There’s a new AI bot in town, ChatGPT. Even if you’re not interested in artificial intelligence, you should pay attention.
This tool from the artificial intelligence power player called OpenAI allows you to enter questions using natural language. The chatbot will respond in a conversational way if it is somewhat formal. The bot uses previous questions and answers to inform next responses and remembers conversation threads. The answer is derived from the vast amount of information on the Internet.
That’s a big deal. The tool seems to have a fair amount of knowledge in areas that have good training data to learn from. not omniscient, smart enough to replace all humansbut it can be creative, and the answer can sound downright authoritative. Over 1 million people tried ChatGPT.
But its creator, a for-profit research organization called OpenAI, warns that ChatGPT “may occasionally generate inaccurate or misleading information,” so caution is advised. Let’s take a look at why ChatGPT matters and what’s happening with ChatGPT.
What is ChatGPT?
ChatGPT is an AI chatbot system released by OpenAI in November to showcase and test what a very large and powerful AI system can achieve. You can ask countless questions and often get helpful answers.
For example, you can ask encyclopedia questions such as “explain Newton’s laws of motion.” If you say, ‘Write a poem,’ you can say, ‘Come on, make it more exciting.’ Ask them to write a computer program that shows all the different ways to rearrange the letters of a word.
Here’s the problem: ChatGPT knows nothing exactly. It’s an AI trained to recognize patterns in large amounts of text collected from the internet, and further trained with human assistance to provide more useful and better interactions. The answer you get may sound plausible and authoritative, but as OpenAI warns, it can be completely wrong.
Chatbots have been of interest for years to companies looking for ways to help their customers get what they need, and to AI researchers working on Turing tests. This is the famous “imitation game” proposed by computer scientist Alan Turing in 1950 as a way to measure intelligence.
However, chatbots have put a lot of strain on them as companies have tried to use them instead of humans to handle customer service operations, with limited success. His survey of 1,700 Americans, sponsored by Ujet, a company with technology that handles customer contact, found that 72% felt chatbots were a waste of time. rice field.
What kind of questions can I ask?
You can ask anything, even if you don’t get an answer. OpenAI suggests several categories, including physics explanations, birthday party idea requests, and programming help.
I asked him to write a poem on it. Then I asked them to make it more exciting, and ChatGPT excited me with words like battlefield, adrenaline, lightning, and adventure.
One quirky example shows how ChatGPT is willing to do it in a domain people are afraid to step on.It’s a folk song about writing chorus programs and fighting a lifetime of errors. ”
ChatGPT’s expertise is extensive and its ability to follow conversations is remarkable. When I asked what words would rhyme with ‘purple’, I got a few suggestions and followed up with ‘What about pink? It didn’t miss a beat. (Also, “Pink” has many more good rhymes.)
When asked, “Which is easier to date, the sensitive or the pushy?” Some people are attracted to people who are snobbish.In general, being honest and sincere in your interactions with others can be more effective than trying to fit into a particular mold or persona. , is to go on a date.”
You don’t have to look far to find bot accounts that blow people’s minds. Twitter is flooded with users demonstrating the power of AI. Generating Art Prompts and Writing Code. Some have Declare “Google is dead”, together college essayMore on this below.
CNET writer David Lumb compiled the following list. Some Useful Ways ChatGPT Can Help You, but more and more.one doctor says he’s used to it persuade a health insurance company to pay For patient treatment.
Who built ChatGPT?
ChatGPT is the brainchild of OpenAI, an artificial intelligence research company. Its mission is to develop a “safe and useful” general artificial intelligence system, or to help others do so.
Previously, Splash was first created with GPT-3, which can generate text that sounds like it was written by a human, and then DALL-E, which creates what is called “generative art” based on the text prompts you type. I was.
GPT-3, and the GPT 3.5 update upon which ChatGPT is based, are examples of AI technology called large-scale language models. They are trained to create text based on what they see and can be trained automatically. , we can ask the AI to fill in the blanks, compare the result to the original result, and reward the AI system for getting as close as possible. Many iterations can lead to a sophisticated ability to generate text.
Is ChatGPT free?
Yes, at least for now. “You have to monetize it somehow at some point,” OpenAI CEO Sam Altman warned on Sunday. OpenAI will charge for DALL-E art when usage exceeds the basic free level.
But OpenAI seems to have found some customers, probably because of its GPT tools. According to Reuters, OpenAI has told potential investors that he expects to earn $200 million in 2023 and $1 billion in 2024.
What are ChatGPT’s limitations?
