
On Tuesday, 14th March 2023, OpenAI announced the launch of ChatGPT-4. GPT-4 provides the latest upgrades in the chatbot released back in November. According to the company, the changes provided in the new system are subtle, but they are there. So what are the latest features of ChatGPT? Let’s dive deep into the news released yesterday.
Some of the most notable features you can expect from GPT-4 is that the new version can read images and provide better creative thinking. The release of ChatGPT back in November was one of the biggest technological advancements in the latest times. Tech geeks all around the world loved the new technology and started making use of it to solve everyday tasks. This prompted big companies such as Microsoft and Google to go into a ‘code red’ situation to start working on their AI chatbots and improving their features.
Since the release of ChatGPT, many more chatbots have stepped up. Bing, Bard, and Claude are to name a few big tech giants. The AI revolution is the future of technology as we know it. It would be costly for any tech company not to integrate it into their system.
With the latest advancements in Chatbots and AI technology, OpenAI decided to improve the features of its existing model and bring it to the world. The news about the GPT-4 was released last week, and naturally, the world was excited to see what more we can do with this chatbot. So, here’s a breakdown of the latest features of ChatGPT and how you can use them.
Image Processing Alongside Text:
We have all used ChatGPT for text processing since the start. You enter a text and ask the bot to rephrase it, edit it, tell you its meaning, and so on. With GPT-4, you go one step ahead and experience image processing as well. One of the biggest differences between GPT-3.5 and GPT-4 is that the new version lets you process images and gives suggestions. The last version could only do that with text.
Here’s an example of how you can use image processing with GPT-4. Take a picture of the contents of your fridge and show it to the chatbot and ask what can be made using just the ingredients inside. GPT-4 will be able to give you a bunch of different options as to what you can prepare with the contents.
Similarly, as released by OpenAI itself, you can create an entire code for a website with just a simple sketch. Furthermore, OpenAI announced that it is working with a startup, Be My Eyes that works in object recognition to help people with vision problems. Be My Eyes would be working alongside OpenAI to improve the company’s app using ChatGPT-4.
Learn more about programming with ChatGPT here.
Creative Thinking and Reasoning:
According to the company, you can expect GPT-4 to be more creative and provide more thoughtful reasoning than before. With the help of GPT-4, you can summarize large texts or blogs into a line of words with a single alphabet.
In a demo by OpenAI, co-founder Greg Brockman asked the bot to summarize a portion of a blog by only using the letter ‘g’. The AI chatbot was able to do so without any problem. This is a new feature added in the chatbot and can help you summarize long texts in a much more creative and fun manner.
Text Processing Limit:
One of the biggest limitations of AI languages is the amount of text they can process. There is only a limited amount of text that can be stored in such models. This is an area that OpenAI has worked on and improved drastically with GPT-4. With the new version, the system can process entire novels, write them in one go, and so on. You can expect GPT-4 to summarize entire scientific papers for you as well. This means that the new version will be able to give answers to more complicated questions and provide more logical solutions.
GPT-4 measures the amount of text that goes in and comes out in the form of ‘Tokens’. Tokenization is a hard concept to understand but all in all, one token is equal to about four characters and so 75 words can be equal to 100 words.
With the older version, that is, GPT-3.5, the maximum number of tokens available was about 4000 for any given query. This equals about 3000 words. With GPT-4, the available tokens are as much as 32,000 for a single query. Meaning you can add text with around 25,000 words and the system will be able to process it and give the desired output you need.
OpenAI has come up with some great features for the new version of ChatGPT. But it is still an AI chatbot. Expecting it to be perfect is just setting yourself up for failure. There is a lot of room to grow and although you can expect some amazing features from this new version, there are still chances that the new version might feel a little underwhelming compared to the hype created around tech geeks. These are the latest features of ChatGPT, which ones are you most excited to use?
Comment (1)
Comments are closed.
Codesy Consulting
March 24, 2023[…] Here’s more on the GPT-4. […]