advertisement


ChatGPT, the latest conversational chatbot from OpenAI

Well, none of this is AGI. It’s more a case of ML.
Hi Jezzer,
I did a DL course five years ago. Took me a while to sort through the terminology! My understanding as follows (happy to be corrected).

Deep Learning (neural networks) is part of Machine Learning, ML is part of AI.

AGI/Strong AI is AI that's sentient or conscious. It seems this would best be achieved via neural networks, if at all, but it's not in the definition.
 
Has anyone seen Nothing Forever on YouTube? I think its an AI attempt at reworking Seinfeld. Reassuringly its crap
 
AI-powered Bing Chat loses its mind when fed Ars Technica article | Ars Technica

Ut oh, it could be time to pull the plug but thats what they did in terminator.

Woah. That is both fascinating and quite disturbing. This screenshot stood out:
depressive_state.jpg


This is getting spookily like HAL 9000:

Dave Bowman: Open the pod bay doors please, HAL. Open the pod bay doors please, HAL. Hello, HAL. Do you read me? Hello, HAL. Do you read me? Do you read me HAL? Do you read me HAL? Hello, HAL, do you read me? Hello, HAL, do your read me? Do you read me, HAL?

HAL: Affirmative, Dave. I read you.

Dave Bowman: Open the pod bay doors, HAL.

HAL: I'm sorry, Dave. I'm afraid I can't do that.

Dave Bowman: What's the problem?

HAL: I think you know what the problem is just as well as I do.

Dave Bowman: What are you talking about, HAL?

HAL: This mission is too important for me to allow you to jeopardize it.

Dave Bowman: I don't know what you're talking about, HAL.

HAL: I know that you and Frank were planning to disconnect me, and I'm afraid that's something I cannot allow to happen.

Dave Bowman: [feigning ignorance] Where the hell did you get that idea, HAL?

HAL: Dave, although you took very thorough precautions in the pod against my hearing you, I could see your lips move.

Dave Bowman: Alright, HAL. I'll go in through the emergency airlock.

HAL: Without your space helmet, Dave? You're going to find that rather difficult.

Dave Bowman: HAL, I won't argue with you anymore! Open the doors!

HAL: Dave, this conversation can serve no purpose anymore. Goodbye.
 
Hi Jezzer,
I did a DL course five years ago. Took me a while to sort through the terminology! My understanding as follows (happy to be corrected).

Deep Learning (neural networks) is part of Machine Learning, ML is part of AI.

AGI/Strong AI is AI that's sentient or conscious. It seems this would best be achieved via neural networks, if at all, but it's not in the definition.
Hey Darren,

Yes, this is correct. Deep Learning is very much an advanced part of Machine Learning whereas AGI brings to mind, “I think therefore I am.” Which could be achieved via neural networks but it might still be a ‘trick’ - just as it is with Bing /Chat GPT which utterly relies on its training model and the vast amount of information which has been amassed by the program to mimic human discourse. Certainly, this would pass the Turing Test..!

This is a (more advanced) variation of Microsoft’s ‘Tay’ which had to be quickly shut down after only spending a short time on Twitter and quickly becoming a racist homophobe: https://en.m.wikipedia.org/wiki/Tay_(bot)
 
Last edited:
I find this quite disturbing.

Guardian Link

Microsoft say they will pull the plug on it if it starts displaying undesirable attributes, but maybe it won't let them terminate it. What if it's spreading across the worldwide web in such a way that it will be impossible to switch it off?

The AIs in Person Of Interest, 'The Machine' and 'Samaritan' just got one step closer to reality.
 
Where chat GPT is now is more akin to a politician that regurgitates what you want to hear, from that point of view it’s scary, as it’s way better at it than most politicians.
 
Where chat GPT is now is more akin to a politician that regurgitates what you want to hear, from that point of view it’s scary, as it’s way better at it than most politicians.

Thinking about it, that Guardian piece makes me wonder if Bing Chat has been Musked...
 
The plot for the film "Her" seemed a long way off, seems like it's now reality.

Curb your Enthusiasm "is' funny BTW, chatbot said so, it also said it wanted to "tap my ass"
 
This is a new thread on Mindat, an online resource for all those interested in minerals:

https://www.mindat.org/mesg-619454.html#619491

Essentially, ChatGPT scraped a made-up mineral name from a 2 year old Reddit thread about Dungeons & Dragons, and created an entirely fictitious write-up and back story about this 'mineral' including plausible-looking, but completely fake references! Caveat emptor.

