I have never used ChatGPT, Claude or Grok. That seems like a weirdly radical statement to make in 2026. I’ve felt a great unease about hyper-scale LLMs for all the usual reasons: data training, multitude of harms, non-determinism, surveillance, hype and overstated capabilities, environmental and e-waste concerns, economic bubble, data centres and disregard of consent. Oh, and don’t forget about the TESCREAL bit. I’ve been quietly not using AI and know a couple of people in the same boat. But recently I’ve noticed it seems to Just Be Everywhere: Most companies and websites describing themselves as AI-powered/enabling or agentic Slop grenades dumped into chat threads Briefs, content, audits, emails increasingly AI-generated Magic wand features prominently inserted into apps Streams of new products that record and transcribe everything around them, without anyone’s consent “It’s not x, it’s y” Exposure to its output feels inescapable. Using these tools is a precipice I haven’t wanted or felt a need to cross. I’ve never felt left behind without them. The few potentially-beneficial-to-me uses I’ve seen seem trivial, especially when weighed against the known harms and other issues. But maybe the problem isn’t non-use but perception of it. If people in hiring positions believe AI is the answer, then maybe you’ll be passed over? I’ve been looking at full-time positions recently, and it’s amazing how many job descriptions state enthusiasm for or established practice using AI as a requirement. I wrote previously about how AI doesn’t magically complete the cheap, fast or good triangle. Now I wonder if much of the general use of AI is for work where good enough will do. Everyone’s interpretation of where that’s appropriate will vary. One part of me thinks maybe this is fine? Another part wonders if that work even needs to be done? And another part remembers the numerous times I, and colleagues, have wasted hours digesting and troubleshooting hallucinated issues reported to clients by LLMs. Everyone agrees that these tools are non-deterministic nature, but the buck on checking output is frequently passed forward. Maybe “intentional AI use” is the new “we care about your privacy”. I see software as my supply chain and don’t particularly want to pollute it with companies occupying a similar ethical ballpark to adtech, arms manufacturers and gambling. My enthusiasm for AI is roughly aligned with how much I’d want to use products from Meta, Flock and Clearview. But it’s tiring feeling this way about something that’s so in-your-face all the time. I’d like to reach a place where I can either move past the discomfort and/or be at peace with the general use of it. I’m not sure how I do that. Do I keep on not using it and hope to one day feel less discomfort from it? Should I declare AI bankruptcy and nihilistically explore using it? Maybe there’s a perspective shift that would be helpful? Or something else? Suggestions welcome.