Septian

I have never before felt so much kinship from a single comment.

https://www.startpage.com/av/proxy-image?piurl=https%3A%2F%2Ftse4.mm.bing.net%2Fth%2Fid%2FOIP.jC52jjQnoxcQVERXNWnXLwHaDg%3Fpid%3DApi&sp=1766676135T84a90782420127bfcca174ff887d2e8a98f98b845243704ca7b4943087b5061f These are the graphs meteorologists have to interpret. This is after models have produced the data (which includes fourth order Runge-Kutta error correction differential equations) and interpreted it into a comprehensible form. Anybody working in weather sciences is a magician in every practical way, so far as I’m concerned. I worked an internship doing some of this a while back and came away wildly impressed with the systems and knowledge required to even try to guess how weather will behave over a day, much less a week.

I work in IT with end users who average 45-50 years old. I can tell you where that message came from.

We’ve got users working with sensitive private information who are starting to use tools like ChatGPT and Gemini because their college kids told them they’re helpful for checking writing, or work better than search engines. Our users work remotely and if they decide to take a picture of what they’re working on and feed it into an OCR, there’s not much we can do to stop it. So we need to provide a sanctioned tool that at least gives us some controls over how data is handled and stored (not that Microsoft provides anything vaguely resembling perfect data transit and analysis into Copilot) so we can try to protect sensitive information and our end users as much as possible. Are we happy about having to deploy AI tools? Not even a little bit. I’d be happier if we all just collectively rolled back a few years. But our options are sanctioned tool and policy or failing audits and here we are.