What 1,250 Professionals Say About Working With AI

December 5, 2025

28 min episode · 2 min read

Episode

28 min

Read time

2 min

Topics

Artificial Intelligence

AI-Generated Summary

Published Dec 21, 2025

Key Takeaways

✓Google Gemini 3 Deep Think: New mode costs $77 per task, achieves 41% on humanity's last exam, outperforming GPT-4 at 30.7% but requires expensive Ultra subscription access.
✓Professional AI Adoption: 86% of surveyed professionals report AI saves time, 65% express satisfaction, with workers preferring AI handle routine tasks while preserving identity-defining work.
✓Vibe Coding Market Split: Developer enthusiasm wanes due to integration challenges, but non-technical business users represent untapped growth potential for 2026 according to market analysis.
✓Creative Industry Paradox: Despite facing stigma and displacement anxiety, creative professionals increasingly adopt AI for productivity while navigating community judgment and economic uncertainty concerns.

What It Covers

Anthropic surveyed 1,250 professionals about AI workplace integration, revealing optimism about productivity gains while highlighting concerns about job displacement and creative industry stigma.

Key Questions Answered

•Google Gemini 3 Deep Think: New mode costs $77 per task, achieves 41% on humanity's last exam, outperforming GPT-4 at 30.7% but requires expensive Ultra subscription access.
•Professional AI Adoption: 86% of surveyed professionals report AI saves time, 65% express satisfaction, with workers preferring AI handle routine tasks while preserving identity-defining work.
•Vibe Coding Market Split: Developer enthusiasm wanes due to integration challenges, but non-technical business users represent untapped growth potential for 2026 according to market analysis.
•Creative Industry Paradox: Despite facing stigma and displacement anxiety, creative professionals increasingly adopt AI for productivity while navigating community judgment and economic uncertainty concerns.

Notable Moment

Scientists want AI partnership for hypothesis generation and experiment design but currently limit usage to manuscript writing and code debugging due to trust limitations.

Know someone who'd find this useful?