia seriously — garrytan · 2026-11-20