I find a commonality between many articles I read online:
1. An interesting headline that draws me in.
2. A long setup/history which I start reading, then skimming, then just scrolling past
3. And finally, the meat of the article.
But by 3. I've decided that the article doesn't contain any interesting info - otherwise, why would I be scrolling past the first 2/3s of it?
I'm aware this is a subconscious decision, a bad heuristic. But in a world filled with info, I rarely fight against it. I didn't read this article, for instance. My coffee break is over, it's time to get back to work.
Maybe it's time for blogs to start doing something similar to what the better recipe sites do, and provide a link at the top to skip the setup and get to the real article.
As someone who agrees with the parent commenter, I don't think (at least for me) it's individual style, but it's what my goal is and what the content is. I think the main problem is blog posts, which basically contain a single main thesis, that are written like short stories.
I think that if, as an author, your goal is to get a primary point across, and you're publishing on the web, for the love of God, get to the point. There is a near infinite amount of content online vying for our attention, and if I'm reading for information, I don't want to spend a ton of time just figuring out if your point is worth my time.
On the other hand, if you're telling a story, I'm fine getting drawn in with elaborate prose. But at least for me, most of the time I'm looking for this style of writing in a book, or at least something like a magazine article that is positioned as a short story. I don't often look for this type of content on the web, and I almost never look for it in blog posts.
No offense taken, of course the common factor is me! That was the entire point of my comment. I don't enjoy articles like this.
I was wondering if other people have the same feeling. It seems like many do, to the point that there's an actual rule of journalism (inverted pyramid) stating that you should avoid crafting articles this way.
The point of this post is the feel, the narrative, in service of strengthening the central metaphor of internet as untilled land dominanted by fortresses (social media). The tl;dr is: don't use Facebook, make websites. Maybe make a decentralized search/discovery protocol (or something?).
What you describe almost the opposite of the classic "inverted pyramid" style of writing journalism. I agree it's a bad style if you want to have busy people reading your site.
Cut to the chase, and put the interesting-to-some-but-not-all stuff at the bottom, or better still, behind a link. We have hypertext, guys. That's what it's for.
> I'm aware this is a subconscious decision, a bad heuristic.
I don't think it's a bad heuristic at all. There are far more web pages than anyone could possibly read in their limited lifespan. It makes sense to optimize for sites that get to the point right away.
Note that I'm speaking here of purely informational sites, not "literary" sites where the beauty of the prose is intended to be part (or even all) of the purpose. Different rules apply there.
Let's stop with the abuse of the word "algorithm". Enough already. Even in a world where it wasn't the wrong word to reach for (i.e. a world that isn't the one we're in), even if it were proper, it is still by this point embarrassingly hackneyed. It's shameless, unrepentant pablum. And that's still ignoring that it's the wrong word to use in >90% of the instances where it seems to show up these days.
I have a pretty simple flowchart for lay people and tech journalists alike. It goes like this:
Are you about to use the word "algorithm" in an article or comment about social media?
No? -> GOOD.
Yes? -> Don't. (And to go one further, since you were already angling that way, maybe you should reconsider whether or not to scrap the whole piece entirely; do you actually have anything new and interesting to say on the topic?)
Even if there were no consensus for an alternative, that wouldn't make "recommendation algorithm" acceptable, because it fails to meet the criteria to be an algorithm. The word means something.
cf "Imagine you're a mathematician, and fellow mathematicians start calling derivatives integrals instead[...] it's being used to describe systems that are almost the exact opposite"
Without even touching how every other thing you said is wrong (esp. about it being accurate), how is "algorithm" more descriptive than "artificial intelligence"?
>: a procedure for solving a mathematical problem (as of finding the greatest common divisor) in a finite number of steps that frequently involves repetition of an operation
>broadly : a step-by-step procedure for solving a problem or accomplishing some end
Right. AI is artificial intelligence. But that's irrelevant; you're answering a question I (explicitly) didn't ask, and not the question I did ask. How is "algorithm" a more descriptive term (your words) than "artificial intelligence"?
Do you have a point here? Whatever you think it is, you're not making it—you're making my case. Social media companies' "algorithms" aren't algorithms.
AI is artificial intelligence, "AI" is artificial but it's not intelligence.
