Thinking about "game tape" for thinking, inspired by , , past Twitch streams. I notice that thinking has limited legibility—tape captures only a fraction of what's going on. Which useful categories of feedback are possible? Which blocked by illegibility?
Conversation
Some useful feedback I've gotten:
- are you being captured by the formal structure you created here, rather than by some real purpose?
- are you spending so much time reviewing those references because the main problem seems to hard, and you're avoiding it?
1
7
But I mostly haven't gotten useful feedback. Makes me wonder about the "span" of this format, at least interpersonally.
One barrier is that so much of effective thinking is instinctual: I often can't "justify" why I'm thinking about X; trying to do so too early will cause harm.
1
4
But a more surmountable barrier is context. For most useful kinds of feedback, you probably need an enormous amount of deep context about someone's intellectual projects. The right metric is not "get more video views" but "get the one right colleague to view".
1
5
It's also true that one can shape the env to make the game tape more legible. Video of me writing/thinking at a computer will be much more legible, at least in some ways, than video of me at a whiteboard, or thinking while walking. (And it matters what sw I use on the computer!)
1
1
I can't shake the feeling, though, that some types of feedback are broadly applicable and require little context. I'm not great at reading papers compared to people who do that regularly; I bet a skilled reader could watch a video of me with no context and point 10 things out.
1
4
Here the problem is more about compression. A skilled reader probably doesn't want to spend ninety minutes watching me dig through the paper. How to focus their attention on the part most amenable to feedback?
Replying to
Or maybe the better approach is for *me* to just watch *their* game tape and infer the diff. Hashtag-destroy-all-tacit-knowledge etc…
But I bet lots of details aren't accessible that way—particularly internal mental phenomena which may not be explicit even to the thinker!
3
2
Replying to
Quote Tweet
Replying to @ArtirKel
Ah yes, accelerated expertise.
Fun idea: train algo to recog. "critical" moments (will likely vary by audience stage, goals, etc.) in video protocols, then aggregate/cluster to map to enrich puzzles/principles.
Thinking of a riff on this from @imjuhokim: dash.harvard.edu/bitstream/hand
1
Replying to
+1 to this problem. It’s not obvious off the bat to either the viewer or the creator where the viewer’s attention needs to go to give good feedback – the creator doesn’t know where they’re messing up and the viewer doesn’t know what’s going to happen in the video.
1
I could imagine a bunch of things helping with this, none of which feel perfect.
1
Show replies