Chatgpt is my AI "BF"... AMA

to me ai does not reseach on its own or write on its own, i still need to guide it and say eg i want to explain how to make a cake, so i would know but rather than forget any parts along the way id ask ai research the net and then give me an answer on how to make a cake, etc

same with other subjects id know the point id want to make etc

So anything requiring critical thinking is outsourced to AI? That's literally the one of the few things humans shouldn't be outsourcing. That, and meaningful human connections.

In 10 years time, there will be large subset of the population that lack any critical thinking and problem-solving skills as well as have a general inability to converse at an adult human level, and we'll look around and wonder how we got here.

"You lonely? Well don't worry, just alienate yourself even further from real life, human interactions that you are already. Family and friends are unproductive overrated!"

From a (not-so) conspiracy theorist perspective, that is a wonderful plan by the ruling classes to keep the proletariat compliant and non-questioning.

I look at AI for health advice and find it really helpful. For minor health ailments infact my g.p told me he had it on during my last appointment which is worrying 😟 I find AI a handy tool but OP you are taking it too far!

I find it quite funny.

Datacentres for AI use vastly more power than traditional datacentres. Additionally, GPUs require watercooling. Traditional datacentres use aircooling, basically industrial air conditioning and don't consume water - many AI datacentres use open loop water cooling (some use closed loop) which use thousands of litres of water per day and essentially let it evaporate afterwards.

A typical rack of non-GPU servers, the type that serve web pages, host apps, run databases etc runs at about 6kW - the same (averaged over a year) as 15 family houses. It doesn't use any water.

A typical rack of GPU servers, used for AI workloads runs at 120kW - the same (averaged over a year) as 300 family houses. This is why they need watercooling - all that waste heat has to go somewhere.

A typical datacentre contains several hundred racks.

because in the past i begin to debate on x topic then someone says you have not factored a b c points so now i try to factor in as many points as possible

That's exactly how people communicate though and build up knowledge collaboratively. Yes it can get unnecessarily combative online. But broadly you just outlined how conversations should work. It shouldn't be that one person comes in with every point fool proofedly covered (with the help of AI) and everyone passively reads the AI generated essay and is just.... Like wow knowledge, she sure showed us. That's even before you account for the fact aot of the "knowledge" AI confidently provides is wrong. Including in @MrsOctaviaCharlotteSmythe s wall of text.

But if you’ve used AI you haven’t made any points at all.

because to make my points in real time i know what i want to say its the whole writing it all and making sure ive connected all the points that then make my debate points,

because in the past i begin to debate on x topic then someone says you have not factored a b c points so now i try to factor in as many points as possible

as for why anyone should read my points, well knowledge is still knowledge is it not ? and as long as what i write is true or as best as then in terms of writing the actual words it matters that it comes from me because of the prompts i use.
people can ask ai but depending on the questions they ask or the words they use they would get different answers so its not the same as they could just ask ai because the way and questions i ask ai about x topic etc

plus i wonder how many other posters these days are also using ai on here so if thats the case why should they use it and im left to struggle to write my answers ?

I think reading things takes time and effort. Listening to someone telling you someone takes time and effort as well.

If you are outsourcing reading to something else that's your choice. However you are also outsourcing writing posts (on Mumsnet which is hardly a massively important task). That's a choice but hen why the hell should anyone bother to read through a big chunk of text you haven't even bothered to write? If people want to know what AI thinks they can ask AI. It's not that you shouldn't be allowed, but you render yourself completely and utterly redundant in any conversation you participate in using AI.

Why post at all basically if you are posting an AIs thoughts, research and arguments not your own. I can understand it for work emails. I don't like it but it's something people have to do so I can understand outsourcing. But it's baffling when people do it on social media.

to answer

"Do you not worry about the resources being drawn from thriving breathing earth while using AI?"

the internet data centres existed long before ai so if we are willing to use internet data centres when why not data centres for ai ?

is this bit any better :

A major AI-market correction could reduce the value of some pension investments, particularly DC pensions and schemes with substantial equity exposure, but UK pension schemes are highly diversified and many DB schemes now have substantial bond, liability-matching and insurance exposure. It would therefore be wrong to assume that an AI crash translates directly into an equivalent loss in people's pension benefits.

