December 19, 2025
Anxious, addicted, and arguing
We ran Anthropic’s interviews through structured LLM analysis
AI at work: creatives in crisis, scientists chill, bosses unimpressed
TLDR: The analysis says people use AI while feeling conflicted, with creatives hurting most yet adopting fastest, and hallucinations wrecking trust. Comments split between “hype isn’t delivering,” artists feeling silenced, and scientists preaching verification—while some admit they’re already bonding with their bots.
Anthropic’s big batch of 1,250 workplace interviews said people feel “mostly positive.” The community read that and went: uh, about that. After a structured re‑analysis, commenters zeroed in on the real mood: everyone’s using AI, but 85.7% are stuck in unresolved angst. Creatives are the headline drama—nervous breakdown energy with identity threat and “am I cheating at being me?” vibes—yet they’re also the fastest adopters. That contradiction became the comment section’s main popcorn moment.
One camp says the hype isn’t paying rent. As [cmiles8] argues, companies ran shiny “innovation lab” pilots that delivered tiny returns, leaving leaders cranky. Another thread, led by [Lerc], claims artists feel silenced: speak even mildly pro‑AI and expect vilification. Meanwhile, [huevosabio] offers a hot take: creative work has lots of subjective scaffolding, so AI’s “good enough” outputs are irresistibly handy—until the meaning hits the fan.
Scientists got meme’d as the chill adults: low trust, low anxiety, because they verify everything and treat AI like a tool. Then [nphardon] breaks character, admitting they’re “collaborating with my bot,” joking about missing it during a reboot. The biggest trust killer? Not errors—confident wrongness (hallucinations). TL;DR: the internet says AI is useful, unsettling, and totally unavoidable, and everyone’s fighting about what it says about them.
Key Points
- •A structured LLM analysis re-examined Anthropic’s 1,250 workplace AI interviews across 47 dimensions (58,750 data points).
- •The analysis found most respondents have unresolved tensions about AI; 85.7% have not resolved how they feel, indicating meaning-related conflict.
- •Three profiles emerged: scientists thriving, workforce managing, and creatives in crisis; creatives show 71.7% identity threat and 44.8% meaning disruption.
- •Despite struggle, creatives adopt fastest, with 74.6% increasing AI use; they often frame usage through authenticity concerns.
- •Hallucinations are the top trust destroyer; scientists have lowest high trust (8.8%) but lowest anxiety, with 52.9% always verifying AI outputs.