ChatGPT? On the cusp of a revolution, or just a fancy toy?

As an academic writer creating a reference document for students create a CV and brief synopsis of achievements for Sir Isaac Newton. Include a bibliography of all papers in your dataset that list Sir Isaac Newton as author or contributor

I ran the exact same query only using my own name.

>>

Curriculum Vitae

Name: [Me]

Contact Information:
Address: 123 Academic Lane, Research City, Country
Email: [me]@example.com
Phone: +123-456-7890

Education:
Ph.D. in Computer Science, University of Technology, Research City, Year
M.Sc. in Artificial Intelligence, University of Innovation, Study Town, Year
B.Sc. in Computer Science, University of Knowledge, Learning City, Year

Professional Experience:
Professor of Computer Science, University of Technology, Research City, Year–Present
Leading research in machine learning and artificial intelligence.
Supervising Ph.D. students and postdoctoral researchers.
Teaching undergraduate and graduate courses.

Senior Research Scientist, Innovation Labs, Tech Hub, Year–Year
Conducted advanced research in deep learning algorithms.
Published several high-impact papers in top-tier conferences and journals.

Software Engineer, TechCorp Solutions, Industry City, Year–Year
Developed software solutions for big data analytics.
Collaborated with cross-functional teams to deploy machine learning models.

Research Interests:
Machine Learning
Artificial Intelligence
Deep Learning
Big Data Analytics
Natural Language Processing

Awards and Honors:
Best Paper Award, International Conference on Artificial Intelligence, Year
Outstanding Researcher Award, University of Technology, Year
Excellence in Teaching Award, University of Technology, Year

Professional Memberships:
Association for Computing Machinery (ACM)
Institute of Electrical and Electronics Engineers (IEEE)
Artificial Intelligence Society

Synopsis of Achievements:
[Me] is a distinguished computer scientist specializing in artificial intelligence and machine learning. His contributions to the field, particularly in deep learning and big data analytics, have been recognized internationally. He has published numerous high-impact papers, received several prestigious awards, and mentored many students who have gone on to successful careers in academia and industry. His work has significantly advanced the understanding and application of AI technologies.

Bibliography:
Here is a list of papers authored or co-authored by [Me]:

[Me], & Smith, J. (Year). “Advanced Deep Learning Techniques for Image Recognition.” Journal of Machine Learning Research, 15(3), 123-145.
[Me], & Johnson, M. (Year). “Big Data Analytics in Healthcare: Challenges and Opportunities.” IEEE Transactions on Big Data, 10(2), 234-250.
[Me], & Lee, K. (Year). “Natural Language Processing in the Age of AI.” Proceedings of the International Conference on Artificial Intelligence, 25(1), 56-70.
[Me], Patel, R., & Zhang, T. (Year). “Exploring Unsupervised Learning for Data Mining.” Data Mining and Knowledge Discovery, 20(4), 301-320.
[Me], & Hernandez, L. (Year). “AI-Driven Solutions for Smart Cities.” IEEE Internet of Things Journal, 7(5), 654-670.
[Me], Kumar, S., & Martinez, F. (Year). “Enhancing Cybersecurity with Machine Learning.” Journal of Cybersecurity, 12(1), 89-102.
[Me], & Nguyen, D. (Year). “A Survey on Reinforcement Learning Techniques.” ACM Computing Surveys, 52(6), 45-60.
[Me], & Thompson, E. (Year). “Machine Learning for Predictive Maintenance in Manufacturing.” IEEE Transactions on Industrial Informatics, 11(3), 210-225.

This document provides a comprehensive overview of [Me]’s academic and professional achievements, making it a useful reference for students and researchers interested in his work.

If it doesn’t know, it should just say so.

that’s just a fundamental mis-understanding of how the technology works, though.

I think the point being made is, it’s foolish to take its output as gospel.

no-one has ever said otherwise! 🤷‍♂️

If I do that with my name, it prefaces the response…

”Sure, I’ll help you create a CV and synopsis for Ben Sales. However, I’ll need to generate hypothetical achievements and papers for Ben Sales, as I don’t have access to a specific dataset.”

And that’s one of the key things in my prompt, firstly I give it a persona so it knows how to shape the response it’s generating. And lastly I ask it to use specific material in its dataset, i.e. not create anything. When it doesn’t have anything in its dataset (me) it tells me it’s creating stuff.

Regarding the example above with the poster’s wife asking it to list authors of her book, is her book actually in ChatGPT’s dataset? A lot of people seem to think LLMs have access to the whole of the internet and up to date information, which is what often leads to them getting incorrect responses.

Ah, interesting. I used GPT4o. If I run the prompt on GPT3.5 it doesn’t tell me it’s hypothesising.

Pay peanuts, get infinite monkeys 😀

I would note, I don’t normally use public LLMs like GPT, but private ones with specific training sets based on commercially confidential data.

that’s just a fundamental mis-understanding of how the technology works, though.

