A minor AI epiphany, a major usage tip. In the guide I wrote, I use a metaphor to describe the current generation of LLMs as having a comfort zone and a twilight zone. As you work with them, it's important to know where you stand. But how can you know when it's not always that obvious? Here's a practical way of getting an indication of it: Regenerate their answer. People do it naturally when using an image generator, but in the back and forth of a chat, it's easy to forget. So — regenerate. Got the same or a similar result? You're probably in their comfort zone. Got very different results or even a collection of slightly strange ones? You're in the AI's twilight zone — tread carefully and be more rigorous with your fact-checking. #ai #work #chatgpt #bard #gpt4 #strategy
I finally got to your guide today, Uri. Very insightful and useful. Thanks for creating it. I love your four-point approach. And your pay-off: "Real thinking starts before artificial intelligence begins and only gets going where it ends. I believe working with AI will teach us more about what makes us unique as human beings. In the long term, this could be a significant part of humanity’s quest in the coming decades. "May we navigate this journey wisely, with trustworthy and ethical platforms at our side. May we avoid bad-faith actors who only wish to 'flood the zone with shit' at the speed of light. "May whatever we create together be meaningful and true."
This is an excellent tip, and a practical value in the regeneration button that I hadn't thought of myself. 🙏
Strategy & Brand Consulting | D&AD Masterclass Trainer
1yThe guide https://docs.google.com/document/d/1Pym9gBxjcT7Ahj4XPZK77rHC_KJnEb4zRxvlSDwlNl0/edit?usp=sharing