Navigating the AI Landscape
Things are moving fast in AI-land. It seems like every week there’s a new model release, each one claiming to be the best we have ever seen.
I’m sick of seeing Twitter threads with the same tired template:
The new model from Y is about to blow your mind! 🤯
10 wild and crazy examples of what people are already building with it >>
It is good to assess the tools you use every now and then (and it is also one of my favorite things to do), but we have to ask ourselves: are we genuinely trying to find the best tool for our work, or are we just distracting ourselves from actually doing the work?
It reminds me a lot of productivity apps. People love to ramble about which task manager is the holy grail that will finally help you get things done, instead of actually doing the work.
Yes, AI models are improving every week, sometimes every day. There are some great advancements. But what matters more is how you use AI, and how to make the most of a given model for your needs.
Take coding, for example. Learning how to code with LLMs is worth investing in, as it is very challenging (some good reading from Simon Willison and David Crawshaw). That’s far more important than saying, “Use the new Claude! It’s amazing,” while someone else replies, “Really? Claude doesn’t even have a 1M context window. Compared to the new Gemini, it’s nothing.”
Learn to use the tools for your work. Do the work. Evaluate every now and then. Repeat.