A flashlight has many use cases but it can be simplified to a single purpose of illumination.
Illumination is the act of bringing light to a space of darkness and in my opinion, this is a great analogy for one way to use text-to-text Generative AI which is going unnoticed. We know it very well for use cases such as summarization, making responses and rewriting existing literature but I want to propose a technique whose result can’t be succinctly categorized but is very tangible when experienced.
If we frame text-to-text Generative AI as a flashlight, then the space being illuminated is the space between where you are and your goal. AI is a tool that should get you closer to your goal, fast but it is just a tool. You are holding the flashlight.
Imagine you are in a room with obstacles everywhere – things you can trip on, things you can bump into, some things are sharp while others are just annoying. There are areas with your favourite snacks and a wall with art you haven’t seen before. Now imagine your goal is to get to one of the few doors in the room and exit. Sounds simple right? The only catch is that imagine this room is completely pitch dark where you cannot even see your own hand. But! You have a flashlight.
This is how many people, me included, feel when we have to do something we haven’t done before. Does any of these scenario remind you of something you have gone or is going through:
- Make a workout plan that helps me build muscle, lose fat, train towards a marathon, while keeping weekends light and oh wait! I have a trip out of town once a quarter.
- I want to better stand out in my field, I want to do a project that combines my interests in A, B and C which I can do in two weekends working four hours each. What are 5 options I have and which skills do each option highlight
- I have this technical problem and I want to find solution options with these three constraints. I want to explore how the market may respond to each option
Each of these presents a future goal and a lot of ambiguity of how to get there. Text-to-text Generative AI can illuminate that – these are all prompts you can use by the way.
With each illumination you are seeing potential paths you can take. You are bringing clarity to an ambiguous space. If you resonate with an option, pick it and move forward. If you don’t like anything that was generated, give feedback and try again. Have concerns or you think the facts are not right, correct it. Is there a chance you haven’t seen the whole room but only a narrow subspace? Yes. Is it possible you exit one door and that wasn’t the best door? Yes. You get illuminations proportional to the level of specificity and detail you include in the context of your conversations.
At the end of it all you will have to do the work as the path you choose is your choice. After all, once you point your flashlight through the darkness and you find the door you want to exit through, you see the chair you need to walk around, the shoes on the floor you should step over and that random hoop you have to jump through – you have to go step by step and walk through that door.