Live data from Hacker News

First, make me care

gwern.net

171–180 of 262 posts

Re: First, make me care

#171
post #4

This is something I find fascinating about TikTok: on that platform you literally get a few seconds to catch the attention of your audience before they skip to the next video. You can't just find one hook that works and reuse it forever because people will get bored of it - including if that hook is heavily used by other accounts. This makes TikTok a fascinating brute-force attack on human psychology, with literally…

> You can't just find one hook that works and reuse it forever

Hmm. I feel exact opposite. Most of successful channels that i see are using exact same formula/structure/often even style time and time again.

Re: First, make me care

#172
post #103

Earlier quoted context omitted.

I feel the same. I found that blog from SSC/ACX and I still much prefer SSC/ACX despite gwern discussing topics that are much much more relevant to my interests (sw dev, Haskell, anime). I can't formulate why but your analysis sounds close enough.

ACX is honestly an incredible writer. I think it's a very high bar to clear.

Yes, Scott has a natural talent for writing (there's an article where he explicitly says he isn't proud of his prose because it just comes naturally to him, but is immensely proud of getting a B in Calculus back in pre-med, or something along these lines).

Re: First, make me care

#173
post #84

Earlier quoted context omitted.

Isn't this pretty much the definition of a meme? I mean before meme just became synonymous with funny cat videos. Like the actual meaning of the word.

Snow Crash explored this much more literally, supposing that there may be memes so powerful they can function basically as magic spells that reprogram people's brains.

The SCP Foundation pages[0] have something similar, a danger classification for "Memetic Hazards" which are basically informational viruses that affect memory, cognition, and perception.

[0] - https://scp-wiki.wikidot.com/understanding-memetics

Re: First, make me care

#174
post #73
post #4

This is something I find fascinating about TikTok: on that platform you literally get a few seconds to catch the attention of your audience before they skip to the next video. You can't just find one hook that works and reuse it forever because people will get bored of it - including if that hook is heavily used by other accounts. This makes TikTok a fascinating brute-force attack on human psychology, with literally…

Except the hooks only attach to the lizard brain while the rational brain just sits there with a palm in its face.

There is no lizard brain. The "triune brain" theory has been debunked by modern neuroscience for years.

Re: First, make me care

#175

Earlier quoted context omitted.

Not classic enough. You need to add an actual explanation if you want that comment to work. What are the multiple meanings to the same phrase? Presumably "weird and fun" is what you're calling out? But to me their post looks like it's using the exact same meaning both times.

> Presumably "weird and fun" Yes, you figured it out, so apparently I didn't need an "actual explanation". > looks like it's using the exact same meaning both times. That's a fascinating equivalence operator you're using.

Guys it's not that serious

Re: First, make me care

#176
post #52

You know I’ve never read an article by Gwern that made me feel like he was sensitive to this idea, one that in my head essentially breaks down to the use of narrative and the leverage of “stakes” that inform the reader of kinds of conflict that make a narrative special. I’m reminded of a remark made by David Foster Wallace (on KCRW? Or oft-repeated elsewhere) about how he had to come to terms with the purpose of writ…

Gwern has hands down one of the worst blogs, readability wise, ever created on the internets. His writing style can be hit or miss too.

I agree, but I think I know why I personally have this opinion. I don't like reading hypermedia, and Gwern is all about hypermedia.

Hypermedia is fine when you're reading reference material, like Wikipedia. We've all done https://xkcd.com/214/, but at some point you just learn to tune out the hyperlinks.

When reading an article written by a single human, I want it to have a well-defined linear structure. I don't even like footnotes or pull-out quotes. Gwern likes to put a "blind" hyperlink or two in literally every sentence. Here's what an "orthoxerox-optimized Gwern article" would look like:

- blind links to literal Wikipedia: gone, I can search for more information myself - blind links to external websites: please just pull the relevant information into the body of the article or, if that's impossible, spend a sentence on why you want me to click it - blind links to other pages on the website: again, if it's some relevant information, please just pull it into the body of the article; if it's self-promotion, I can live with links like these if they are the only blind ones left, but the "Similar links" box under the article is already there

Re: First, make me care

#177
What I find extremely off-putting and overused is the pattern of making you care about an article by saying something about the person being interviewed, usually related to the interview itself. Think "he was a balding man[...] drinking his matcha latte[...]" It's always something which has zero bearing on the situation in question.

Whenever I see this, I immediately turn to cmd-a + cmd-c + `pbpaste | llm 'summarize this'`

Re: First, make me care

#178
That’s basically what tv shows do every time. The pilot is great/awesome then the 10 episodes are boring and the very last one get exciting enough for you to wait the second season.

Best example: the walking dead season tv shows.

Re: First, make me care

#179
For the last 30 years I've decided that the best stuff (most engaging books, stories, experiences in life) require investment. It gets worse (you go through some pain while exercising) before it gets better. That's essentially a definition of a good life to me - finding the things worth sacrificing resources and getting the payoff.

So 'first make me care' to me is a manifest of Gen Z - tiktok - brainrot approach. From my perspective you miss most of the really good stuff by cultivating this approach. I.e. my favorite books - Tai Pan, Noble House; tv series - Better Call Saul! - require you to go through so much of initial boredom. It's also the same discussion as 'learn to code vs only do AI Slop' or 'learn math and algos vs only import functions from libs and never check what's inside'.

*Exceptions apply, ofc. There are things that hook you and progressively ad depth, but it's really rare. I.e. Arcane tv show is both easy to access and quite deep.

Edit: ...so I can imagine math teacher that first tell you what are some amazing uses of derivatives and integrals - PIDs, SGD, better estimation, wave functions, generalized description of problems, accessing interesting physics etc. And after that they make you grind. I think it would be quite great. But it is so rare, that you have to make a leap of faith and assume most of the good stuff is boring initially.

Post reply on HN