Do you thank AI after getting a useful response?

[Bicurico]

I used to thank GPT and made my requests longer to include some politeness.

Now I abbreviate my requests to the bare minimum. For instance, when I am analysing code, I will say:
"Explain this code: ..."

If I then need further help, I just paste the next part of the code, without any request and GPT explains the code, knowing this is what I want.

BUT:
When GPT really outperforms my expectation, I will enter a "Thank you" or "That was great".

I am not sure if such feedback goes back into the system at this point of time, but I am absolutely sure it will in a near future, so why not get used to it, in order to improve the behaviour and to train the neural network (which is currently not the case - the neural network, as far as I understand is, is static).

To be, after Mobile Phones, Internet and Smart Phones, AI is the next big thing in disruptive tecnologies. It is so amazing, that I, as a technical versed person, cannot believe how GPT is actually possible. I read several documents and saw videos on how it works, but it's just amazing. One of the reason I find it hard to understand and believe is probably due to the sheer amount of processed data and number of neurons in the network.

[Smokey]

...
Now I abbreviate my requests to the bare minimum. For instance, when I am analysing code, I will say:
"Explain this code: ..."
...

I feel like the LLMs give better responses if you give it as much specific context as possible up front.  So I'll start with stuff like "I am a professional software engineer working on a project for my business.... "

[Sal Ammoniac]

Of course not. Why would I? It's just a machine.

[Sal Ammoniac]

If anyone should be thanking AI it's Nvidia.

https://www.cnbc.com/2024/05/22/nvidia-nvda-earnings-report-q1-2025-.html

[MT]

Do you thank your hammer after hitting a few nails?

I thank the nails for accepting to be hit by a hammer. They newer reply, i assume the nails have a tremendous headache.

[pbernardi]

I just read a study which says that polited questions receive more useful answers from IA. The base is that the LLVM used for training shows better answers for polited questions.

https://arxiv.org/abs/2402.14531

[SiliconWizard]

I just read a study which says that polited questions receive more useful answers from IA. The base is that the LLVM used for training shows better answers for polited questions.

https://arxiv.org/abs/2402.14531

That looks like public money well invested.

[coppercone2]

Do you thank your hammer after hitting a few nails?

I thank the nails for accepting to be hit by a hammer. They newer reply, i assume the nails have a tremendous headache.

no, this is like calling japan to thank a hammer. If you are like right there next to it, ok. but online its mega crazy inefficient, I dunno about bladerunner AI girlfriends, but you should not be thanking the robot librarian from 10000 miles away

i also noticed if its at work, and you really get alot of querys, sometimes I wish it would just end abruptly instead of seeing 1 more "......" . I did my job OK thank me at a meeting or something, I don't need it for a 20 second request.. i feel like im always getting prompted for attention by something when its not necessary.  Actually yeah thank me at the meeting with the bosses not in a conversation no one will ever see. put the ty ammended to the last useful message. don't do another line. stop the god damn blinking

[PlainName]

If you are like right there next to it, ok. but online its mega crazy inefficient

What difference does distance make? Do you not say thanks to someone who has done you a good deed simply because they are online instead of in person? Does it matter if they are in the next street, town, country, celestial body? Of course not - it is interaction, and good manners, regardless of how you communicate.

Whether you should (or can be bothered to) thank an AI is something else, but using distance as an arbiter is conflating two very different things.

[coppercone2]

because you activate how many logic gates to send the message? With billions of people it adds up.

[xrunner]

In case I have forgotton over the years -

THANKS TO ALL HUMAN BEINGS ON THE FORUM THAT HAVE HELPED ME!

Note: I hereby retract all thanks I gave you if you were in fact an AI forum member

[m k]

Of course not. Why would I? It's just a machine.

Learning machine, so you should.

When AI answer start being a regular answer new content start diminishing.
Many current contents are financed by the network, when that start fading so does the new content.

On the other hand, is diversity staying up longer with less targeted AI learning.

[EPAIII]

I don't know as I have never gotten a useful response from an AI.

I will report back when I do.

[PlainName]

because you activate how many logic gates to send the message? With billions of people it adds up.

Wood/tree issue. If saying thanks (or anything) makes that much difference then we'd best turn off the internet - adverts restricted to a static PNG would surely save far more that a single person could waste.

But the problem with AI is that it's already too late. The LLM has burned a HUGE amount just getting to the point where you can send your 9-character "Thank you":

According to OpenAI, GPT-2, which has 1.5 billion parameters, required 355 years of single-processor computing time and consumed 28,000 kWh of energy to train. In comparison, GPT-3, which has 175 billion parameters, required 355 years of single-processor computing time and consumed 284,000 kWh of energy to train, which is 10 times more energy than GPT-2.

I think the trivial energy used in comparison to reinforce the habit of saying thanks is well worth it.

[dietert1]

I am not using AI for entertainment, but more as a search engine. I let it know which of its considerations were useful. Don't know whether it finds that feedback useful or not, though.

Regards, Dieter

[coppercone2]

seriously, it does not need to do MORE work on something useless

https://www.bloomberg.com/news/articles/2024-01-25/ai-needs-so-much-power-that-old-coal-plants-are-sticking-around

that power consumption is no joke. AI is the kinda guy that smogs you after you say thank you