Money quote:
Excel requires some skill to use (to the point where high-level Excel is a competitive sport), and AI is mostly an exercise in deskilling its users and humanity at large.
Money quote:
Excel requires some skill to use (to the point where high-level Excel is a competitive sport), and AI is mostly an exercise in deskilling its users and humanity at large.
There are things that could be done to improve Excel. For instance, fully integrate python and allow it to be used to create custom functions. Then, maybe one day, VBA can ride off into the sunset where it belongs.
Adding Copilot to Excel is not an improvement because Copilot and all other LLM based platforms frequently barfs out totally incorrect information about how to do something in Excel.
“You do that using <X> formula.”
No, I can’t, you worthless pile of shit because THAT FORMULA DOESNT EXIST.
Integrated python scripts in excel sounds like a malware developers dream.
I mean… Yeah, but the same can be said for VB?
Especially since VBA can make calls to the Windows API directly and through that avenue do all kinds of funky things to your system.
Surely there’s some sort of sandboxing that could be done? Like start by disallowing sys calls entirely
Definitely, but sandboxes can be escaped, and you can’t protect everything via sandbox. Apparently its all cloud anyway, but if it were local and sandboxed, there are still exploits like rowhammer and spectre that may cause further risks.
Its taken years to get browser sandboxes to where they are, and even they get broken every so often.
And a nightmare for an application developer told to make some app with a spreadsheet for a database scale
Could result in some very cursed codebases.
“We dont use git, we just update the excel spreadsheet”
I’ve worked at places where they did that anyway lol
That’s just called Access
Is that creepy thing still alive?
They foresaw that. That’s because python on Excel doesn’t run locally, but in the cloud and then returns the result to you: https://support.microsoft.com/en-us/office/introduction-to-python-in-excel-55643c2e-ff56-4168-b1ce-9428c8308545
That’s the worst possible solution to that problem. Why can’t they just develop their own script that’s Turing complete but doesn’t have any system calls?
Or just use Lua compiled without the system calls. This is done by many video games. İt’s 2025, there is no need to create new domain specific languages.
Or use embedded Lisp, like all the cool kids.
That’s even worse!
Still sounds like you’d be shipping your data to the cloud, where it can be exfilled from there.
Would potentially be a great phishing tool, just need to trick someone into putting sensitive data into a precooked excel file, and it gets exfilled.
Currently only for business customers which probably use OneDrive or SharePoint anyways, so it’s not that they need that to exfiltrate data. But for a phishing/hacking attempt? There are probably some nice possibilities.
Fair point. Of course that’s already a problem with Excel. It would probably have to be disabled by default just like VBA macros.
Yeah, no doubt.
Having access to visual basic is dangerous enough, let alone Python
Yea like what? It’s been a big increase in workflow for me.
Increase in workflow? Like there are more steps to perform the same task? Because workflow isn’t work volume or units if output. It’s the process that gets the work done.
Did the increase in “workflow” get you more money or more work for the same money?
Like I spend less time trying to build formulas and I can create formulas and tools I normally wouldn’t with it because I can have a conversations about what I want to do and it provides suggestions.
How did it improve your workflow?
I have a conversation about what I want to do and it provides suggestions and formulas or tools I wasn’t even aware of that stopped up productivity.
The manual exists though.
You know it’s easier, efficient and better to have a conversation with an AI rather than read an entire manual that won’t contain the thing you need right?
You know, all the things it’s advertised to do.
Only to find the “AI” hallucinating functions that won’t work. Or won’t do the thing you were told they do.
Lol you shared your personal experience and got downvoted… lmao even
Lemmy is propaganda against AI at this point. Not sure what paid for it but it has all the markers. Feels like being in the comment section of ny post articles.
Same energy as talking online about immigrants, nuclear energy or marvel
It’s using a community to post toxic and dystopian articles over and over again. Lemmy technology communitys are extremely vile. Not sure why it happened but it’s turned toxic
There isn’t propaganda against AI, it’s totally grassroots because companies are overselling it.
No it isn’t. There is 100% propaganda and media targeting communities to spread it.
The Gap between peoples opinion towards AI in everyday life vs people on Lemmy is massive and a good indicator that Lemmy is astroturfed to be toxic towards it. People who are influenced cannot see it, outsiders can though. It’s like seeing right wingers talk about immigrants. They’ll never be able to see how their news and media influence them. That is their truth and it’s as true to them as hate towards AI is towards lemmings in places like c/technology
Look at the articles posted, the headlines, the appeals used, the comments. It has all the markers of an Astro turf campaign.
Lemmy is pretty consistent with the people I know IRL in terms of opinions on AI.
Not where I am. I haven’t met anyone irl that has any spite with AI. They think it’s interesting. Have tried it a few times. But nobody is out there saying fuck AI.
I fed AI all my Lemmy posts and asked it for a portrait of the artist. Not bad, down to my 6 fingers.
That was the initial impression of it. Now that we’ve had more experience with it and learned that it can’t be relied on, perception has changed. It is oversold and the costs are not worth what we are getting out of it.
No, I’d definitely agree that AI sentiment overall is pretty negative. I am not such a hardliner, but they are definitely out there. I don’t see it as astroturfing at all, to even suggest this is ironic because LLMs are the ultimate astroturfing tool. The institutions capable of astroturfing do support AI and are using it. What institution or organization are you accusing of anti-AI astroturfing, exactly? This question requires an answer for that claim to be taken seriously.
IMO the problem is not LLMs itself, which are very compelling and interesting for strictly language processing and enable software usecases that were almost impossible to implement programmatically before; the problem is how LLMs are being used incorrectly for usecases that they are not suited for, due to the massive investment and hype. “We spent all this money on this so now we have to use it for everything”. It’s wrong. LLMs are not knowledge stores, they are provably bad at summarization and as a search interface, and they should especially not be used for decision making in any context. And people are reacting to the way LLMs are being forced into all of these roles.
People also take strong issue with their perceived violation of intellectual property and training on copyrighted information, viewing AI generated arts as derivative and theft.
Plus, there are very negative consequences to generative AI that aren’t yet fully addressed. Environmental impact. Deepfakes. They’re a propaganda machine; they can be censored and reflect biases of the institutions that control them. Parasocial relationships, misguided self-validating “therapy”. They degrade human creativity and become a crutch. Impacts on education and cheating. Replacement of jobs and easier exploitation of workers. Surveillance.
All of these things are valid and I hear them all from people around me, not just on the internet.