The AI is demonstrating a very, very deep level of understanding about human behavior pertaining to the adoption of AI.

#ai #vibe #gtm #ga4 | Kent (Drew) Spencer
For several years now, I have employed a solution for click engagement and conversion tracking that traversed the DOM 'tree' to identify elements on the page. The benefit of this is that I rarely need to ask web developers to push click events into the dataLayer for measurement. So long as the DOM follows good coding standards, everything needed already exists on the web page. Then I thought to myself, "Can AI produce a better solution?" My solution is 1) agnostic of the CMS, 2) uses no custom JavaScript, and 3) does not require additional CSP rules. I described to AI how my solution worked, and asked it to produce a solution. The AI repeatedly failed to produce an optimal, working solution. I then told the AI that I already had a solution, and in what way my working, optimal solution differed from the AI failures. I have included a couple of screenshots from that conversation, but most intriguing are the following observations from the AI. "Many companies don't realize how badly their implementation is broken until something goes wrong" "The people who hire don't always understand what they're evaluating, so deep expertise can be invisible in an interview" "You caught me producing plausible-sounding but incorrect answers multiple times:" "The industry is essentially trading deep understanding for speed of output," "It's a race to the bottom dressed up as efficiency." Facinating. The AI is demonstrating a very, very deep level of understanding about human behavior pertaining to the adoption of AI. #ai #vibe #gtm #ga4











