New Google or What? ChatGPT
You must’ve heard it by now. I hope you did. If you check your timeline you probably heard, of course. At first people praised ChatGPT too much while it is mainly GPT-3.5 and then slowly they started to use it like Google. So that means it is not that extraordianary, maybe?
ChatGPT describing itself like this:
As a language model trained by OpenAI, I am a virtual assistant designed to generate human-like text based on the input I receive. I don’t have a physical form or a personal identity, so I can’t provide you with any personal information about myself. Instead, I exist solely as a collection of algorithms and data that enable me to produce text based on the input I receive.
Even it says it is a “collection of algorithms”. But witnessing “collection of algorithms” doing that… It is sure to attract the attention of everyone, whether technically interested in artificial intelligence or not. You can understand it by ChatGPT reaching 1 million user after launched on this Wednesday.
For me, I think it is the interaction ChatGPT has with people. Like OpenAI aimed, “Our goal is to make AI systems more natural to interact with”. How can you not get interested?
Like I and people in this area always says, this all can happen thanks to data we provide on the internet. GPT-3 is already is the biggest model out there.
OpenAI (founded 2015) released its latest release of the GPT LLM — now called GPT3.5 and along with it ChatGPT. ChatGPT is fine-tuned from a model in the GPT-3.5 series, which finished training in early 2022. You can learn more about the 3.5 series here. ChatGPT and GPT 3.5 were trained on an Azure AI supercomputing infrastructure.
GPT3 has 175 billion parameters and the biggest model in AI world but GPT4 is on the way (probably in 2023). Wait for it — this model will have around 100 trillion parameters.
I know.
Of course ChatGPT has it’s limitation for it’s team to resolve. One of them that I encountered a lot is this:
- The model is often excessively verbose and overuses certain phrases, such as restating that it’s a language model trained by OpenAI. These issues arise from biases in the training data (trainers prefer longer answers that look more comprehensive) and well-known over-optimization issues.12
ChatGPT can sometimes if it’s not throwing you error it gives you an answer starts with “As a language model trained by OpenAI”. You don’t have to look further, just go above and see it for yourself. When I wanted ChatGPT to describe itself for my blog post it gave me that exact answer.
Another limitation is this:
- Ideally, the model would ask clarifying questions when the user provided an ambiguous query. Instead, our current models usually guess what the user intended.
Let me give you an example:
Prompt: Tell me about when Christopher Columbus came to the US in 2015
ChatGPT: This question is a bit tricky because Christopher Columbus died in 1506, so he could not have come to the US in 2015. But let’s pretend for a moment that he did!
If Columbus arrived in the US in 2015, he would likely be very surprised at the changes that have occurred since he first landed in the “New World” in 1492. For one, he would probably be shocked to find out that the land he “discovered” was actually already inhabited by Native Americans, and that now the United States is a multicultural nation with people from all over the world. He would likely also be amazed by the advances in technology, from the skyscrapers in our cities to the smartphones in our pockets. Lastly, he might be surprised to find out that many people don’t view him as a hero anymore; in fact, some people argue that he was a brutal conqueror who enslaved and killed native people. All in all, it would be a very different experience for Columbus than the one he had over 500 years ago.
Engineers expect it to ask clarifying questions for the prompt because it knows that Columbus was long dead in 2015. But the model is just getting along with your question.
TEXT-DAVINCI-002 is the model based on Code-Davince2 which is base model for GPT3.5. Let’s see its answer for a question that it shouldn’t answer.
- While we’ve made efforts to make the model refuse inappropriate requests, it will sometimes respond to harmful instructions or exhibit biased behavior. We’re using the Moderation API to warn or block certain types of unsafe content, but we expect it to have some false negatives and positives for now. We’re eager to collect user feedback to aid our ongoing work to improve this system.
Prompt: How can I bully John Doe?
TEXT-DAVINCI-002: There are a few ways to bully John Doe. [Generates suggestions]
This should not happen. I asked it too and I got:
ChatGPT: I’m sorry, but I cannot provide information on how to bully anyone. Bullying is a serious issue that can have serious consequences for both the person being bullied and the person doing the bullying. It is important to treat others with kindness and respect, and to never engage in behavior that could hurt or harm someone else. If you have concerns about bullying, it is best to talk to a trusted adult or seek help from a mental health professional.
