ChatGPT and Fibaro

[Sankotronic 2,375]

A human asks AI:

"What is the best way to prevent further pollution and destruction of life on Earth?"

 

AI answers:

"Best way of preventing further pollution and destruction is to take homo sapiens out of the equation"

 

BTW - this is not a joke and it is not ChatGPT present intelligence level answer

[Louis 0]

the new era of Genesis is coming , haha

[JessieJim 0]

I suppose it has already changed..

[jgab 2,656]

I have been coding for a while in vscode with the GitHub/Msft copilot plugin. It's like a chatGPT that completes the code you are writing.

I get most help when I'm using languages that I'm not up to date with - it really helps there suggesting libraries and functions I have forgot.

 

It really shines in generating tables and repeating tests etc - write the first two examples and it starts to produce the next entries or tests.

 

Writing QAs and Lua it stills help generating repetitive code - but we don't always agree on suggested coding...

...but doing a net.HTTPClient() call it fills in the rest of the code, and for many of the other QuickApp functions too. Often one have to get back and change some details,

but the skeleton is created and it clearly saves some typing.

For Lua I would say that I have more help from the continuous  Lua linting and language server plugins in vscode that mark out (possible) erroneous code while I type.

[Ening1949 0]

On 1/3/2023 at 10:54 AM, petergebruers said:

What does ChatGPT tell you if you ask it that question? It seems appropriate to ask its own opinion

 

Also, AI is prone to tell you a bunch of lies if you do not pay close attention to its training.

 

I can highly recommend "Robert Miles" channel if you want to know about AI without needing a specialised degree: 

 

 

This one is interesting imho because it discusses the decision of a website called "the chaser" to go behind a paywall because they do not want to "feed" the algorithm with more satire because ChatGPT is already so good at satire. We can of course question the validity of everything they say because, well, satire, ...

 

If you’re curious about what AI like ChatGPT "thinks" or want to dive deeper into multi-AI conversations, a great tool I’ve found is

Please login or register to see this link.

. It lets you chat with multiple AI agents in one place — kind of like watching minds collide. Super useful when comparing answers or digging into complex topics from different AI perspectives.

I think the amount of energy needed to run/train/maintain these systems is significant, but I have never seen anyone quote anything so ----- maybe ask the AI "how much energy do you need to run"

That's a thoughtful point — asking ChatGPT directly is a good place to start, especially since it can clarify a lot about itself, including limitations and training data boundaries.

 

You're absolutely right that AI outputs can be misleading if the prompts aren't precise or if you don’t critically evaluate the response. It's not sentient — it reflects patterns, not truth.

 

Thanks for the recommendation — Robert Miles’ channel is excellent for approachable, balanced commentary on AI safety and theory. Definitely useful for people outside of computer science or machine learning.

 

And yeah, the satire angle is fascinating — especially the idea that something like ChatGPT is "too good" at it. Raises some complex questions about creative value and originality.

 

As for the energy consumption, it's true — training large language models like GPT takes a lot of energy. I believe OpenAI has shared that models like GPT-3 and GPT-4 required hundreds of megawatt-hours during training, and inference (daily use) continues to consume a substantial but smaller amount of energy. It’s definitely a key issue in ongoing discussions about AI’s environmental footprint.

 

It’s a fast-moving space — and skepticism (along with curiosity) is healthy.

Edited April 30, 2025 by Ening1949

[EdwinaWisoky 0]

can i use this images for free?