You must log in or # to comment.
Look, I know you’ve been using this software for twenty years, and love how it works and what it does, but hear me out: what if we fuck it up? Like, a lot.
Tbf, Microsoft has been doing that since vista
Go on
I’m gonna have plenty of job security. :D
My favorite is scrolling Instagram and then all of a sudden seeing somebody being like," I wanted to bake an apple pie, so I partnered with Meta AI to bake a pie!"
Because Jesus fucking Christ
“Partnered” is such a bizarre word to use to mean, “I read a recipe”. Even if that recipe was regurgitated in the most inefficient way possible.
Imagine people in the past saying shit like, “I wanted to bake a pie so I partnered with Betty Crocker.”
Oh really Brenda? You “partnered” with Betty Crocker? And what exactly did Betty do to help? Oh. That’s right. Nothing. Because she’s not an actual person, just a branded cookbook. Maybe make sure more of the rum makes it in the pie next time.
I keep reading this “nobody wants AI” talking point. Is that really your experience in the real world? I’m surrounded by people who absolutely love that shit and use it for everything.
I’m by no means an AI enthusiast, but acting like people don’t want it, why it’s quite definitely the other way around is a bit weird. I feel like critics are inclined to replay the Blockchain playbook of “haha, a product with 0 use cases they try to push into everything”. But with AI it just isn’t applicable.
There’s so much valid criticism for AI, but saying it’s useless and nobody wants it makes you seem like a lunatic.
You’re right, but sadly I think the reason people like it so much is because of how illiterate they are.
While I’ve found uses for local LLMs (ie I host) and even the bigger chat bots:
I feel the gen “AI” everywhere is just an excuse to collect data since they always use the online models… And if you are either not paying or paying api rates (per token… Or usually million token rate) it’s basically being provided significantly below cost.
Yeah I use chatgpt and Gemini for targeted tasks, data collection and organizing for blog posts and work proposals (yes I verify the data, I know about A.I. hallucinations), and how to’s for tasks I don’t do often or have never done.
Just yesterday I told it “I just bought the PC (linked to product page) I want to install Linux Mint but first I want to disable secure boot, provide a step by step guide to accomplish this” and after about a minute it spit out and detailed guide referencing the manufacturer’s website, their forum, and a third party guide. It even pulled tips and tricks from the forum with things like “tip: users report if you see this screen do this to correct it (link to forum thread)”
I’m not giving these access to my email or calendar or anything like that, once I can have a locally hosted LLM that can access that stuff then sure, as long as I can control my data. Until then I just use commercial LLMs as tools that I don’t fully trust.
are you even saving time, or productivity if you have double-triple check what the LLM even spit out.
When I asit for data it’s generally data that would take hours to find, ready through, accept or reject, then take notes and summarize. When I tell an AI to do it, it takes a minute or two for all the data collection and summarization. Then all I have to do is click the link provided with each piece of data and verify it, which only takes a few minutes generally.
So yes, AI saves me hours of researching and summarizing even with checking the source data.