ChatGPT, Gemini, Stable Diffusion, LLAMA, et al. are not AI, they are "AI".
>Do you have a point here? ... Social media companies' "algorithms" aren't algorithms.
Social media run on servers, servers run software, software is quite literally "a procedure for solving a mathematical problem in a finite number of steps that frequently involves repetition of an operation" and "a step-by-step procedure for solving a problem or accomplishing some end"
Please come back when you at least have a basic understanding of how computers work.
Nice try, but no. Everything that runs on a server is not ipso facto an algorithm. It's a word that has a specific meaning, which you would (I'd hope, at least) be able to grasp if you weren't ignoring half the words in the definition you already quoted (and the specific meanings attached to those words, too).
> AI is artificial intelligence, "AI" is artificial but it's not intelligence.
And artificial tears aren't tears. Likewise with artificial vanilla, artificial plants, etc. Which is beside the point—because you have yet again attempted to provide a (crummy) response to the question I didn't ask and avoided answering the one I actually did ask.
Please come back when you can demonstrate that you can carry on a conversation without this type of rhetorically dishonest, empty, pigeon-shitting-on-the-chessboard condescending bluster. (And maybe take a course on the theory of computation and the fundamentals of computer science in the meantime.)
> You proposed we stop using the word "algorithm", I argued to the contrary
K, thanks for the review. It was neither necessary, informative, nor otherwise a response to anything that I asked.
You can make as many bare assertions about being right about the accuracy question as you want. No amount of them is going to transmute these sorts of replies into one that can't be plainly identified as avoiding the question I did ask, but you can keep trying if you want.
> You asked:¶ >Let's stop with the abuse of the word "algorithm".
That's not a question. For someone with such strong feedback in a discussion about words and their meanings, you sure have a lot of difficulty actually working with them, huh?
* * *
In one double-barreled remark, you managed to pack in four claims—three explicit, and one implicit. The three explicit ones being that, in comparison to "artificial intelligence", "algorithm" is:
- "more accurate"
- "more descriptive"
- a "better"(?) word choice.
Ignoring the dumb false dichotomy that you pulled out of thin air and the ahistoricity of the other claim about "AI" succeeding "algorithm" when it comes to social media, I have repeatedly asked you to defend the second claim. You insist on doing anything but that—mainly by focusing on the first (and then reverting to posting low-effort, overconfident insults like, "Please come back when you at least have a basic understanding of how computers work").
> Without even touching how every other thing you said is wrong (esp. about it being accurate), how is "algorithm" more descriptive than "artificial intelligence"? <https://news.ycombinator.com/item?id=40235433>
> you're answering a question I (explicitly) didn't ask, and not the question I did ask. How is "algorithm" a more descriptive term (your words) than "artificial intelligence"? <https://news.ycombinator.com/item?id=40242056>
>> And those who cultivate those plots outside of these walls face pressures to conform to the whims of the businesses in hopes that the pathways remain open. Otherwise, they might toil away in silence, rarely seeing visitors like they one day used to.
I feel like this paragraph sums up the root of the issue. IF your main concern is "I want lots of people to like my work, and by extension like me, and by extension validate my existence" then naturally we gravitate (literally a force of gravity) towards single common places.
Since most people most definitely fall into this camp, common platforms win. If no-ones gonna read my post on Google+, why on earth would I share my profound ideas there? If the worth of my product is based on the number of followers and likes then it's logical to go to the place where those metrics are amplified.
The nostalgia for the "good old days" is ultimately rooted in golden memories (sometimes, admittedly, delusional) of when we felt noticed, and valued, and cared for. For many of us that is "before adulthood" for some definition of "adult".
Put another way, there are a LOT more high-school football stars than NFL stars.
Of course the web of our youth -hasn't gone away- . You are as free as ever to build a blog, write your thoughts, as you ever were. In fact it's easier than it ever was. But how will people like it? Friend you? Follow you?
Not literally. Gravity is not pulling me towards Facebook and Twitter. Maybe it's similar if you consider big social networks as having lots of mass, but it's still not literally a force of gravity.
