#6: To-do lists as predictions

A reframing of schedules and tasks that I've found helpful lately.

I’ve written before about my perilous relationship with to-do lists and time management. The summary is: task lists and schedules stress me out. They bring me back to my stretched-thin, over-committed college days when I had no choice but to cram an unreasonable number of tasks in one day just to stay afloat.

I encountered an idea recently, from Carlos’s episode on the Do Explain podcast, that has helped me embrace schedules and task lists in a new way. In the episode, Carlos’s suggestion was to treat schedules as predictions rather than commitments.

Some personal context: I start each day with a new page in my physical notebook outlining plans for that day. I create a list of things I’d like to get done, e.g.

  • stretch & meditate

  • read The Midnight Library

  • [work things, e.g. make PR to fix UI bug]

  • exercise & shower

  • read Logic of Scientific Discovery

  • write blog post about causality

  • call brother

  • send email about tax returns

  • watch Theories of Everything interview

There are two ways to go about such a list. I could treat it as an “unordered bag”, meaning I do what I feel like, for however long I want. Or I could treat it as an ordered list with rigid start and end times. I’ve tended towards the “unordered, do-as-you-please” approach for a while, because as mentioned, schedules can take the fun out of things. But the “unordered” approach poses a different problem: you have to constantly make decisions about what to do, and it’s easy to feel like a day was “wasted” in retrospect if you’d never set clear plans for how to spend it.

The prediction vs commitment dichotomy has helped me solve some of these problems, getting me past my aversion to schedules to see the utility in them.

By default, we think of to-do lists and schedules as commitments. We say we’ll do task 1 for half an hour, then task 2 for the next half hour, and so on. We start with the set of things we feel like we should do, and then try to figure out how we can squeeze all those things into the day, and promise ourselves we’ll actually do those things. We succumb to wishful thinking about how long things take us. And when we ultimately fall short of these commitments, we feel bad and get frustrated with ourselves.

The alternative is to make a schedule based on what you genuinely expect you will do today. Do you think you'll spend half an hour procrastinating? Write that down. Will you probably spend an hour making dinner even though you could technically squeeze it into twenty minutes? Schedule a whole hour. And so on.

Then as you go through your day, you can assess how accurate your predictions are. Crucially, incorrect predictions are not a moral failing—they're just a signal to consider tinkering with your model of yourself. Perhaps you need to update your understanding of your motivations, or your environment, or your proficiency at a particular kind of task. There's also inherent unpredictability in the world and in human behavior, so a missed prediction is often just noise.1

The key distinction is that in the commitment-based approach, the goal is to get as much done as possible, which incentivizes you to delude yourself about how much you will do. In the prediction-based approach, the point is not to get more things done, but rather to be accurate about your future behavior. The goal is learning and understanding.

I've been doing this more and more when I make my daily schedules, and it's helped me be more realistic with myself and feel less stressed. I accept that I won’t be able to do more than one or two “productive” things in the evening. That even though I could publish a post every week if I was optimally efficient, a better prediction is that I’ll publish every month. If you're like me and tend to set ambitious daily goals which you subsequently miss, this might also be helpful for you.

The fun part is that you can apply this to much more than a daily to-do list. You can take a predictive mindset towards much of your life and even your long-term plans. "Ok, I know I want to go through twenty apartment viewings in the next month, but how many do I actually expect to get through?" "I know I should have this difficult conversation soon, but when do I genuinely believe I'll initiate it?" It's about letting go of the should's momentarily and treating yourself with compassion and forgiveness. It's about trusting that you will do the things that must be done, but being honest with yourself that it might not be on the most efficient schedule.2

They say plans are useless, but planning is essential. The process of making a plan—a tentative prediction of what your future behavior will be—helps you get a grasp on your life and direction, and it forces you to test how well you really understand yourself. A plan also helps reduce decision fatigue around what you’ll do from one hour to the next.

And here’s the thing: the plan doesn't actually need to be followed to the letter to be useful. In fact, deviations from the plan could be a good thing. They might be an indication that you have something to learn about yourself, or just that your own goals and desires have changed over time (which is good!).

So try this next time you're planning out your day or week. Take the role of intrigued scientist researching yourself, rather than unforgiving self-dictator. Make predictions rather than commitments. And ponder the gap between expectation and reality with curiosity: it’s an opportunity to learn something new.

Other recent finds

A few things I’ve discovered recently that have been enriching and fun:

  • The Rationally Speaking Podcast by Julia Galef. I’ve enjoyed her recent episodes with Jonathan Haidt and Matthew Yglesias.

  • The ideas of Bernardo Kastrup. He makes some very interesting, very speculative metaphysical claims about consciousness and the universe. From his interview with Curt Jaimungal, he seems like an incredibly articulate and astute thinker. He actually disagrees with a number of the philosophical stances I hold (e.g. he’s a reductionist, and he believes there’s an ‘ultimate’ level of explanation/reality) which has been a refreshing challenge to my views!

  • This video by Visa. I’ve watched it twice and it captures so much of my learnings about self-trust, ambition, goals, and productivity. It dovetails well with the “to-do list as predictions” idea.

Other writings by me

I feel honored that you’ve chosen to subscribe to this newsletter and would be remiss to waste your time. That means I post with much more abandon on my personal blog, and I also get deeper into the weeds there with some of the academic (and personal) topics I’m interested in. I might use this newsletter as a little “wrapper” for all my other writing. So in case you’re interested, here are a few recent posts there:

See you next time,



When it comes to evaluating predictions, I personally don’t do any explicit, detailed assessment of their accuracy. I just look at how my predictions differed from reality, ponder them for a few seconds, and allow my subconscious brain to pick out patterns over time if it notices them. Over-analyzing past predictions is a mistake in itself.


Here’s one more example of a situation I love applying this framework to: moments of paralyzing dread. Let’s say you have some task you believe you should do, but for whatever reason, you feel a dread that is likely to make you procrastinate. Previously my internal monologue in this scenario would look like: ugh don’t wanna do this, but okay c’mon let’s do it, gotta do it, I keep avoiding it but gotta do it. It’s all struggle and resistance. My updated internal monologue looks closer to: ok, I don’t want to do this. what’s my best prediction about when (and whether) this will actually get done? what if I predicted that I’ll spend 5 minutes on it and then stop and do something else? And so on.