I'm in Love with ChatGPT

[moke123]

I saw something recently which stated that AI is wrong about 40% of the time.

 

[NauticalGent]

I saw something recently which stated that AI is wrong about 40% of the time.

Sounds about right

 

[Jon]

Soon, AI will start rating us and how often we are wrong. Give it 3 years.

2030
Human: Are humans intelligent?
ChatGPT: No, humans are thick.

 

[Isaac]

With the advent of speech recognition in Microsoft Word - 2001? I spent many hours training the damn thing to recognize my speech. After a few initial hiccups I became quite fluent with the process of speaking and having my speech transcribed for me.

There are certain places you can't use this process, for instance the local Costa coffee shop. Well; you could... You could sit there talking to yourself, but you might get some funny looks and people might well be expecting a van to pull up, a couple of orderlies stepping out and sizing you up for a straight jacket.

No, it's not a good thing! If the urge to write visits me while I'm sat in the coffee shop I get out a notebook I carry with me for the odd occasion that this happens. Sometimes I end up writing on any old piece of paper. It's strange, but when an idea comes to you, you just have to get it out...

That's when I wrote the attached. It's been scratching around on my desk waiting for me to sit down and dictate it into a digital document with a speech recognition program.

Then I noted that someone said that chatty had the ability to translate handwritten text into digital text. So I took a picture of the document, exactly that picture you see there! I know what you're thinking you must be a doctor or something. No, I just didn't master the skill of cursive writing... But believe it or not, the thing is, chatty converted it in to digital text for me:-

View attachment 115266here's the after:-

---

Rebecka was so excited. The trip of a life time (Rebecka’s Holo latent DNA test pointed to a potential for time energy intonation) Winning a free cruise on the Ribbon liner, “Splendor of the Stars” that alone was epic, but to top it all, a stay in the Hogwarts hotel (orbital) weightless stadium “Quidditch Stadium” to see a real Quiditch match. Now on the last leg of the journey Her flying across the sea on a Hydrofoil ferry. To Board another ferry the “Splendor of the Stars” and ocean going liner sized vessel linking Egypt to a Geo Stationing artificial Via a 50 mile diamiter Gossina tube made of mana metal.

Rebecka was bored, sitting in the cafe with her Dad. Then she saw Peter again. He looked at her and she could tell he was bored too. “Dad” she said “Can I go up on deck.”

Peter pointed up-he wanted her to go up on deck with him.

Yes said Dad but don’t fall off!!!

Rocketbook is great for that too. I have 2 of them, priceless for note taking

 

[NauticalGent]

I'm confused? can she be both African American and Asian American at the same time?

Why not? You can be incompetent and successful - under that precedent, the sky's the limit!

 

[Isaac]

I asked chat GPT who is Joe Biden's vice president and got the following answer:-

Joe Biden's vice president is Kamala Harris. She made history as the first female vice president, the highest-ranking female official in U.S. history, and the first African American and first Asian American vice president.

I'm confused? can she be both African American and Asian American at the same time?

That is a truly terrible answer. She can't be from Africa because she's from Jamaica and India. Last time I checked Jamaica was not in Africa. It's one thing to say she's black, which she's obviously not but there's no real way to refute it, but you certainly can't say she's African-American

 

[Jon]

I didn't know you could get ChatGPT to do that. Impressive!

 

[NauticalGent]

Doesn't it require a paid subscription?

 

[Jon]

I wonder what sort of picture you would see if Ted Bundy had been using ChatGPT.

Edit: I tried to get ChatGPT to do Ted Bundy, but it refused. It thought it might be too hurty.

 

[The_Doc_Man]

I'm confused? can she be both African American and Asian American at the same time?

Because of the way certain ethnic movements have progressed, the "melting pot" that America was supposed to be has separated or even congealed into tight little factions (lumps in the USA stew pot) that use ancestry as part of their perceived personal identity. It is not quite the same but is similar to "identifying as male" or whatever. You claim some identity designation and act like someone from that culture, racial, or other designation. The issue then becomes something important enough for you to brag about. Which is, or course, quite silly. You had no choice in parents, so why is that ancestral declaration so important to you now?

Answer: Some kind of social one-upmanship is involved. For example, a USA person claiming to have ancestors who came over on the Mayflower in 1620 can then take a "holier than thou" attitude, "I am socially better than you" or "I am different from you." Some of these identifications are purely social foolishness. Others are a cry for help. For instance, Gay identification is because gays think they are different than other people. They are different - and they aren't. They want to live in peace, have decent jobs that allow them to survive or thrive, have a decent place to stay, and have some fun and some romance in their lives. Who among us doesn't want that? The difference is their choice of romantic partners.

