Most people in tech aren't in AI, not even adjacently. They're building/maintaining web apps and legacy systems or corporate IT infrastructure or Salesforce instances or Wordpress sites. These fields lag behind the cool kids like OpenAI/Anthropic. They may be sold some AI solution but not seeing it as a threat as they've been sold off the shelf software for decades. It'll take time.
It won't take time. The directors will replace them with 20$/month subscriptions and it will happen suddenly much like what happens with textbook cases of normalcy bias (Wikipedia).
I have gone through more than one cycle of automation already, and I'm not scared by the prospect of another. Most of the work programmers used to do back when I first entered the industry has long since been taken over by software, but there are far more programmers than ever.
If a significant share of the work today's programmers are doing is routine enough that new automation can handle it, so much the better! Human programmers will simply move on up to the next level of abstraction, as we have done before, and get on with the more interesting work of managing the robots.
I think my attitude is less "denial" than "cautious skepticism". I have heard many, many, many announcements of radical world-changing new technologies over the years, but as bright and shiny as the prospects appear to be, the realities invariably feature shortcomings and inconveniences and new problems that people then have to work around and deal with. I see no fundamental reason to expect that new LLM-driven automations will be any different.