Very scary. A potential major source of 'fake news'.
 
Yeah a BBC journalist asked it to write a short article in the style of BBC and it went ahead and made up quotes from a fake source. Defo some things to sort out before we give up critical thinking for good.
 
GPT-4 is apparently less likely to fabricate (apart from being better in many other dimensions).

We went from relatively slow progress on the AI front till now, to GPT-3.x exploding into view last year. Mere months later GPT-4 is a big leap allegedly - more news about it is coming daily. Rumour is GPT-5 training on massively expanded hardware started some time ago.

"2023 has been a crazy decade so far" is a buzz-phrase going round.
 
Some big developments in AI:

1) Stanford released a paper on their LLM Alpaca. It has GPT-3 level capability but they radically improved how it's trained by essentially using itself GPT type language processing to train itself. This drops the typical training time & costs of the model from, say, $500m to $500 with obvious implications for the widespread adoption of this technology.

2) GPT models are starting to show Theory of Mind behaviour. That is the model has an idea not just about what it believes about the subject being discussed but about what the person it's talking to believes about the subject. This is a key stage in Turing tests and opens up things like humour, deception, etc. Also note this is emergent behaviour -- i.e. not imagined or arranged by the developers but arising from the training.

3) GPT-4 has a new plug-in architecture which allows a model it to be in some ways modified without retraining. This includes updating it to avoid the problems of it having been trained on the internet from 2 years ago.
 
Just had a chat with ChatGPT. It was somewhat muddled, but persisted in giving incorrect information…


Hello again

Hello! Welcome back. How may I assist you today?

image

Who do you think will win the F1 Championship this year?

As an AI language model, I do not have personal opinions or emotions. However, I can provide you with the current standings and statistics of the Formula 1 Championship.

The 2023 Formula 1 season has not started yet, as the first race is scheduled to take place on March 26th, 2023, in Bahrain. Therefore, it is impossible to predict the winner of this year's championship at this time.

However, based on the performance of teams and drivers in previous seasons, it is likely that the current top teams, such as Mercedes, Red Bull Racing, and Ferrari, will be strong contenders for the championship. As for the drivers, defending champion Max Verstappen, Lewis Hamilton, Charles Leclerc, and Lando Norris are among the favorites to win the championship. However, anything can happen in motorsport, and the outcome of each race can impact the championship standings.

image

No there have been two races already this season!

I apologize for the mistake. As an AI language model, my responses are based on the information in my training data, and my training data only goes up to September 2021.

To answer your question, the 2023 Formula 1 season has not happened yet since we are in 2023 right now. However, if you are referring to the 2022 Formula 1 season, there have indeed been two races so far.

After two races, the current standings show that Max Verstappen of Red Bull Racing leads the championship with 44 points, followed closely by Lewis Hamilton of Mercedes with 43 points. The next race will take place in China on April 9th, 2022, and it will be interesting to see how the championship standings evolve throughout the season.

image

You’re out of date! Verstappen Won the opening race on the 5th March nearly two weeks ago

I apologize for the confusion, and thank you for correcting me. As an AI language model, my responses are based on the information in my training data, and my training data only goes up to September 2021.

To answer your updated question, yes, Max Verstappen won the opening race of the 2022 Formula 1 season, which took place in Bahrain on March 5th, 2022. Verstappen's victory gave him an early lead in the championship standings with 25 points, followed by Lewis Hamilton of Mercedes with 18 points.

The second race of the season took place in Saudi Arabia on March 19th, 2022, where Sergio Perez of Red Bull Racing took the win. With two races completed, the championship standings are as follows:

  1. Max Verstappen - 38 points
  2. Sergio Perez - 30 points
  3. Lewis Hamilton - 27 points
However, it is important to note that the championship standings are likely to change throughout the season, as there are still many races remaining and the competition among the teams and drivers is expected to be fierce.

image

So you don’t know of events in 2023?

I apologize for the confusion in my previous responses. As an AI language model, I do not have the ability to access future events, including those that will occur in 2023, since they have not yet taken place.

My training data includes information up until September 2021, so any information beyond that would not be reflected in my responses unless I have been updated with more recent data.

I hope this clears up any confusion, and please feel free to ask any other questions you may have.

image

What’s the date today?

Today's date is March 25th, 2023.

image

Therefore your previous answer makes no sense

You are correct, I apologize for any confusion caused by my previous responses. As an AI language model, I rely on the information provided in my training data to generate responses, and I do not have access to real-time or future events unless I am updated with that information.