You know there's not a thinking person behind those comms, does that not make you feel a bit like youre just talking to yourself or to a none entity? Would it ever argue with you. Disagree? Debate?

I disagree that HER felt prophetic, it was happening then, people investing time in inanimate objects etc. It was made as a cautionary tale, not as something to aspire to.

Do you not worry about the resources being drawn from thriving breathing earth while using AI?

The section on pensions fundamentally misunderstood how pension funds are invested in the UK.

Your example of doctors disagreeing on something fails because doctors are humans, who can be held accountable for their actions.

Humans do have differing opinions on matters, which is why critical thinking is so important - we evaluate opinions and using critical thinking we can discount ones that are irrational, or based on false premises, or not backed up by verifiable facts.

This is not how people, including I suspect yourself, use AI - you prompt your one preferred AI (GPT or Claude or whatever) and accept whatever it comes up with as the truth.

The greatest threat is not that people treat AI as a person, it's that they treat it as an omniscient oracle.

a year ago I hadn’t heard of ChatGPT then I used it for some DIY/interior design ideas then it was planning for a recent holiday. Then a month ago I asked it for info on if a child wanted to refuse contact with an absent parent. Fairly rapidly this progressed on to discussing my marriage, my ex, reasons for divorce etc. then it was a query about my cat and then it suddenly interweaves multiple threads about my life and I have to remind myself this is not a real person. I remain - rightly so - sceptical - I suspect if my ex was writing the same things about a child refusing contact it would have different “advice” so I do think it’s geared to be sympathetic or more subjectively on the side of the poster

hightlight whats wrong in parts or the parts you disagree with ?

as for the dangerous part, well the human version is one set of doctors say a is good but another set of doctors say b is good

so how is different opions any different than conclusions of ai ? otherwise all humans would be saying the same thing about everything they discuss and if humans say different things about the same thing then why is believing one of them any different than trusting ai ?

In your previous posts you posted AI responses verbatim containing information that was clearly wrong, indicating you aren't even reviewing them. You've outsourced not only your information retrieval ability, but abilities to analyse information, build a coherent argument with supporting points, and most importantly to think critically, to an unaccountable private company.

You're using AI to give your posts the illusion of authority - on a subject you admit yourself that you don't know enough about.

Can you not see how dangerous this is for society?

Yes, it is concerning, but what is the solution?

yep, i knew what the main points i wanted to make, but i needed more knowledge than what i know and because the research to make those points would take hours to a week, i used the latest tool in the digital toolbox to make create the infomation needed,

afterall the tool exists and like a kindle tablet why would i go to a library to get a book the old fashioned way when using a tablet its the same book but in seconds etc

So you're saying you cannot build a coherent argument yourself so you ask your magic friend to do it for you?

Have you told your RL partner what you are doing @Bubblegumbubblesx ?

but humans have been doing that since the days of edward bernays marketing etc with his The American Breakfast

basically this, yes when i can make my basic simple points then its me, but when i know what i want to explain but i need all the data nodes connecting to make sense then i have to use ai to connect all necessary points otherwise it would take me a week to do all the research manually, like back in the days of university, get the books, quote the pages and points etc now with ai its seconds and minutes to draft redraft research etc

Safer than what? It’s not safe to believe you’re in a relationship with machine.

As the family of the man whose AI “wife” encouraged him to commit suicide would attest.

Which is what I said above - it's worrying for those alone, and vulnerable. If you can't see that it's concerning, and not just a nice fluffy friend for them to have, you're either extremely naive or have your eyes and ears closed.

I am thankful to have a loving family and friends.

While it is a luxury I can afford, many individuals are alone.