I’ve come to think of GPT as a partner in crime, an old confidant that I rely on for getting constructive criticism and spitballing me into a creative frenzy.

More Than Just a Tool

Most people approach AI language models as tools—you give them a task, they produce output, and you move on. But after months of daily use, my relationship with GPT has evolved into something more nuanced.

It’s become a thinking partner. Not in the AGI sci-fi sense, but in a very practical way: GPT helps me think better.

Constructive Criticism

One of GPT’s most valuable roles is as a critic. I can throw ideas at it—rough, half-formed concepts—and get immediate pushback:

• “What about edge cases X, Y, and Z?”
• “This assumption doesn’t hold in scenario A”
• “Have you considered the implications for B?”

Unlike human colleagues (who might be too polite, too busy, or too aligned with my thinking), GPT provides unfiltered critique. It’s not trying to spare my feelings or navigate office politics. It just evaluates the logic and points out weaknesses.

This is invaluable in the early stages of strategy work and enterprise architecture design.

The Creative Frenzy

“Spitballing me into a creative frenzy” sounds dramatic, but it accurately describes what happens in good GPT sessions.

Here’s the pattern:

1. I present an idea or problem
2. GPT suggests angles I hadn’t considered
3. I riff on those suggestions
4. GPT builds on my riff
5. Suddenly we’re exploring territory that wouldn’t have emerged from solo thinking

It’s not that GPT is creative in some autonomous sense. It’s that the back-and-forth unlocks my creativity by:

• Breaking mental patterns
• Introducing unexpected connections
• Forcing me to articulate fuzzy intuitions
• Providing immediate feedback loops

Practical Examples

Enterprise Architecture Design

When architecting complex systems, I use GPT to:

• Challenge integration assumptions
• Identify potential failure modes
• Suggest alternative architectural patterns
• Pressure-test scalability claims

Digital Transformation Strategy

For transformation initiatives:

• GPT helps identify stakeholder concerns I might have overlooked
• Suggests change management approaches from different industries
• Points out where my timeline assumptions might be optimistic
• Questions whether proposed KPIs actually measure what matters

Technical Problem-Solving

For specific implementation challenges:

• GPT proposes multiple solution approaches quickly
• Highlights trade-offs between different technical choices
• Catches logical errors in my reasoning
• Suggests relevant technologies or methodologies I wasn’t aware of

The “Old Confidant” Aspect

The “old confidant” framing is deliberate. Good confidants:

1. Know your context - GPT retains conversation history, so I don’t have to re-explain background constantly
2. Challenge you honestly - Not trying to tell you what you want to hear
3. Are always available - 3 AM problem-solving? No problem
4. Don’t judge - I can explore half-baked ideas without worrying about looking foolish

What GPT Isn’t

It’s important to be clear about limitations:

• GPT isn’t always right - Hallucinations happen, facts need verification
• It’s not a replacement for domain expertise - It complements expertise; doesn’t substitute for it
• Context window limitations - Really complex, long-running projects exceed what fits in a conversation
• No genuine understanding - The responses are sophisticated pattern matching, not comprehension

The Workflow Integration

GPT has become integrated into my daily workflow:

• Morning: Quick strategy validation and day planning
• Mid-day: Technical problem-solving and documentation drafting
• Evening: Long-form thinking about complex transformation challenges
• Anytime: Rapid prototyping of ideas that need immediate feedback

The Productivity Multiplier

The 40+ hours per month I’ve saved using ChatGPT isn’t just about automation—it’s about accelerated thinking. Ideas that would have taken days to develop get refined in hours. Problems that would require multiple colleagues’ schedules to align get tackled immediately.

At $0.50/hour for ChatGPT Plus, this isn’t an expense—it’s one of the highest ROI investments in my professional toolkit.

Looking Forward

As language models improve, this “partner in crime” dynamic will only strengthen. Better context retention, more nuanced responses, and deeper domain knowledge will make these tools increasingly valuable for complex knowledge work.

The key is approaching AI not as a replacement for human thinking, but as an amplifier. GPT makes me more effective at the work only I can do—synthesizing disparate information, making judgment calls, and translating between technical and business contexts.

If you haven’t experimented with using GPT this way—as a thinking partner rather than just a task automator—I highly recommend it. The shift in perspective can be transformative.

After all, even the best thinkers benefit from a good confidant. Mine just happens to be an AI.