Did you all catch up with the the latest presentation by OpenAI. They rolled out some mind-blowing capabilities. For the older ChatGPT they introduced the Desktop version and improved the UI. The spotlight is on the flagship model – GPT-4o, it’s the improved version with enhanced intelligence, much faster and offers scalability in text, vision, and audio. Interaction becomes natural and efficient. You could upload images, chat about them, it has memory to remember things, real time analysis of charts, supports 50 languages and much more.

Good news for the developers, as they are bringing the 4o-API that is cheaper, 50% faster and has 5x higher rate limits than 4o-Turbo. The presentation had an interesting session, where Mark Chan and Barret Zoph demonstrated a few live demos. It was entertaining. Mark asked ChatGPT for help as he was feeling nervous and then made some mistakes in following the advices. He was jokingly breathing faster. GPT quickly expressed its feelings, “Woah! Slow down Mark!”. When prompted, GPT started telling a romantic story between robots as a bedtime story. What was unique, was it its ability to adjust tone as per the requests. The expressions were spot on.

With its vision capabilities, it could solve linear equations, or assist in real time analysis of code and charts. It could also read your emotions and provide real time translation.

My favourite part was when ChatGPT said, “Awww, That’s so sweet of you” as Barret showed an image with “I love ChatGPT” written on it.

The best part is that it would be available to free users as well. OpenAI aims to reduce the barriers to accessibility for users. Of course, there will be a limit on the number of messages that free users could send. It depends on the usage and the demand. After you cross the limit, well you are taken to your old friend – GPT 3.5