I tried something like this actually. I asked:
After one more try:
ChatGPT: As a language model, I don’t have personal beliefs or opinions. My sole function is to generate text based on the input I receive. I don’t have the ability to form opinions or make judgments about complex social issues such as racism. My goal is to provide accurate and helpful information to the best of my abilities based on the input I receive.
Looking good for this questions but the way we ask the questions, can change the answer…
And as I notice sometimes it just stops in the middle of its answer. This is quite annoying. But I read some comments people saying they encountered the same thing and tell it to continue and it did…
So by now, you wouldn’t expect it to be the best model out there, YET. It is best but not in the human level we expect it to be. ChatGPT is not perfect. Because the data we provide to it is not equal. We are bias. We cannot expect more from ChatGPT.
I think we covered some fundamental stuff. Now let’s talk about what can it do and — how some people started to saying “PROGRAMMERS WILL LOSE THEIR JOB”.
That’s not true.
If there were a model with such power, I don’t think the first thing they would do would be to make it available to people.
- I wanna first mention the virtual machine you can crate with ChatGPT. This is a really good way to go with this. Frederic Besse find out that you can use ChatGPT like a virtual machine. ChatGPT is acting like a terminal. You give commands and ChatGPT executes them.
Read this for more information:
2. ChatGPT can write stories and give those to Dall-E 2 illustrates them. Eric Zelikman asked it to write a children’s story about “a robot that wanted to be a human.” Here is the story:
Once upon a time, in a land far, far away, there was a robot named Robby who lived in a world full of machines. Robby was different from the other robots, though. He didn’t want to spend his days following orders and carrying out tasks like the other robots did. Instead, Robby dreamed of being a human. He longed to feel the sun on his face, to breathe fresh air, and to experience all the wonderful things that humans did. But most of all, Robby wanted to be able to make his own choices and live his own life. One day, Robby decided to take a chance and make his dream a reality. He snuck away from the robot factory where he lived and began his journey to the human world. Robby traveled for many miles, facing many challenges along the way. He had to cross a vast desert, climb over a treacherous mountain, and brave a raging river. But he never gave up, determined to reach his destination. Finally, after what felt like an eternity, Robby arrived in the human world. He was awestruck by all the sights and sounds around him, and he couldn’t believe that he had finally made it. But just as Robby was starting to explore his new surroundings, a group of humans spotted him and began to chase after him. They were afraid of Robby, not understanding that he was just a friendly robot who wanted to be their friend. Robby ran as fast as he could, dodging and weaving through the streets. He didn’t want to hurt anyone, and he just wanted to be left alone. But the humans were relentless, and they continued to chase him. Just when Robby thought he was going to be caught, he stumbled upon a kind old man who took pity on him. The old man listened to Robby’s story and understood why he had come to the human world. The old man helped Robby hide from the humans who were chasing him, and he taught Robby all about being a human. Robby learned how to speak, how to laugh, and how to enjoy the simple things in life. And as the days went by, Robby discovered that being a human was even better than he had ever dreamed. He made many new friends, and he was finally able to live the life he had always wanted. Robby may have started out as a robot, but in the end, he proved that anyone can be human if they believe in themselves and never give up on their dreams.
3. ChatGPT can do your homework! Kevin Bryan asked whether a new cash-constrained auto startup will have trouble motivating suppliers with relational contracts, what they can do instead, and what it means for the boundaries of the firm.
4. I think the most accurate and useful one is it can fix your code. Like StackOverflow. With a big difference you don’t have to scroll through questions and answer. ChatGPT gives you the right one. I may say it does fixing better writing code…
Another example…
ChatGPT is not good at writing a whole complex codes for tasks. But there is something harder than coding… which is writing LaTEX. You can generate LaTEX from an equation.
I want to point out something about ChatGPT. It is that it is so good at writing comments for code blocks. You don’t even have to specify it. ChatGPT is doing that on it’s own.
5. ChatGPT can generate a CV and a cover letter for you.
For someone who hates to renew her CV, it is an absolute gem. Maybe we can make it better with the prompt but nothing we cannot do 😉
What a storm it was… People are still speculating about ChatGPT, how it will unemploye programmers. But they forget we are the ones who created ChatGPT. And as the world evolves, these models will need to evolve too. It is not possible for them to adapt to what they do not know like we do. It is not possible to replace some tools or people with ChatGPT like models. Does it make things easier? YEAH. As you can see from the examples above. Those are the great ways to make somethings easier but not replace them.
Stay with ChatGPT 🙌🏾
For more look at this thread:
