During his trial in 399 BC, Socrates famously told the jury, "The unexamined life is not worth living." This not being a philosophy essay, I instead want to apply this line of thinking to something that would have been foreign to the Athenian gadfly: projects. Just as the unexamined life is not worth living, the unexamined project is not worth doing.
When you finish a project, there is tremendous value in taking out a pen and reflecting on it. This means a short write up, 1 to 2 pages, that explains the project, nuances about how it went, what you learned, and what you would do differently next time. The exact structure is less important than just sitting down and writing it.
What if you’re a bad writer? This wasn’t a good excuse before ChatGPT, and it’s even less of one now. Your paper will not be on the front page of the New York Times. In fact, there is a good chance you will be the only human to ever read it. This is fine. I personally would suggest writing it yourself, but an LLM-written and human-edited essay works too.
This simple exercise is beneficial for two reasons: learning and AI training. Let’s examine each in detail.
Reason #1 - Learning
I don’t have a great memory. It’s extremely frustrating when I forget 90% of a book just a month after reading it. There are a few lucky people who can read something once and remember it forever, but most of us cannot.
I found the best way to fix this was to write about whatever I had read. I started my blog almost exclusively for this purpose. And guess what? It worked. My recall and overall understanding of books that I review is orders of magnitude greater than for reading alone.
What is it about writing that dramatically improves retention? It has to do with teaching. It's common knowledge that the best way to learn something is to teach it. The act of breaking-down, structuring, and articulating a concept to an audience with no context deepens your understanding on a neurological level. This is the power of writing: it compels you to teach.
Your “career experience” is largely made up of what you learned from dozens and dozens of projects. The better you can remember what you did, the more experience you will inherently have. If you take two people with identical project experience, but one writes summaries and the other does not, you will end up with two people with vastly different levels of experience, despite everything else being the same.
Admittedly, I did not do this for almost the first decade of my career. What I’m left with is a very rough understanding of dozens of projects without the ability to recall what went wrong, what went right, and what I learned from them. Don’t be like me.
Summarizing projects goes beyond your personal ability to remember, it also improves organizational learning. The lifeblood of every business is the project. One of the most important things a business can do is to improve their ability to execute on projects and initiatives, finishing each one quicker, cheaper, and more successfully than the last. And as we have said, the way to improve at anything is by learning and reflecting on what you have done in the past.
Reason #2 - AI Training
Just as writing a summary helps humans learn, it will also help machines to learn. Corporate LLMs will soon have large enough context windows to be able to intake the entirety of a company's internal data. This includes the structured data that exists in databases, and unstructured data like emails, DMs, and documents. The “better” this data is the more ways AI can help us with projects, or even eventually do projects on its own.
Emails and DMs will give the AI good context into how a project went, but it will most certainly fail to give the entire picture. This should not come as a surprise. Everyone reading this has probably seen how important context is to an effective LLM prompt. Forcing the AI to infer insight based on a rough collection of conversations is the quickest way to have misguided inference.
Improving corporate LLMs through reflective project summaries is what will give you “better data” and it will separate those that just use LLMs and those that harness their full potential.
Concluding Thoughts
The astute reader may have picked up on the irony of using Socratic wisdom to promote writing. Socrates famously never wrote anything down. Despite this, his framework on life revolved around learning and reflection, something we all could do more of.
Passively experiencing projects without reflecting on them is like collecting a ton of valuable data but never analyzing it. The best time to start was 5 years ago, the next best time is today. At OVG, we recently began pushing everyone at the company to write more. Do yourself and your organization a favor and start writing.