Dear Reader,
This week, I’ve been juggling two big challenges: fine-tuning Claude's MCP server tools for a client project and gearing up for a hackathon in Estonia. It’s been a balancing act of practical problem-solving and forward-thinking innovation.
When the Right Tool Makes All the Difference
I ran into a tricky JavaScript problem this week—nothing too advanced, but just complex enough to be frustrating. I tried debugging it with ChatGPT for an hour, but the suggestions, while close, didn’t quite solve the issue.
On a whim, I copied the exact same prompt into Claude and received working code within seconds. The difference was night and day. It wasn't that ChatGPT is "bad" - it's brilliant at many things - but in this specific instance, Claude's code reasoning was simply better aligned with my particular problem.
This experience reminded me of something important: there is no "best" AI assistant, only the right tool for specific tasks.
The AI FOMO Trap
If you've been following AI news, you've likely noticed how quickly these models are evolving. One week Claude has better reasoning, the next week Grok has real-time web access, then ChatGPT releases a new feature, and Google's Gemini leapfrogs them all with something else.
This constant innovation can trigger FOMO (Fear Of Missing Out)—that nagging feeling that you’re missing something essential if you’re not using the latest and greatest AI model.
Research suggests that while 72% of companies globally have incorporated AI in some way, only about 4% are leveraging it to drive meaningful innovation across their operations. This disconnect highlights how many AI adoptions are driven more by competitive pressure than by strategic alignment with actual needs.
The Current AI Assistant Landscape
Let's take a quick look at the unique strengths each major assistant brings to the table:
ChatGPT (OpenAI)
Shines at
- General knowledge
- Creative writing,
- Explaining concepts
- Integration with other tools
|
|
Not so great at:
- Complex reasoning tasks
- Certain types of coding (as I discovered)
- Handling very nuanced ethical questions
|
Claude (Anthropic)
Shines at
- Nuanced reasoning
- Code generation and explanation
- Document analysis
- Long-form content
|
|
Not so great at:
- Real-time information (without plugins)
- Handling very ambiguous requests
- Remembering chat topics long term
|
Grok (xAI)
Shines at
- Real-time data access,
Less filtered responses on controversial topics (for better or worse)
- Personality-driven interactions
|
|
Not so great at:
- Deep analytical tasks
- Having the same level of refinement as more established models
|
Gemini (Google)
Shines at
- Knowledge tasks
- Integration with Google's ecosystem
- Multimodal capabilities
- Notebook LM is also pretty awesome on its own
|
|
Not so great at:
- Creative writing
- Handling some types of complex reasoning
|
Alexa+ (Amazon)
Shines at
- Agentic capabilities (completing tasks autonomously)
- Voice interaction
- Integration with Amazon's ecosystem
|
|
Not so great at:
- Text-based detailed analysis
- Some aspects of creative content generation
|
Finding the Right AI Assistant: A Practical Approach
Instead of chasing the latest AI model, focus on these four steps:
- Define Your Needs: Are you writing, coding, researching, or analysing data? Prioritise your tasks.
- Limit Your Toolkit: Choose 2-3 assistants that match your needs and test them.
- Play to Their Strengths: Use each tool where it shines—ChatGPT for creative writing, Claude for code, Perplexity for research.
- Outcome Over Features: Focus on results rather than flashy features.
The Multi-Tool Approach
I’ve adopted a “multi-tool approach”—choosing the right AI for each task:
- Claude: Code generation, document analysis, and nuanced writing.
- Perplexity: Research and up-to-date information.
- ChatGPT: Creative ideation and quick explanations.
- DALL-E / SORA: Visual content creation.
- Gemma & Mistral: Local, offline knowledge work.
This approach has significantly boosted my productivity compared to when I was trying to make a single AI tool do everything.
The best AI assistant isn’t always the latest or most feature-rich—it’s the one that works most effectively for your specific needs. While these platforms continue to leapfrog each other with new capabilities, they all eventually converge on similar feature sets. Instead of giving in to AI FOMO, evaluate AI assistants based on how well they handle your specific inputs and deliver the outcomes you need. The most valuable AI tool isn’t the one with the most features—it’s the one that consistently delivers results for you.
Ready to Master Your AI Workflow?
Feeling overwhelmed by the endless AI options? I can help you find the perfect mix for your needs. Book a 90-minute AI Strategy Session and discover your ideal setup.
🚀 AI Toolkit: Ready-Made Solutions to Save You Time
Want to skip straight to proven solutions? Check out my AI Toolkit at tools.informatic.ai. It's packed with ready-to-use AI tools I've personally tested and refined to save you hours each week. Try it for free with 10 credits to get started.
Next Week: Estonia and the Valencia DANA Project
Next week I'll be writing to you from Estonia, where my team is participating in the Valencia DANA Project hackathon in Tallinn. This innovative collaboration between Startup Valencia and Estonia's e-Residency program brings together entrepreneurs and tech experts to develop solutions that help cities anticipate and respond to extreme weather events.
After the devastating floods that hit Valencia last October, this initiative feels especially meaningful. We'll be working alongside teams from across Europe to create practical technologies for emergency management and urban resilience that can be implemented back in Valencia. I'm looking forward to sharing insights from this unique cross-border innovation effort with you!
What's your experience with different AI assistants? Have you found certain ones excel at specific tasks? I'd love to hear your thoughts!
Stay curious and stay informed,
Jim