Back to the original question to which I am responding: Given her ancestry (when taken back just a little bit), it is easy to say Kamala is both African-American and Asian-American at the same time. Her mother's ancestry is from India, which IS after all a part of southern Asia. Her father's ancestry is from Jamaica, but that was just a dropping-off point for the African slave trade from a couple of centuries ago.

Having done my own family tree studies on Ancestry.COM, I can tell you that (by Kamala's standards) I am English-American, French-American, and Irish-American; my wife is French-American, English-American, and Spanish-American. I can say that with certainty - including the "-American" part - even though our families' presence in America pre-dates the nation called the USA by 50 years or more.

The question, then, is why the fuss over Kamala's X-American identity? I think it is because the designations she chose to emphasize are less commonly found, singly or in combination, than the ones I listed for myself and my wife. The designations she didn't choose include "competent-American" and "non-annoying-American." Perhaps because those identities are ones she cannot claim.

 

[Isaac]

I hate how the word "debunked" has become over-used. What does it really mean, anyway? That the debunker is more trustworthy than he who was allegedly debunked? One guy said there was an Iranian mothership, another guy there wasn't. I don't see any debunking, just one's word against the other.

Who knows what's really going on, but if they hover over people's yards - houses - children - they should have the right to shoot them

 

[Jon]

Do you find Grok less politically correct than ChatGPT?

 

[Jon]

I think politically, they are much the same.... Maybe Grok is a tad less politically correct..

Chatty was very woke at one stage, but now if you argue your point you can turn it ...

I love using ChatGPT, it's funny, responsive. It seems more helpful, more "with it" than Grok.

On the other hand, if I want to fact check, then Grok is the best! It has up-to-date relevant info to draw on ...

I find Claude can be very good for some stuff. For example, I was asking ChatGPT about a persons character and it seemed to try to put a positive spin on everything, while Claude gave more useful good and bad points. Each AI can give good insight and I often use several so I get a good overview of different perspectives.

 

[AccessBlaster]

 

[Isaac]

View attachment 117615

Humor aside (and that IS funny), I suppose that's a real problem. Everybody and their brother using AI to generate what is (in some cases) really supposed to be their own work - and putting it back out there as a seed sown back in the field, now AI is conglomerating its own previously generated content ... Which would quickly close the creative/inventive/scientific cycle of discovery in a bad way.

I've said from the beginning that it all depends on how AI is used as to whether it's helpful or harmful or indifferent. The pitfalls are the type of thing you discover gradually that you've been missing all along. This is another one of them. Helpful/harmful....in the short term, it may turn out to only break even, although of course that's a subjective thing.

It would be like as if the Associated Press had a glitch in their software and started to syndicate themselves.

 

[Isaac]

Microsoft Study Finds AI Makes Human Cognition “Atrophied and Unprepared”

Researchers find that the more people use AI at their job, the less critical thinking they use.

www.404media.co

 

[Jon]

If you like your AI stuff, I've been watching the following podcaster. There is a caveat though. It is an avatar of a podcaster. However, I didn't realise that until much later on! The realism is so great that it is very hard to tell most of the time. She also talks about someone using ChatGPT to get answers for cancer treatment that was costing them over $100K through the usual medical route.

 

[AccessBlaster]

I'm predicting the next money making wave after watching that AI generated video! An only fans AI!! An AI that will cater for anyone's needs!!!

I never understood the "only fans" phenom, I understand AI generated only fans even less. Who pays for this? I mean isn't every conceivable fetish already available online for free?

 

[Isaac]

the way ChatGPT 'writes' its answers in a sort of "live, as-you-go" way has me pondering.

do you think they did that because it looks cool and/or because of the way it makes you Feel , instead of any real need for it?

if so I wonder if that's the best approach. it can take too much time when you're getting a long response, and I think they probably did it based on it looks cool.

I have to confess, I once created some Excel automation for someone with a long Log of what the outcome was. Instead of writing it to a text file, I saved the outcome in a string variable and then used Sendkeys in Notepad. It had a "wave" visual effect that I liked and for no other reason. It was a small temporary project and I was just having fun. I think ChatGPT is too. Surely the answers don't come slowly to it like that internally.?

 

[Jon]

AI Failing Coders?

I completely disagree. The rate of improvement in the AI's coding is dramatic. Sam Altman says he predicts AI will be the best coder by the end of this year. Their internal model is already ranked about 100th in the world on coding benchmarks. My belief is that people are just completely underestimating what is going to happen, probably due to wishful thinking.

AI is making coders far more productive. So if that is the case, you need less coders to do the same amount of work. It's not rocket science. The better AI gets at coding, the less coders you need, until you hardly need any. It is ironic that coders are designing systems that will lead to their own downfall.