The last few years have been full of news about AI. There are some wild mainstream articles for sure, but even in the tech world we see many people experimenting with everything from vibe coding to managing agents to speculation about replacing people with AI agents. Even my boss has used Loveable to build a schedule app my team uses instead of asking our IT group or purchasing a service. It actually works pretty well, too.

I don't think that AI will get rid of a lot of jobs for developers, and I'm unsure of how well the GenAI chat systems with agents and MCP servers will increase productivity. Hopefully, these tools will help improve the quality of code from more junior developers, but that remains to be seen.

Despite all the GenAI hype, there are some great AI successes in the world. This week I ran across an article on digital twins and ML-based image recognition that is an example of where AI-tech that isn't the Copilot/ChatGPT stuff can really shine. In the article, AI technology is examining images and videos for safe practices by workers. This isn't something easily programmed traditionally, where we might hang safety equipment on hooks with sensors that detect is workers have picked them up. Instead, this is a real-time look at people using equipment and trying to detect if they are using safety equipment.

It seems a little big-brother-ish, but as someone who has performed their share of stupid things and skipped using some safety equipment, this could be useful. A real-time reminder broadcast aloud, through a phone, or some other way can reduce the mistakes that humans make while shortcutting procedures that prevent issues. Of course, sometimes we need to skip a step, but those should be rare. Most of the time, we humans just do dumb things because of laziness or a lack of focus.

AI includes a lot of different technologies, and some of them have worked very well. Image recognition applications are everywhere and work very well. Not perfect, but I think they work about as well as humans, especially over time, when there are tedious tasks where humans' attention and focus might wander. Safety-related tasks, like looking for drowning victims in a pool, checking for protective gear, or just catching those errant pets that escape from a yard, are places where AI can perform very well.