Post

It’s fair to disagree. But “has never used these models” is an odd takeaway. I use them every day. The post wasn’t about replacing management with AI. It was about building systems around my own blind spots, which, as a neurodivergent person was something I was doing long before AI.

Long before AI it was notebooks, docs, recurring reminders, sticky notes… whatever worked. Now some of those systems use MCP. The tool changed. The philosophy didn’t.

I suppose that's what I'm getting at. Systems have existed long before LLMs. And they had varying levels of effectiveness. Maybe the point I didn't express well enough is both that LLMs aren't as effective as the article claims compared to these previous methods

While also destroying our environment, access to water, and access to electricity

And if, you were to claim that it is more effective than previous methods at preventing things going missing or lapses in management of your work or life, I would ask you to deeply audit your process. Our management has started to rely on LLMs to manage the release process, and we've never had more

I’m not using an LLM to make management decisions. I’m using it to aggregate information from Slack, Teams, email, docs, and GitHub that I’d otherwise spend a couple of hours collecting by hand. I still read it. I still verify it. I still make every decision.

You cannot possibly know that it is not rewording, overly paraphrasing, or straight up lying to you without directly validating the sources yourself. It does not copy and paste. It reads your documents, and does the best it can to regurgitate it and it has a faulty memory

At my job my direct manager has been using it to summarize teams conversations. And it mixes up products constantly. We nearly had a leak of a major feature for one customer, to their direct competitor (another customer of ours) because the LLM told him it was owned by the wrong customer

I'm upset with your management too. I am in no way high in our company to raise a formal complaint, and i keep this account mostly anonymous/disconnected, but we have a large and very expensive enterprise plan, and the number of code freezes weve had because GitHub has been down is on the rise

That is part of the reason I'm so frustrated to see GitHub pushing this sort of thing. The error rate is going up, downtime rate is going up, and the timing is either hella coincidental or something specific is causing it in y'all's company

Your trauma, bad experiences, and the actions of *other* people are all not reasons to lash out at someone who is not the cause of any of that Using women as a punching bag for your unprocessed feelings is not a good look for you, regardless of your feelings about AI or her workplace. Do better