The coming decade will reward people who learn how to drive these models.
I have no doubt that this is true.
I also note that you have shown a great example of the use of this mechanism as
an assistive tool, wherein it is the human in the loop (yourself) who decides on the actionability of any given alert. This is absolutely the correct use of these things: by all means, let them filter through the noise, and present what their training and prompting indicates might be signal to you, for you to review. You can then apply your hard-earned experience to make a decision as to priority, and whether or not to follow their suggestions with respect to solving the issue.
This, I think, is exactly right and proper: it is an application of this technology that I can respect, and indeed hope for. You have revealed yourself to be an educated, conscientious, and well-reasoned user of the tool.
What I cannot respect would be the situation where some manager with no actual experience in the art decides to simply delegate the entire process to some LLM, including the critical decision making. While that is clearly not the use in your particular case, I'm quite convinced that that is
worryingly common across the shambolic rollout of these tools.
I'd happily let an employee of mine use this mechanism as an assistive tool- hell, I might eventually give it a shot myself, should I live so long. However, the moment I caught wind of them letting the tool
do their jobs for them, in terms of allowing its nonexistent "judgement" to override theirs (out of laziness or any other misguided motivator), would be the time for a very pointed discussion. The presence of the human in the loop is sacrosanct, to my way of thinking.
The thing of which we must all be aware is that there really do exist layers of management within corporations who know nothing whatsoever of the core competencies of the company, but focus only on "controlling costs" and "improving shareholder value", often allowing the actual product itself (whatever it might be) to suck hind tit. These are the ones to watch out for: those who would be perfectly happy to simply have somebody write a prompt for them that appears to render the assertions of human judgement and creativity expensively moot, to "save the headcount". They are are fools, of course, and always have been. This is not new: I've ridden a couple of startups right into the ground, as the upper management followed only the siren song of the shareholder-value precept (along with the expansion of their own salaries), and ignored the details of the original core business.
Ahh, well. In this brave new world, every company will find their own position to occupy on the human versus machine continuum, as will every employee who retains their job. It will always be an individual, local decision. But it must be a
decision: and in both cases, I will continue to have no respect whatsoever for those who choose to simply delegate everything to a non-human entity, unsupervised, so that they can maximize their income, and their time on the golf course or in the bar. I'm particularly thinking of the author of that article I linked to a few posts earlier as a great example. But I have no doubt that there are thousands of MBAs with VP titles in the C-suites of a thousand businesses who are making the same mistake, even as we speak. Frankly, I expect a lot of situations closely akin to Boeing's "acquisition" of McDonnell-Douglas to play out in the short term.
Hope that helps a little, in the process of thinking about these tools. But once again, your mileage will almost certainly vary.