Search engines ruined by their SEO spam designed "to pull in money via ads and affiliate links", Social networks and their "constant push to satisfy...an algorithm tuned to keeping eyeballs on the platform", newspapers and their "aggressive paywalls and expensive subscription models", the ads, the trackers, the popups, and the chatbots - the enshittification of the internet in all of its forms can be traced back to single source. The root of all evil. Greed is the problem and we're no better equipped to deal with it online than we are offline.
It seems like everywhere people act selfishly and try to take as much as they can from everyone else while giving back as little as possible things only get worse for everyone, but you can't stop people from falling into that trap. It's like some kind of mental disorder. A person suffering from greed will knowingly poison themselves and their families. It's caused people to pollute the air they need to breathe and the water they need to drink. We'll never convince someone afflicted with Greed to stay away from our websites.
Any new internet protocol we create, any new platform we develop, any new community we build will be slowly corrupted the moment it gets popular enough to be useful. It will probably start even before then. Speculators will try to get an early foothold to gain an edge over their soon-to-be competitors. The playbook has been written, and now it's just a matter of following the steps. Embrace, Extend, and Extinguish.
We have many advantages compared to the people who created the internet. We have better tech. We have better access to that tech. We have the examples they left us to build from. We also have things they never had to deal with. The technology they used was designed to work for them. All of our technology is designed to be used against us by other people.
They were creating the infrastructure to communicate freely. All of our infrastructure is already captured and monitored. It's designed to get in our way if we aren't using it the way others want us to. Simply hosting a website likely violates your ISPs TOS. Using Tor or a VPN means some webpages never load and those that do harass you to solve riddles to improve the AI they'll use against you.
The creators of the internet never had to deal with AI bots pretending to be normal users spreading astroturf and shilling for their masters. Real content takes time and effort but ads can be spammed by bots which never sleep until those ads drown out everything else.
The only way to get back something like the internet before it was infested with greed is to keep most people off of it. Small communities can be controlled and moderated. In a small community the identities of users can be verified to keep bots away. It doesn't solve the problem of carefully planted shills or paid off members, but it keeps the problem a little more contained and manageable. The problem is that what we'd be left with in that case wouldn't really be the internet. Just another walled garden, one to keep the money away. At best we might be able to pull off going back to the time when you could find a really cool BBS, but the old internet was a once in a lifetime thing.
counter-counterpoint: you already have a different web, it's called Geminispace[0]. Brutally simple by design, text and hypertext only, built on a protocol frozen in amber. Antimodern, antiprogress, anticapitalist, antinormal. It will never grow beyond niche because it's too simple to be useful to the mainstream. Hell, it's too simple to be useful to most of Hacker News, which is saying something.
If it's gonna focus on text only, that tiny low-contrast default sans font is a terrible, terrible choice. After reading a few pages of that, your readers will all be in bed with eyestrain.
Then you can change the font (and colours). A Gemini file does not specify fonts/colours, so it is possible for the user to specify them (according to the client software being used).
Well, yeah, obviously. That's not the point. That won't make the linked site itself any more legible.
I was interested enough to click through to the second page, but stopped reading after a few paragraphs. It's been a long day and my eyeballs are already pretty toasted.
If they want people to use (or even consider) their stuff, they need to make it not look like ass.
Edit:
Consider the non-technical reader who stumbles across this. Likely reaction: "Sounds interesting, but man, that looks like ass."
> It will never grow beyond niche because it's too simple to be useful to the mainstream. Hell, it's too simple to be useful to most of Hacker News
Different protocols, file formats, programming languages, etc can be useful for different things. If you are only writing such text that is just a simple text rather than all of the fancy stuff like HTML, CSS, etc, then it will work very well OK.
Of course, many people (including myself) do have criticisms, e.g. some people dislike mandatory TLS. (That is (one reason) why I made up my own (which also has many other differences from Gemini (and from WWW)), but that doesn't mean that Gemini cannot be used.)
Note that simple explanations are possible, e.g: Use TLS with port 1965; send the full URL and then carriage return and line feed. (In this case, the URL is "gemini://geminiprotocol.net/")
Okay, I'm interested. I'm not convinced Gemini won't suffer the same fate as everything else if it really catches on, but I am interested and it might be a lot of fun while it lasts.