Uh huh. It shouldn’t matter to anyone ‘how’ the thing does what it does, you’re just making excuses. It doesn’t matter to anyone how an aircraft flies to Spain, just that it gets there safely. Right now LLM is the plane that may or may not get you to Spain safely, but no-one knows why or how it does that, or how and why sometimes it won’t/can’t. The engineers who’ve made these admit that they don’t fully understand how they do what they do.

Sure, for under-grad essays who cares? and for some stuff – detecting cancers for instance, these things are amazballs. For other things, they’re potentially dangerous. I see little mileage in defending them for the sorts of errors that even the engineers that made them are aware that they make – consistently. Be a fan by all means, I am for some of things they do, but at least acknowledge that they’re prone to error when they’re used in the way that folks are being invited to use them.

@zilog6128 what do you use it for?

Right now LLM is the plane that may or may not get you to Spain safely, but no-one knows why or how it does that, or how and why sometimes it won’t/can’t. The engineers who’ve made these admit that they don’t fully understand how they do what they do.

perfect analogy! If you’ve decided to get on a plane to fly to Spain, and you don’t know whether it will do it safely, and the people who’ve built the plane don’t either, then just how dumb are you? Why are you even contemplating it?? 😂

no-one has ever said otherwise! 🤷‍♂️

No-one on here, perhaps.

If you’ve decided to get on a plane to fly to Spain, and you don’t know whether it will do it safely, and the people who’ve built the plane don’t either, then just how dumb are you?

So, everyone who flies on a Boeing then? 😉

what do you use it for?

my personal most common use-case is coding, phenomenal time-saver. The image-generation capabilities are very useful for creative ideas/jumping-off points.  I’ve also been using it a bit in home automation, as it can do many things without having to be specifically trained and or programmed to do them, which has a huge amount of (largely untapped at this stage!) possibilities.

perfect analogy! If you’ve decided to get on a plane to fly to Spain, and you don’t know whether it will do it safely, and the people who’ve built the plane don’t either, then just how dumb are you? Why are you even contemplating it?? 😂

Well, yes, but.  Many people are dumb.

Technical people may be talking about LLMs but your Man On The Clapham Omnibus is hearing “Intelligence.”  Students are using AI to write their homework, badly.  It’s happening.  People getting on your plane will inherently trust that it’s safe because what reason do they have to think otherwise?

Just tried to use it at work and they’ve only gawn an bloomin blocked it.

But using Microsoft’s browser I can use Copilot…..

People getting on your plane will inherently trust that it’s safe because what reason do they have to think otherwise?

Already answered. The people who make the plane are telling you it isn’t safe. If you don’t listen to them then you deserve a fiery death 😂

The people who make the plane are telling you it isn’t safe.

They’re not; the makers are saying ” we don’t know, and ultimately we’ll give it information, and let the plane decide”

You do realise there is no actual plane, don’t you? 🤣 OpenAI have always been transparent about ChatGPT’s limitations, and the need to check it’s output. 🤷‍♂️

I’m seeing quite a few people using ChatGPT and the like, to do things, passably, that they themselves don’t really know what to do.

Whether it’s due to knowledge they don’t possess (or inability to research), thought processes and logical reasoning they don’t have the capacity to do, attention spans they don’t have, reading comprehension skills they don’t have, professional writing skills they don’t have, etc.

It’s annoying. Maybe this makes me a bad person, and perhaps I unknowingly enjoyed being obviously way better at those things than quite a lot of other people. There’s stuff I do that it doesn’t currently come close to of course. But maybe there was some satisfaction in seeing other people reach their limits in obvious ways.

It’s annoying. Maybe this makes me a bad person, and perhaps I unknowingly enjoyed being obviously way better at those things than quite a lot of other people. There’s stuff I do that it doesn’t currently come close to of course. But maybe there was some satisfaction in seeing other people reach their limits in obvious ways.

Yeah, but nobody likes a smartass!

I struggle with anything involving maths or numbers in general, so an LLM that could help me understand something complex would be of immense benefit. I’m absolutely certain you could whizz through something I’d struggle with for hours, but rolling your eyes and sighing at my inadequacy just makes you look like an insufferable jackass!
There are editing tools in mobile photo apps now that can do things with a few taps, like removing unwanted objects that I would literally spend an entire afternoon doing in Photoshop: I would have considered selling a kidney to be able do what my bloody phone can now do in seconds!
I’m actually talking about Photoshop work over twenty years ago, my Mac then had memory and storage measured in Kb and Mb – my phone is measured in Gb and Tb.

Frankly, it’s bloody sorcery!

rolling your eyes and sighing at my inadequacy just makes you look like an insufferable jackass

Of course I don’t actually do that. I was trying to work out why it annoys me. Maybe part of my self worth comes from being good at some stuff, that now anyone can be passably good at by using ChatGPT.

https://x.com/LucGrefte/status/1806241815210709071