Is there ever a concern that those who use AI, and use it well, will eventually become so dependent that they are unable to remember basic principles and strategy without it?

I agree. Many times we go in extreme directions without doing a proper forecast to analyze the effects it may have in the future. I have witnessed many shortcomings such as biased as well as inaccurate information being dispensed without being properly vetted. We must have near full control over our output.

The IDEs were filling in where persons with some technical skillset fell short. Furthermore, they were then made to fill in the gaps since the entire code wasn't written but "spell-checked" (or code-checked rather) for errors. AI today can completely take hold of a situation from a clueless user.

That is a very good point to consider since to get quality output, the input must be of a particular standard. However, in an organization or worse yet an entire industry where the people around seldom know better, how is that recognized? Many times, AI can make mediocre input look slightly less mediocre and therefore enough to be granted a passing grade. Once given the comfort of a "thumbs up," laziness and erosion of standards can kick in.

This is an EXCELLENT parallel Guy. I hope, just like with the calculators, teachers mandate their students to practice without it in their early technical learning years instead of having it as a reference from the beginning without a proper foundation.

The calculator analogy is a good one, although two additional thoughts came to my mind right away:

1) Even with calculators, there has always been a downside that it makes us lazy. Students do not exercise their math muscles enough to be able to do simple calculations as they over-rely on calculators. Common example is when I see people struggle with calculating how much they should tip for different services. That said, this by itself is never a sufficient argument to disregard the advantages of using calculators. Same argument applies for GenAI too.

2) One key difference between GenAI and calculators is that calculators are definitive - the inputs are structured, the output is certain. That is not true with GenAI, which is both good and bad. It just shifts the balance towards being creative and logical in a slightly different way - with GenAI, if you can write logically "airtight" prompts, you'd get better results. What concerns me though is that when people would simply outsource things to GenAI and eventually not cite their sources. Imagine people using GenAI for creative endeavors (e.g., writing stories, movie scripts, songs, etc.) - it is one thing to use GenAI for ideation and motivation, but another thing to claim credit for something generated by a machine in 10 seconds!

Guy, thanks for that brilliant example with calculators.

Particularly love what you said about how calculators have "actually allowed for instruction to move to higher level concepts faster."

That's exactly how I see the use of AI panning out. Intelligent use of AI can help more people elevate their critical thinking skills quicker.

Thanks again for sharing.

Raman Chadha, thanks for raising the two add'l points. I particularly liked #2 with the calculator inputs being structured and the output certain. This is why keeping "humans in the loop" and, as you mention, citing your sources and including your prompt(s). This can lead to what the scientific community has been doing for a century or more with peer reviews.