In this episode of the Inspiring Tech Leaders podcast, I continue the conversation about Generative AI tools and the enhancements to ChatGPT with the release of GPT4.
I share my thoughts and insights on how AI technology is rapidly evolving and also some of the precautions we should consider as Tech Leaders when using this incredible technology.
Please let me know what other technology topics you would like to see covered in future episodes of the podcast and remember to subscribe and stay tuned for more Inspiring Tech Leaders.
[00:00:10] Welcome to the Inspiring Tech Leaders Podcast with me, Dave Roberts. Early in the year, I spoke about the introduction of chat GPT and the buzz that it has created across the tech community. ChatGPT is a language model that makes it feel like you're interacting with a person rather than a machine. So if you're interested to learn more about the background of the initial launch of ChatGPT, make sure to listen back to that initial episode.
The fast pace of technology means that there is now a newer version of GPT which I want to talk more about in this. GPT4 was launched on the 14th of March and was made publicly available in a limited capacity via ChatGPT Plus. While access to the commercial API is still via wait list, ChatGPT plus now costs $20 per month, which offers faster response times and provides priority access to new features and improvements.
GPT4 is [00:01:00] now able to generate edits and iterates on creative and technical tasks, including writing screenplays, composing songs, and even a particular author style of writing. It can also now accept images as inputs and generate subtitles, categorizations, or any other relevant content or analysis.
GPT4 is also capable of handling over 25,000 words of text compared to just 3000 with GPT 3.5. This enables users to create longer content and improve document search and analysis capability. All of these new features means that tool is now considered as being multi-modal technology and is something that is unique in this new world of artificial intelligence and machine.
It's being predicted that chat GPT4 might have 570 times the amount of data compared to the training data size included in GPT3, meaning the results are now more accurate than the previous version. New applications are now taking advantage of the technology used to create ChatGPT, and enabling us to have [00:02:00] better knowledge management systems that help us to unlock the value of our collective know.
Just two days after the launch of ChatGPT4, we saw Microsoft announce the 365 co-pilot product. It's still in preview at the moment for a small number of enterprises and individuals for testing purposes, but co-pilot uses AI technology based on GPT4. This will assist with a wide range of productivity enhancing tasks and analysis that is embedded into the familiar Microsoft 365 suite of tools.
But let's remember, there are still limitations to what the technology is capable of providing. We already know that ChatGPT has only been trading on data up to September, 2021, which means it can't answer questions about current issues or events. There are also accuracy issues, which means that ChatGPT doesn't always provide the right answer, so expect some inaccuracies..
At the moment ChatGPT can also find it difficult to understand context, meaning that it can have difficulty with things like sarcasm and humor. In a world where we're [00:03:00] increasingly aware of privacy issues, we should be careful about what data we share and any generative AI tools should not be excluded from this precaution.
Consider how much data you want to share and what you might consider to be sensitive data as your input when working with generative AI tools such as ChatGPT. If you want to test ChatGPT for yourself, then that's possible. If you visit chat.openai.com and register an account. I'd be interested to hear about your comments and experiences today with ChatGPT, as well as other generative AI tools.
I hope you found this podcast helpful, and I'm sure there'll be further podcasts on this subject. I get the feeling that we're just getting started on this exciting technology. Thanks again for listening.
Please remember to subscribe to the podcast and stay tuned for more inspiring tech leaders.