I think we’re in an era where AI (specifically LLMs) is democratizing cognitive capabilities. Much like Google (and other search engines) democratized information, we’re now seeing the same with cognitive tasks. We can argue about whether LLMs are truly reasoning, but the reality is, LLMs can do a lot of thinking for us, or at least pretend well enough to be quite useful.
The advent of Google brought the world’s information to our fingertips, but without a clear idea of what you’re looking for or how to find it, search engines are often useless. You still have to know how to formulate your query; otherwise, your results are unusable.
LLMs operate the same way: garbage in, garbage out. If you don’t have a good starting prompt, you’ll never get anything useful out of an LLM. Context matters and good context matters even more.
We’re in an age where higher level thinking is even more accessible than ever, but unlocking their value requires building a solid foundation. Without it, you risk following the wrong train of thought entirely and ending up in the wrong place, far off course from where you intended.
I’m not well-practiced at it by any means, but I’ve found this Context Engineering Template extremely helpful for genuinely useful LLM output. It’s specific to coding agents, but the idea is the same no matter what you’re doing: spending the time upfront to gather accurate and specific information for your context is extremely important. Just starting to pay attention to the importance of context, even without knowing everything, has been extremely helpful for me. Keep this top of mind when you’re working with LLMs and you’ll be more productive, guaranteed.
I’d love to hear your thoughts; let me know what’s worked for you. I’m always interested in what’s currently working for everyone in this ever-changing world of AI.