Consider the following:
- There are already more tools available than you can possibly use
- The pace of new tool development is lightning fast (and accelerating)
- Many specialized tools may become obsolete as foundation models improve
Given these realities, it's helpful to develop a three-part approach:
- Maintain a core go-to toolkit (a handful of tools you know exceptionally well)
- Stay loosely aware of what else exists (which can be as simple as asking Perplexity if any tools solve your specific problem)
- Focus on developing the ability to learn new tools quickly rather than trying to master everything
When you hit a problem:
- Work with your core tools first - try different approaches and prompts
- Reach for specialized tools when you're genuinely stuck
- Use your core toolkit to research alternatives: "I tried general AI assistants for [problem] but need something more specialized. What tools are best for this?"