For today’s use-case, we’ve turned to a different LLM nicknamed Claude. Claude was created by Anthropic with a focus on safely deploying AI throughout the world. As I continue to explore new practical AI integrations I am going to shuffle around language models as well.
My friends and I love puzzles. Trivia, Wordle, Crossword puzzles – you name it and we’ll play it. As we are eyeing down the start to NFL kickoff I decided that developing an NFL-themed crossword should be next on the docket. With an assist from Claude and an additional tool to be discussed later, we were able to come up with a playable crossword puzzle.
Rather than coming up with a summary on my own, I figured we might as well turn to Claude directly for a highlight-reel of its abilities:
To start, I had Claude come up with a list of 10 answers and 10 corresponding clues that were NFL themed. Despite my best efforts, neither Claude or GPT were able to handle the next request of putting together a crossword puzzle. This involves trial and error trying to ensure the answers overlap in a particular pattern. Neither AI was able to complete the task on their own, so I ended up turning to “Oh My Dots!” – a site that is also powered by AI. I took the prompts & answers from Claude and copied them into Oh My Dots to generate the puzzle and was provided the output in seconds.
Oh my Dots provided us a blank puzzle that could be interacted with on the website or printed out for offline use. Although it would have been nice to make a static html page with Claude, adapting on the fly was the lesson for today. It’s important to leverage these tools for the value they can provide us and turning to other resources with narrower focuses when necessary.
mike murphy 