In the interests of taking your words to heart, I agree that EAs (and literally everyone) are bad at steelmanning criticisms.
However, I think that saying the âand literally everyoneâ part out loud is important. Usually when people say âX is bad at Yâ they mean that X is worse than typical at Y. If I said, âDetroit-style pizza is unhealthy,â then there is a Gricean implicature that Detroit-style pizza is less healthy than other pizzas. Otherwise, I should just say âpizza is unhealthyâ.
Likewise, when you say âEAs seem particularly bad at steelmanning criticisms,â the Gricean implication is that EAs are worse at this than average. In another thread above, you seemed to imply that you arenât familiar with communities that are better at incorporating and steelmanning criticism (please correct me if Iâm mistaken here).
There is an important difference between âeveryone is bad at taking criticismâ/ââEAs and everyone else are bad at taking criticismâ/ââEAs are bad at taking criticismâ. The first two statements implies that this is a widespread problem that weâll have to work hard to address, as the default is getting it wrong. The last statement implies that we are making a surprising mistake, and it should be comparatively easy to fix (as others are doing better than us).
I donât generally like steelmanning, for reasons that this blog post does a decent job of summarizing. When folks read what I write, Iâd rather that they assume that I thought about using a weaker or stronger version of a statement, and instead went with the strength I did because I believe it to be true.
If an issue is framed as black or white, and I believe it to be grey, then I assume we have a disagreement. I try to assume that if an author decided to frame an issue in a particular way, itâs because thatâs what they believe to be true.
Apologies, I donât mean to imply that EA is unique in getting things wrong /â being bad at steelmanning. Agree that the âand everyone elseâ part is important for clarity.
I think whether steelmanning makes sense depends on your immediate goal when reading things.
If the immediate goal is to improve the accuracy of your beliefs and work out how you can have more impact, then I think steelmanning makes sense.
If the immediate goal is to offer useful feedback to the author and better understand the authorâs view, steelmanning isnât a good idea.
There is a place for both of these goals, and importantly the second goal can be a means to achieving the first goal, but generally I think it makes sense for EAs to prioritise the first goal over the second.
Thanks, I think this is an excellent response and I agree both are important goals.
Iâm curious to learn more about why you think that steelmanning is good for improving oneâs beliefs/âimpact. It seems to me that that would be true if you believe yourself to be much more likely to be correct than the author of a post. Otherwise, it seems that trying to understand their original argument is better than trying to steelman it.
I could see that perhaps you should try to do both (ie, both the authorâs literal intent and whether they are directionally correct)?
[EDIT: Iâm particularly curious because I think that my current understanding seems to imply that steelmanning like this would be hubristic, and I think that probably thatâs not what youâre going for. So almost certainly Iâm missing some piece of what youâre saying!]
In the interests of taking your words to heart, I agree that EAs (and literally everyone) are bad at steelmanning criticisms.
However, I think that saying the âand literally everyoneâ part out loud is important. Usually when people say âX is bad at Yâ they mean that X is worse than typical at Y. If I said, âDetroit-style pizza is unhealthy,â then there is a Gricean implicature that Detroit-style pizza is less healthy than other pizzas. Otherwise, I should just say âpizza is unhealthyâ.
Likewise, when you say âEAs seem particularly bad at steelmanning criticisms,â the Gricean implication is that EAs are worse at this than average. In another thread above, you seemed to imply that you arenât familiar with communities that are better at incorporating and steelmanning criticism (please correct me if Iâm mistaken here).
There is an important difference between âeveryone is bad at taking criticismâ/ââEAs and everyone else are bad at taking criticismâ/ââEAs are bad at taking criticismâ. The first two statements implies that this is a widespread problem that weâll have to work hard to address, as the default is getting it wrong. The last statement implies that we are making a surprising mistake, and it should be comparatively easy to fix (as others are doing better than us).
I donât generally like steelmanning, for reasons that this blog post does a decent job of summarizing. When folks read what I write, Iâd rather that they assume that I thought about using a weaker or stronger version of a statement, and instead went with the strength I did because I believe it to be true.
If an issue is framed as black or white, and I believe it to be grey, then I assume we have a disagreement. I try to assume that if an author decided to frame an issue in a particular way, itâs because thatâs what they believe to be true.
Apologies, I donât mean to imply that EA is unique in getting things wrong /â being bad at steelmanning. Agree that the âand everyone elseâ part is important for clarity.
I think whether steelmanning makes sense depends on your immediate goal when reading things.
If the immediate goal is to improve the accuracy of your beliefs and work out how you can have more impact, then I think steelmanning makes sense.
If the immediate goal is to offer useful feedback to the author and better understand the authorâs view, steelmanning isnât a good idea.
There is a place for both of these goals, and importantly the second goal can be a means to achieving the first goal, but generally I think it makes sense for EAs to prioritise the first goal over the second.
Thanks, I think this is an excellent response and I agree both are important goals.
Iâm curious to learn more about why you think that steelmanning is good for improving oneâs beliefs/âimpact. It seems to me that that would be true if you believe yourself to be much more likely to be correct than the author of a post. Otherwise, it seems that trying to understand their original argument is better than trying to steelman it.
I could see that perhaps you should try to do both (ie, both the authorâs literal intent and whether they are directionally correct)?
[EDIT: Iâm particularly curious because I think that my current understanding seems to imply that steelmanning like this would be hubristic, and I think that probably thatâs not what youâre going for. So almost certainly Iâm missing some piece of what youâre saying!]