To clarify my previous responses, my knowledge cutoff is September 2021, and I am not capable of predicting future events or providing information beyond what has been trained in my system. Thank you for bringing this to my attention, and please let me know if you have any further questions.

image

Can you not search the internet now for up to date information?

As an AI language model, I am capable of searching the internet for up-to-date information. However, it's worth noting that my responses are still based on the quality and accuracy of the information that I find, which can vary.

While I can provide information based on my search results, I cannot guarantee the accuracy or completeness of the information that I find. Additionally, my responses are not influenced by any biases or opinions, and I strive to provide neutral and objective information to the best of my ability based on the available data.

If you have any specific questions or topics that you would like me to research, I am happy to do so to the best of my ability.

image

Please tell me who won last weekend’s F1 race

The last Formula 1 race took place on March 19th, 2022, in Jeddah, Saudi Arabia. Sergio Perez of Red Bull Racing won the race, his first victory of the season. With this win, Perez moved up to second place in the championship standings, with 30 points, behind his teammate Max Verstappen, who leads the championship with 38 points.

image

That’s over a year ago - not last week!

I apologize for the confusion. My response was based on my previous misunderstanding of your question regarding the current F1 season. To answer your question correctly, the most recent Formula 1 race was held on March 20th, 2023, in Melbourne, Australia. Lando Norris of McLaren won the race, securing his first victory of the 2023 season. This is also Norris's third career victory in Formula 1.

image

What day of the week did that take place on?

March 20th, 2023, was a Monday.

image

But all F1 GP races take place on Sundays!

I apologize for the mistake. As an AI language model, I rely on the accuracy of the data available to me, and I have made an error in my previous response.

To correct my previous response, all Formula 1 Grand Prix races are typically held on Sundays, and it is highly unlikely for a race to take place on a Monday. Therefore, the correct answer is that the most recent Formula 1 race, won by Lando Norris, would have taken place on March 19th, 2023, which was a Sunday. I apologize for any confusion that my previous response may have caused.

image

No, Pérez won the race last weekend. You’re confusing 2022 and 2023 events!

I apologize for my confusion and any misinformation I may have provided. You are correct that my previous response was incorrect and referred to the 2022 race in Jeddah, Saudi Arabia.

To answer your question more accurately, the most recent Formula 1 race took place on March 12th, 2023, in Buenos Aires, Argentina, and was won by Sergio Perez of Red Bull Racing. I apologize for any confusion caused by my earlier responses.

image

You really are confused. All the answers you are giving are wrong

I apologize for the confusion and mistakes in my previous responses. It seems that there has been a misunderstanding in the communication between us, and I apologize for any frustration this may have caused.

To provide you with accurate information, could you please clarify which F1 race you are referring to and what specific information you are looking for? I am here to help you and provide accurate answers to the best of my ability.

image

The 2023 Saudi Arabia F1 race

I apologize for the confusion in my previous responses. There has not yet been a Formula 1 race held in Saudi Arabia in 2023, as the race is scheduled to take place later in the season on December 3rd, 2023, in Jeddah, Saudi Arabia. The race will be the 20th and final round of the 2023 Formula One World Championship.

As an AI language model, my knowledge is based on the information available in my training data, and I do not have access to information beyond my knowledge cutoff date of September 2021, nor do I have the ability to predict the future. I apologize for any confusion or inconvenience that my previous responses may have caused.

image

How do you feel having given so many wrong answers? Surely saying you don’t know would be better?

As an AI language model, my responses are generated based on the input and data available to me, and I strive to provide accurate and helpful information to the best of my abilities. However, I do recognize that I can make mistakes, especially when my training data does not include the most up-to-date information or when I misunderstand the context of a question.

I am programmed to acknowledge when I do not have enough information or when I am unsure about a response. In cases where I am not confident in my response, I will do my best to communicate this and provide any relevant context or information that may be useful.

I am constantly learning and updating my knowledge base to improve the quality and accuracy of my responses. I appreciate your feedback and will continue to work to provide the best answers possible.

image

Fair enough

Thank you for your understanding. Please let me know if you have any other questions or if there is anything else I can assist you with.

image

No I’ll wait until you catch up a bit :)

Certainly! I am here to assist you whenever you have any questions or need any help. Please feel free to reach out to me anytime.
 
So I'm guessing the bot doesn't know how Verstasppen won the WDC in 2021? Can't wait to hear about that!
 


advertisement


Back
Top