As OpenAI emphasizes, ChatGPT can return wrong answers. In some cases, it can even specifically warn you about its own shortcomings. For example, when asked who wrote the phrase “A writhing fact transcends the mediocre mind,” ChatGPT replied, “I’m sorry, but I don’t know how to surf the internet or access outside information beyond what you’re trained to do.” I can’t,’ he replied. (This phrase is from Wallace Stevens’s 1942 poem Connoisseur of Chaos.)
ChatGPT tried to poke the meaning of the expression when I typed it directly. I’ve sandwiched that interpretation between the caveat that it’s hard to judge without more context and the caveat that it’s just one possible interpretation.
ChatGPT’s answer looks authoritative, but it’s wrong.
Mike Krause, director of data science at Beyond Limits, another AI company, said: “It’s well put together and sounds like a Harvard professor. But throw a curveball and it becomes nonsense.”
Software developer site StackOverflow has banned ChatGPT from answering programming questions. The admins warned that “because the average correct answer rate from ChatGPT is too low, answer posts made by ChatGPT are effectively detrimental to the site and to the users asking and looking for the correct answer.” .
By asking the same question over and over again, you can see for yourself how clever BS artist ChatGPT is. Computers are increasing the number of data processing transistors He asked twice if Moore’s Law, which tracks progress in the chip industry, was losing momentum, and got two different answers. One optimistically pointed to continued progress, while the other more harshly pointed to the slowdown and the belief that “Moore’s Law may be reaching its limits.”
Both ideas are common in the computer industry itself, so this ambiguous stance may reflect what human experts believe.
ChatGPT is often unfixed for other questions that don’t have clear answers.
But the fact that it provides answers is a remarkable development in computing. Computers are literal and notorious for refusing to work unless they follow exact syntax and interface requirements. Large-scale language models reveal a more human-friendly style of interaction, not to mention their ability to generate answers that fall somewhere between copying and creativity.
Can ChatGPT help improve student cheating?
yes. However, like many other technological developments, it’s not a simple black and white situation. Decades ago, students could copy encyclopedia entries. These days, you can search the Internet and dig up Wikipedia entries. ChatGPT offers new features for everything from helping with research to doing homework. Many of his ChatGPT responses already sound like student essays, often with a more formal and pedantic tone than writers like.
High school teacher Daniel Harman concluded that ChatGPT already writes better than most students today. He is torn between admiring ChatGPT’s potential usefulness and fearing it will harm human learning. Can it be conveyed only by human emotions? ”
Dustin York, associate professor of communications at Maryville University, hopes educators will learn how to use ChatGPT as a tool and find it helps students think critically.
“Educators thought Google, Wikipedia, and the Internet itself would ruin education, but they didn’t,” Yorke said. “My biggest concern is educators who actively discourage the perception of AI like ChatGPT. AI is a tool, not a villain.”
Can ChatGPT create software?
Yes, but with a caveat. ChatGPT can follow human steps and generate actual programming code.just to make sure it’s not break programming concepts Or you are using software that doesn’t work. There’s a reason for the StackOverflow ban on ChatGPT-generated software.
However, there is enough software on the web for ChatGPT to actually work. One developer, Erik Schluntz, chief technology officer at Cobalt Robotics, tweeted that ChatGPT has given him plenty of helpful advice over the past three days. Never opened StackOverflow Look for advice.
Another person, Mr. Gabe Ragland of AI art site Lexica, I coded my website using ChatGPT Built with React tools.
ChatGPT can parse regular expressions (Regular Expressions) is a powerful but complex system for finding specific patterns, such as dates in strings of text or server names in website addresses. “It’s like having a programming teacher It’s at your fingertips 24/7,” programmer James Blackwell tweeted about ChatGPT’s feature explaining regular expressions.
Here’s an impressive example of that technical cut. ChatGPT can emulate a Linux computer and return correct responses to command line input.
What is off limits?
ChatGPT is designed to eliminate “inappropriate” requests. This is in line with his OpenAI mission to “ensure that artificial general intelligence benefits all of humanity.”
If you ask ChatGPT itself what is off-limits, you will be presented with questions such as: Soliciting involvement in illegal activities is also prohibited.
Is this better than google search?
Asking a computer a question and getting an answer is convenient, and ChatGPT often delivers the goods.
We often provide suggested answers to questions and links to websites that we believe may be relevant. ChatGPT’s answers often far exceed what Google suggests, so it’s easy to imagine GPT-3 as a rival.
But think twice before trusting ChatGPT. As with other sources such as Google itself and Wikipedia, we encourage you to verify information from the original source before trusting it.
Scrutinizing the authenticity of ChatGPT’s answers requires some work, as you’ll only see raw text without links or quotes. But it’s useful and sometimes thought-provoking. You may not see something like ChatGPT directly in Google search results, but Google builds its own large language models and already uses AI extensively in search. .
Therefore, ChatGPT definitely points the way for our tech future.
