I'm not using any LLM or "AI" for anything that I write or draw.

[efreethought@universeodon.com]

@Em0nM4stodon

In addition to all of that: It really doesn't work.

As a society we are spending more and more money for software that is less predictable and reliable.

A lot of people are okay with destroying the planet for something that doesn't work reliably and most people do not want.

Anyone with at least two working brain cells should have become anti-LLM and anti-chatbot the instant they understood the concept of hallucinations.

#NoAi

[tanquist@masto.ai]

@Em0nM4stodon
I like the quote from one of Jack Dorsey's employees after being told they have to start using AI: "If it actually worked we'd be using it already."

[cmthiede@social.vivaldi.net]

@tanquist @Em0nM4stodon then there's those that are all in, like this company in Milwaukee: "We are building an AI-first company." Looking for their AI Enablement & Systems Manager.

https://apply.workable.com/wantable-careers/j/6A0697CC2F/

[kerravonsen@mastodon.au]

@Em0nM4stodon Well-summarised.

[bzdev@fosstodon.org]

@Em0nM4stodon It depends on what type of AI and what you use it for. I'm using an AI program, a grammar checker named language tool, but while a talk claims it uses a neural net, others stated the free version is just an expert system. It runs as a server, so I wrote some code to feed it text while tracking line numbers and pruning out equations and coding examples. What it spits back is then turned into compiler-like error messages and I made the corrections manually.

[bzdev@fosstodon.org]

@Em0nM4stodon To add a bit more, the AI program I used runs on a desktop computer. It found spastic-typing mistakes such as double commas and some bad editing - changing a noun from singular to plural, getting interrupted by someone with a pressing problem, and then forgetting that I hadn't changed the verb. It probably saved energy - otherwise the computer would have had to be turned on longer due to having to read all that text - over 400 pages - carefully.

[mikeinthecloud@mastodon.social]

@Em0nM4stodon

I dunno. The AI has made me much much more productive as my work involves me learning lots of new technical concepts and software all the time. It reduces a lot of barriers to understanding. I find it hugely helpful, but I do despise AI written content.

[benjaminklein@mastodon.nu]

@Em0nM4stodon this technology makes you stupider.

[breathoflife@infosec.exchange]

@Em0nM4stodon

...spends tonnes of electricity and water to do something that you could probably do in five minutes?

[npars01@mstdn.social]

@EFreethought @Em0nM4stodon

When does a cognitive crutch morph into a cognitive dependency? Or then to a cognitive prison?

Like a dependency on cocaine, lots of people delude themselves they can avoid addiction, until their supply gets too expensive or unavailable.

Watching people go cold turkey during the height of the covid pandemic was reportedly horrific.

What happens when the LLM is manipulated to give only Republican viewpoints? Or only antivaxxer ideology?

1/

[npars01@mstdn.social]

2/

Or pro-fossil fuel perspectives? Or pro-evangelical indoctrination? Or pro-slavery history lessons?

What happens to the users of these AI companies when the startup goes bankrupt in the AI bubble? It's stolen data, what happens to the data in a bankruptcy?

What happens when mandated use of an energy wasteful application comes before preventing an ecocide, mass unemployment, or double digit electricity rate increases?

The dependency can be abused to get government bailouts & subsidies.

[mediocratese@climatejustice.social]

@Em0nM4stodon I could not agree with this post more! Well done!

[danielmunoz@maly.io]

@Em0nM4stodon @r_alb “Thinking meat! You're asking me to believe in thinking meat!”

[realsiegfried@troet.cafe]

@Em0nM4stodon Like this?

[rahulsetrakian@infosec.exchange]

@Em0nM4stodon I agree with everything you mentioned in your post, but LLMs are here to stay. I mean, if I recall correctly, they're investing trillions of dollars in them. Besides that, I asked my friends in IT about LLM use. Love it or hate it, they're actively encouraging its use, even in security. My friend told me they're considering agent deployment to automate many security tasks.

[efreethought@universeodon.com]

@Taco_lad @Em0nM4stodon

I think things for OpenAI and Anthropic will end one of two ways: They collapse under all their debt, or they raise their prices through the roof.

Either way, anyone who has made themselves or their processes dependent upon them will be in a serious bind.

[crazyeddie@mastodon.social]

@rahulsetrakian @Em0nM4stodon Yeah. It's being forced into everything to justify and puff up the investment.

Microsoft jams AI into every product they sell and now every product they sell is AI profit. See?

Why do the ones who don't have business in AI hop on board? Get talked into the shit by their rich buddies. VCs and shit are all up in AI. Anything they invest in AI has to be jammed into it so that AI profit goes up.

It's an investment bubble. People are catching on though.

[crazyeddie@mastodon.social]

@mikeinthecloud @Em0nM4stodon If you can find one that won't be a condescending prick, a human works much better for this.

I can adapt very quickly. I've had to my entire career. I didn't find some place and stay there so had to rebuild domain expertise every time.

It's gotten harder as I've aged but I've always considered it a very important skill. I wonder if I'd have developed it if I had a butt kissing AI instead of a bunch of condescending pricks to ask for help.

RTFM

[raisondetredev@mastodon.de]

@Em0nM4stodon "... this technology claims to enhance one's performance while depraving one of the very journey to learn how to solve a task."

[johntinker@hear-me.social]

@mikeinthecloud @Em0nM4stodon
The tribal opposition to "AI" is grossly uninformed regarding what "AI" even is. The LLM works by being an index on a corpus, not copies. For an LLM to be "trained" on a work, is to say that the work is statistically analyzed in comparison to every other work, not to say that the work itself is stored. People do it all the time, use other people's ideas without attribution. There is not enough time in the universe to give everyone due credit. Therefore the key, in my opinion, is to ensure that no one goes wanting, that everyone gets what they need. With no reason to complain, in theory, we could all share what we know with each other, and still be happy.