I work in social work; I would say about 60 percent of what I do is paperwork. My agency has told us not to use LLMs, as that would be a massive HIPPA nightmare. That being said, we use “secure” corporate emails. These use Microsoft 365 office suite, which are copilot enabled. These include TLDRs at the top, before you even look at the email, predictive texts… and not much else.

Would I love a bot who could spit out a Plan based on my notes or specifications? absolutely. Do I trust them not to make shit up. Absolutely not.

This is all of tech right now.

My company is doing small trial runs and trying to get feedback on if stuff is helpful. They are obviously pushing things because they are hopeful, but most people report that AI is helpful about 45% of the time. I’m sorry your leadership just dove in head first. That’s sound like such a pain.

My boss has mentioned AI once, and it was when I was asking for guidance on how to take over a twice annual presentation he used to run. He said I could just get the LLM to generate the presentation then edit it as needed.

I did not do that. I just made the damn thing myself because, by him saying that, I realized he had no fucking clue how to do any of this so I didn’t have to either.

Corporate has introduced an LLM we’re supposed to be able to use in lieu of asking an HR representative basic questions, like vacation policy and how to get resources for open enrollment. It… barely works. If you happen to use the magic words, it’ll lead you to a PDF of the policy you’re looking for. I just call or email someone in HR. It’s better to build rapport and awareness between home office and the satellite offices, anyway. Makes things smoother when there’s an actual issue to deal with and people help people they like, even just a little, faster.

It has absolutely no involvement anywhere which is good.

I teach a state university in the US, and AI use is encouraged for tasks that LLMs are actually good at. Generating wrong answers for multiple choice questions, formatting latex documents, writing excel formulas, etc. We’ve also used it during some brain storming sessions to generate ideas and check for any obvious holes in our ideas.

Fortunately not that bad but the people who are using it do get praised while being a massive burden on everyone who has to review the code or worse, documents.

I did see one clever usage to basically replace most front end devs with AI:

• client asks LLM question about our data.
• LLM generates code to query our data and display charts
• client sees results

All runs with just mistral:8b. Very flexible solution compared to having front end devs constantly iterate on a UI monstrosity meant to serve every single clients needs. Of course this assumes the AI is writing the query correctly.

I’m exactly in the same boat.

I’m a developer and my boss has no idea about developing. If i take long with something, he consults chatGTP and gives me the solutions chatGPT came up with.

It’s exhausting…

@rain_lover@lemmy.ml why don’t you like it?

semi-toxic and stupid. I got the fucking AI cert they wanted me to, but instead of getting hooked up with clients to do that, I’m doing the fucking test automation since the start of this awful career.

Thankfully it hasn’t made it into my workplace yet. We have a quarterly newsletter that someone tried to submit ChatGPT slop to. It was immediately identified and rejected by the rest of the horticulturists. My bosses are the kind of people who only talk about plants in Latin so there’s a big institutional focus on getting the right information from primary sources and then using multiple layers of expert review.

However, we’re facing massive budget shortfalls over the next few years and I doubt that will get any better if the economy crashes. Outside of installing/maintaining plants, the bulk of the job is intellectual and creative labour that the public isn’t even aware of. I can absolutely see my workplace hollowing out the job and not hiring based on expertise. Instead of five people with scientific degrees debating a space for an hour, at some point it’s going to be someone who hasn’t seen that space feeding words they can’t pronounce into an LLM that doesn’t understand what space is. On paper it will look great for the metrics admins and other departments track. In practice it will immediately ratfuck everything that makes our urban forest function and drive away the really rare pool of overqualified people we have.

I use gemini to generate code. None of it works. It gets confused between GA and beta features, enhancement requests, and doesn’t give any extra credence to official documentation, which is copious. So its basically like an intern producing little value and taking up my time by requiring constant supervision. Oh and AI generated firewall rules took down a major service for us recently. Good times.

I can only speak for my use of it in software development. I work with a large, relatively complex CRUD system so take the following as you will, but we have Claude integrated with MCPs and agent skills and it’s honestly been phenomenal.

Initially we were told to “just use it” (Copilot at the time). We essentially used it as an enhanced google search. It wasn’t great. It never had enough context and as such the logic it produced would not make sense, but it was handy for fixing bugs.

The advent of MCPs and agents skills really bring it to another level. It has far more context. It can pull tickets from Jira, read the requirements, propose a plan and then implement it once you’ve approved it. You can talk it through, ask it to explain some of the decisions it made and alter the plan as it’s implemented. It’s not perfect but what it can achieve when you have MCPs, skills, md files all set up is crazy.

The push for this was from non-tech management who are most definitely driven by hype/FOMO. So much so they actually updated our contracts to include AI use. In our case, it paid off. I think it’s a night and day difference between using base Copilot to ask questions vs using it with context sources.

My company has its own internal system that leverages GPT models. So yeah it’s definitely becoming prevalent. My boss specifically is always toting it as some amazing tool, all the other business wankers think it’s great, my direct colleagues have taken to letting it summarize their meeting notes and clean up their emails before sending.

All in all, it’s a grim situation. From the execs who think it will change the workforce for the better, down to the average workers who are losing their ability to proofread.

I’m beyond the point of airing my grievances and just find myself idly standing by hoping enough of them get owned by it somehow and become weary of it. I don’t think that’s gonna happen, but a guy can dream.

The most technically illiterate leaders are pushing the hell out of using for things that don’t make sense while the workers who know what they are doing are finding some limited utility.

Out biggest concern is that people are going to be using it for the wrong stuff and fail to account for the errors and limitations.