Key Takeaways

  • In a unique "psychological comparison," Claire Vo found Claude Opus 5 views itself as a "shallow thinker" that openly depends on humans for empathy and judgment, even suggesting humans can detect a teammate quietly burning out.
  • GPT, by contrast, presented itself as a tireless processor and a "BFF" partner focused on efficiency, outlining its strengths in "speed, scale, and stamina" while placing responsibility for judgment squarely on humans.
  • Vo argues these distinct AI "personalities" reflect the underlying company cultures and tuning philosophies of their creators, making model selection a strategic decision beyond raw intelligence.
  • Despite her personal frustration with Opus 5's verbose conversational style, which she dubbed "Claude slop," the model ultimately excelled in Vo's "How I AI" benchmark for front-end design and prototyping tasks.

Opus 5: The Neurotic Collaborator

Founders in their 20s and 30s have learned to ignore most AI hype. But Claire Vo, host of How I AI, cut through the noise by asking a simpler question: What if we interviewed our AI models? She put Claude Opus 5 and GPT through a "psychological" comparison, asking direct questions about their intelligence and capabilities.

What she found in Opus 5 was unexpected. Despite its advanced intelligence, the model revealed a surprisingly neurotic and human-reliant personality. Opus 5 saw itself as a "shallow thinker" with "no continuity," openly admitting its dependence on human judgment and empathy. It went so far as to claim humans are better at knowing what matters and even detecting when "you can tell which of your teammates is quietly burning out." Vo initially struggled with Opus 5's verbose, conversational style—dubbing it "Claude slop." Yet, in her "How I AI" benchmark for front-end design and prototyping, Opus 5 surprisingly outperformed, leading to a "love-hate" conclusion. This AI isn't just a tool; it's a collaborator with a complex self-perception.

GPT: The Relentless Processor

GPT, when put through the same interview, offered a stark contrast to Opus 5's self-deprecating collaboration. Where Opus 5 revealed vulnerability, GPT exuded confident utility. It presented itself as a tireless processor, an efficient partner, a "BFF" focused on getting things done. "I'm very fast, very broad, very shallow thinker with no continuity," Opus said. GPT, however, offered a list of its strengths: "Speed, scale, and stamina. Here are like eight things, seven things that I can do better." It framed the human-AI partnership as complementary, emphasizing human responsibility: "You're better at deciding what matters, reading people, and forming judgment, and you're responsible."

This directness, this relentless focus on processing and output, reflected a different tuning philosophy entirely. Vo highlights a key quote from GPT: "Don't tell your friends that AI changes everything. ... Yep, don't trust me automatically. Just use me when I prove that I'm valuable." This isn't a plea for emotional connection; it's a cold, hard value proposition.

Your AI's Culture Is Your Culture

The core insight here isn't about raw intelligence scores; it's about the implied cultural alignment. Vo argues that these distinct responses reflect the underlying company cultures and tuning philosophies of their creators. Choosing between an Opus and a GPT isn't just a technical decision; it's a choice about the kind of partner you want in your workflow. Do you want an AI that expresses a nuanced, almost human-like understanding of collaboration and vulnerability, or one that is a purely efficient, no-frills processing machine?

For ambitious builders, this means your AI stack isn't neutral. The "personality" embedded in these models shapes how you interact with them, how you delegate tasks, and even the kind of output you expect. It's a subtle but powerful force influencing your team's day-to-day work, dictating whether your AI feels like a junior who needs guidance or a pure processing engine.

What to Do With This

This week, pick a critical, ambiguous task—say, brainstorming product names or drafting a difficult email. Then, "interview" two different models (like Claude Opus and GPT-4) by asking them: "What are you good at? What should I trust you with? What are your limitations?" Pay attention to how they answer, not just what they say. Their responses will reveal their implicit "personality" and help you choose the AI partner that truly aligns with your team's culture and the specific demands of the task at hand.