reidac.bsky.social
Scientist, cyclist, urbanist, Linux and HPC enthusiast.
Washington DC.
197 posts
211 followers
460 following
Prolific Poster
Conversation Starter
comment in response to
post
The thing I want most from social media is smart people with a different frame of reference – often due to race, gender, disability, sexual orientation, country of origin, class, academic training - to tell me what the world looks like to them right now. Bluesky does this. /3
comment in response to
post
This is related to my somewhat vague intuition that an understandable AI is one I can reliably lie to to advance my own ends at its expense, which shares the idea that "understanding" a thing suggests the ability to manipulate it at more than one level and anticipate its behavior.
comment in response to
post
I feel like understanding chess would include being able to also play variants of chess, or solve weird chess-adjacent puzzles, like play on an infinite board or with a range limit on the queen or something. It would get that the standard 8x8 32-piece game is a special case of "chess-ness".
comment in response to
post
Ooh, another relevant hook for one of my favorite essays!
theoutline.com/post/1611/th...
comment in response to
post
But also just now found a PDF archive of them.
blogs.cornell.edu/collinslab/2...
Not as mathy as the arxiv thing, but feels adjacent, at least to me.
comment in response to
post
This sent me on a wild goose chase to find a "metabolic pathways" chart that I was sure I had bookmarked, which also has strong "bewildering complexity" vibes, except it's real.
I found the site, the chart is gone.
Please accept this SMBC comic in its stead:
smbc-comics.com/comic/2013-1...
comment in response to
post
Are we so sure it hasn't? Who is reading those five-bullet e-mails every week?
comment in response to
post
I keep thinking about this scene from the original ROLLERBALL, the “libraries” of the future, that peddle in summaries of approved books. Everything summarized by computer. The interaction with the clerk is so friendly and chilling.
youtu.be/g6-rcIjKFFs?...
comment in response to
post
That was lovely, thanks. It reminds me, particularly in the efficiency-skepticism part, of a favorite essay of mine on the value of learning to code.
To be fair, many things remind me of this essay, and I rarely pass up a chance to repost it.
theoutline.com/post/1611/th...
comment in response to
post
Bookmarks!
comment in response to
post
With alt text.
comment in response to
post
Probably there's a doctoral thesis in there to gather evidence for or against this model. I'd love for someone to dig in some day, or to learn if someone already has.
(5/5)
comment in response to
post
And they have bills to pay too, and they want to scale and automate.
"Any sufficiently commercial technology is indistinguishable from television" happens, and now it's all flat and bland and family-friendly, and here we are.
(4/n)
comment in response to
post
First-movers do OK for a while, then Google shows up with their algorithm, and it's awesome, until everyone starts to game it, with the same flattening effect -- they're just another aggregator, really.
(3/n)
comment in response to
post
Then, once there are aggregators, they have an incentive to scale to broaden the audience for the ads that they serve to cover their costs.
Incentive to scale means incentive to flatten for the most broadly desirable content.
(2/n)