Hey friends,

This week I started writing an "indiepreneurs log". The idea is: At the end of each week, jot down what I worked on that week, what I learnt and current thinking on my projects.

Hoping this can be helpful in the short-term, plus a cool thing to look back on in a year.

🦶A step-wise increase in productivity

ChatGPT Pro became available in UK last week, and for me it was an instant-buy.

So this has been the first week where I've been able to access it on-demand without any friction.

The main things I'm doing right now are:

Building ML models on health data for a few different companies and research groups
Writing articles around data science, careers, healthcare and knowledge processing
Coding web apps using large language models
Preparing lectures and coding exercises for a module I'm teaching at Imperial College London in May

And for every single one of these, chatGPT has concretely enabled me to get more done in less time.

ChatGPT is a great "pair programmer". I'll constantly hop between writing my own code and asking chatGPT how to program something. It doesn't fundamentally change what I'm writing, but it enables me to write the code much faster.

When writing (articles or lecture material), chatGPT is great at highlighting other things I should include - particularly for sections where you're recapping things and not saying anything that new.

I know there's a lot of hype about large language models right now, and there are definitely things that chatGPT is not good at. But in my experience, for the type of work that I do, this has been huge.

Chris Lovejoy (chrislovejoy.eth) @ChrisLovejoy_

Today was first today using chatGPT for "serious" use (now that chatGPT pro is available in UK). Asked chatGPT to write functions for data science work with pandas. (In collaboration with GitHub Copilot in VSCode.) I have to say: I'm very impressed.

👀Avoiding the pursuit of ever-increasing productivity

There's a never-ending stream of new tools and technologies with the promise of "boosting our productivity".

I'm keen not to fall into the trap of continually trying out new tools - at the expense of actually doing the thing.

Some principles I try and follow to help with this are:

(1) Maintaining a good explore-exploit ratio

For me, this is spending a maximum of ~5-10% of my time exploring new tools. When it goes above that, I start to question myself.

(2) Not being in a rush

The Lindy Effect describes how the longer something has been around, the longer you can expect it to last.

Roam Research is an example: it was an innovative tool which got a lot of attention and was valuated very highly. But now, only a few years later, it seems to be dying out. I never really hopped on it, and now I'm glad I didn't.

(3) Optimising for adaptable tools

I want to use tools that can be used for a diverse range of things - and which you can adapt to your own specific needs.

For me, both chatGPT and obsidian are prime examples of this.

🕯From date night to chatGPT

Last night, I was having a date night with my wife. She mentioned some tasks that were causing problems at work, and before long we were testing out chatGPT's ability to help.

She was impressed with the result - and woke up today excited to play around with it more.

She'd heard a lot about chatGPT, but hadn't really seen a why for using it. Now she does.

I think this why can only be discovered through testing things out and experiencing them first hand.

So with that in mind: feel free to check out chatGPT here.

Have a great week :)

Chris

(PS. I'm hoping this email didn't read like a long sales pitch for chatGPT 😅)

About Me

👋 Hi, I'm Chris Lovejoy.

I'm a medical doctor 🩺 -> machine learning engineer 👨‍💻 -> start-up founder 💡

I'm on a mission to improve how we manage our health - and share my learnings and experiences here, on my personal website and on YouTube.

I also throw in my favourite things from the internet, and the occasional joke (humour is work-in-progress).

Chris Lovejoy's Mailing List

80: Explore vs exploit for productivity tools