My Experience Using AI to Boost Productivity

AI assistants like Claude and ChatGPT have been game-changers for my productivity in many of my daily routines. I've been experimenting with them for note-taking, writing, brainstorming, and more. In this post, I'll share my learnings on how to use AI to work smarter.

Finding the Right AI Assistant

Not all AI assistants are created equal. Below are a few of the popular ones that I've tried and my thoughts on each:

• ChatGPT: Great for creative tasks, idea generation, and coding, but its writing can be inconsistent. It can be clear at times that it was produced by AI.
• Claude: My go-to for writing and prose. It produces very natural, high-quality text, which makes for a great starter point for drafts.
• Bing Chat: Decent, but I don't get as much success. It acts as a ChatGPT clone with additional guardrails that mostly hinder its ability to be helpful.
• Perplexity: Helpful for researching and finding sources. I often use Perplexity instead of Google to find relevant resources related to niche topics.

I test new assistants as additional ones pop up and product updates will continue to change which tools I use and when. The key is finding what works best for you to produce your ideal output for each task.

Crafting Prompts for Success

The prompts you write are crucial. Here are some things I've learned:

• Give very detailed instructions upfront on what you want.
• Share your personal style and preferences.
• Adjust and refine prompts over time.

I often iterate on the prompt I am using several times before getting an output I use. Over time, I have been refining the information I provide to craft prompts that work the best for me.

For example, I commonly take textbook notes and want them in a certain style, so my notes prompt specifies that I like bullet points and hierarchies. I still find that actively reading the source is the best way to absorb information, but AI has been great as a quick way to make sure I am not missing anything.

Prompt engineering takes experimentation but pays off.

My Favorite AI Workflows

Here are some of the workflows I've created:

• For project meetings, I record audio, transcribe with Whisper, and then have Claude summarize key insights. This has been great for my class meetings as I can quickly generate specific action items based on our discussions.
• I upload textbook and article PDFs to Claude to take notes for me.
• I give ChatGPT creative exercises for brainstorming. I especially love using the voice feature on the app to quickly get feedback on ideas.

The more context I provide, the better the results. Let me know if you are interested in learning more about any of the specific flows in detail.

Key Takeaways

• Try different AI assistants for different tasks. Find which excels where.
• Invest time in crafting effective prompts. Iterate as you learn.
• Don't be afraid to experiment. Track results and adjust.

Adding AI to your toolkit can seem intimidating, but with testing and patience, you can build workflows that make you far more productive.