Slate Star Codex on Willpower

Slate Star Codex has a characteristically intelligent review of Baumeister’s book on Willpower today.

If you grep “beeminder” in the comments you’ll see it’s so far all anecdotes about how beeminding just adds insult to injury since you suck and are going to fail anyway.

I’m going to excerpt some of my counterarguments as replies here…


Hey TeMPOraL, highly valuable anecdote and I’m so sorry to hear that. Here are our thoughts on how to beat Beeminder burnout, as we call it: . Very interested to hear if that would work or if precommitments just don’t jibe with your personality.

I should also mention that we’re extremely averse to people paying money to Beeminder that they don’t feel was worth it (gratifyingly it’s almost universally the opposite) and if you just reply to the legit check to explain then we’ll just cancel the charge. (There’s also a “weaselproof me” box you can check if it sounds like that would defeat the whole point. :))

Could a mental reframing fix the negative impact Beeminder had (anxiety, stress)? Namely, if you value beeminding things you don’t need to beemind, for the graphs and data, and you averaged something like $10/month in derailments, maybe that’s actually a reasonable fee and can be emotionally associated with paying for a service rather than punishment for failing to follow through on commitments.

I should also repeat the advice from the post I linked TeMPOraL to, about beeminding very conservatively so there’s little risk of derailing unless egregious akrasia is at work. Of course for many people stress is the whole point. If you’re severely akratic then then it sucks to be beeminded but not as much as it sucks to, say, waste money on a gym membership you never use, or make zero progress learning new skills that have a critical long-term career impact because you’re in a perpetual state of “one more day doesn’t matter”. (Basically, if you’re deadline-driven then Beeminder’s a way to have a tiny but hard deadline for an epsilon of progress every day.)

In your case though what I’m most curious about is if it was Beeminder that led you to the insight that negative reinforcement is ineffective for you (and your interesting point that it could have to do with you be self-punishing). Or was it predictable but you wanted to be sure?

(Really appreciate this feedback, everyone!)

[two more people – Doug S. and ddreytes – chimed in to say they figure Beeminder would backfire for them]

Ooh, yes, and many people find HabitRPG and Beeminder nicely complementary. In fact, we have an official HabitRPG/Beeminder integration: &

  1. I accept that criticism. I think Beeminder might indeed be neglecting people who have the pain brain reaction.

  2. Exactly! Beeminder is absolutely in the systems camp (keep all your datapoints on this “yellow brick road” indefinitely) and it’s unfortunate that we use the term “goal” so much. I often say “graph” instead. I agree with Scott Adams’s point that goals as commonly conceived are for losers. But his criticism doesn’t apply to S.M.A.R.T.(E.R.) goals, as Katja Grace explains:

Your systems approach mindset actually sounds beautifully aligned with Beeminder.

  1. Beemind all the things! (Or, more realistically, as I’ve said in other comments here, beemind things very conservatively, focusing on fixing egregious instances of akrasia.)

Finally, the article you link to about, basically, the value alignment problem for goal setting is really good but I believe that goals one sets for oneself are reasonably immune to that criticism. Sometimes people define elaborate fine print for themselves but mostly it seems to be unnecessary. You may even purposefully leave huge loopholes (like “I will touch the door of my gym 3 times per week”) because you know yourself and what loopholes you’ll tend to abuse (like you may know that once you arrive at the gym you’ll have no trouble going in and actually working out).

Here’s my top-level reply to the Slate Star Codex post:

Love this (as always) and just want to cast my vote to review Carol Dweck’s book next:

So far I’m on her side in the Baumeister vs Dweck debate, so I’m eager to hear what’s behind that “if Dweck were an oncologist” jab (or maybe it was entirely good-natured).

To argue against your million-dollar analogy: With physical endurance you approach a physical limit asymptotically. The feeling that you can always eke out more with the right inducement is an illusion. Eventually one more straw will in fact break a camel’s back. As for willpower… With the right inducement (say, continued employment) you can exert superhuman willpower, like waking up early and going to work every day for years or decades. Which is to say that with the right incentives, willpower doesn’t even need to be invoked. So maybe I’m saying that it doesn’t matter if willpower is unlimited, you can route around it and find creative ways to induce yourself to do what you really want to do.

Dammit, I degenerated immediately to pitching Beeminder again!

But to finish the thought, as I’ve argued in the comments above, the best way to use Beeminder, at least initially, is not to probe the hard limits of willpower but to fix egregious instances of akrasia — to do a bit more than the bupkes you’d do if left to your own devices. You can then gradually dial up the steepness of your graph, but certainly stop doing that before feeling overwhelmed by stress and anxiety.

In other words, make a measurable improvement well below the point that the limits of willpower are even a question (if you don’t think of it as routing around willpower altogether). Some people — like the productivity-ueber-alles types who try polyphasic sleep and whatnot — thrive on adding stress and Beeminder can accommodate that. But using it in moderation can reduce stress, like by getting you to spread your studying out over a semester instead of cramming for exams, or by making you pay attention to your fitbit just enough to actually get in 10k steps a day, or getting yourself to bed on time instead of repeatedly getting in “one more point” in a comment thread till 2am.


Another thing that came up in the Slate Star Codex discussion:

More fundamental to Beeminder than magnifying negative consequences is, exactly as Gudamor says, bringing the negative consequences SOONER. You have to have your datapoints on Beeminder’s yellow brick road every day, making slow but steady measurable progress.

That’s in fact Beeminder’s key difference from all the other commitment device apps we know of, which we list here, btw:

1 Like

Seems to me like he’s shooting the messenger. Beeminder is giving
him a smack upside the head saying either he doesn’t really want that
goal (even if it seems nice) or that the current methods he’s using
to get to that goal just aren’t effective.


Also, this may be a little harsh (especially from one little comment on a blog), but I think the people who are crying about anxiety are just seeing the anxiety manifest itself extra hard through beeminder. When they don’t measure how lazy they’re being, it’s easier to be nice to yourself. A good dose of therapy might help them cope in general.

Don’t worry though, Beeminder isn’t keeping me up at night (probably true for most other users). My wife on the other hand thinks Beeminder and its system is completely stupid. Then again, she never seems to procrastinate EVER. She also spends her work day being empathetic towards people with severe mental health issues (psychologist) and dealing with our 2 year old daughter (actually very similar to an adult with severe mental health issues!) which ends up leaving little room for understanding for me and my desires to become better.


love this comment!! beeminder doesn’t magically make me do things. but it certainly doesn’t make me not do things that i would have done without beeminder. if i’m derailing on my beeminder goals, i obviously wasn’t going to be doing the thing in the first place.