My initial thoughts are that it seems just as vulnerable to the problem of bots. Popular search engines and aggregators could suffer the same issues as their new web counterparts (even leading to SEO for Geminispace), it still needs certificates, and it's a little concerning that the spec isn't finalized yet, and the whole thing is in the hands of a single person (Solderpunk, I didn't see a real name) who could sell out if Greed gets to him, but overall I like the idea of it.
Andrew Stephens rang a bell. “Save the Web by Being Nice”
My rant: read it or don’t. The way to have the new web is to be the new web. The web is here to talk to folks. I am one of the ones. So are you. Write it like you mean it. Start a blog, join a forum, add talk. It takes time. Do it. Say nice things on other blogs. Post good reviews. And learn to write well.
Check Hemingway’s 5 hints. Write short sentences. Write short first paragraphs. Use vigorous language. Be positive, not negative. Never have just five rules. (As for the fifth, Omar said, “It is my firm belief that it is a mistake to hold firm beliefs.”)
The point being, forget algorithms. Create! Make the new web. Try small words: let large words earn their keep. Be nice to those for whom you write and make the new web. For grins, I wrote my blog for 16 years; I sent 14 to archive. It’s your turn.
39 comments
[ 3.3 ms ] story [ 101 ms ] thread1. An interesting headline that draws me in.
2. A long setup/history which I start reading, then skimming, then just scrolling past
3. And finally, the meat of the article.
But by 3. I've decided that the article doesn't contain any interesting info - otherwise, why would I be scrolling past the first 2/3s of it?
I'm aware this is a subconscious decision, a bad heuristic. But in a world filled with info, I rarely fight against it. I didn't read this article, for instance. My coffee break is over, it's time to get back to work.
Maybe it's time for blogs to start doing something similar to what the better recipe sites do, and provide a link at the top to skip the setup and get to the real article.
Could it just be you don’t enjoy this style of writing?
I think that if, as an author, your goal is to get a primary point across, and you're publishing on the web, for the love of God, get to the point. There is a near infinite amount of content online vying for our attention, and if I'm reading for information, I don't want to spend a ton of time just figuring out if your point is worth my time.
On the other hand, if you're telling a story, I'm fine getting drawn in with elaborate prose. But at least for me, most of the time I'm looking for this style of writing in a book, or at least something like a magazine article that is positioned as a short story. I don't often look for this type of content on the web, and I almost never look for it in blog posts.
This. There is a reason TL;DR became a proper word.[1]
[1]: https://www.merriam-webster.com/dictionary/TL%3BDR
I was wondering if other people have the same feeling. It seems like many do, to the point that there's an actual rule of journalism (inverted pyramid) stating that you should avoid crafting articles this way.
https://en.wikipedia.org/wiki/Inverted_pyramid_(journalism)
Cut to the chase, and put the interesting-to-some-but-not-all stuff at the bottom, or better still, behind a link. We have hypertext, guys. That's what it's for.
> I'm aware this is a subconscious decision, a bad heuristic.
I don't think it's a bad heuristic at all. There are far more web pages than anyone could possibly read in their limited lifespan. It makes sense to optimize for sites that get to the point right away.
Note that I'm speaking here of purely informational sites, not "literary" sites where the beauty of the prose is intended to be part (or even all) of the purpose. Different rules apply there.
I have a pretty simple flowchart for lay people and tech journalists alike. It goes like this:
Are you about to use the word "algorithm" in an article or comment about social media?
No? -> GOOD.
Yes? -> Don't. (And to go one further, since you were already angling that way, maybe you should reconsider whether or not to scrap the whole piece entirely; do you actually have anything new and interesting to say on the topic?)
"Recommender system", e.g. <https://www.trulia.com/blog/tech/recommender-engine/>, <https://github.com/microsoft/RecAI>, etc.
Even if there were no consensus for an alternative, that wouldn't make "recommendation algorithm" acceptable, because it fails to meet the criteria to be an algorithm. The word means something.
cf "Imagine you're a mathematician, and fellow mathematicians start calling derivatives integrals instead[...] it's being used to describe systems that are almost the exact opposite"
<https://news.ycombinator.com/item?id=23672561>
2. https://www.merriam-webster.com/dictionary/algorithm
>: a procedure for solving a mathematical problem (as of finding the greatest common divisor) in a finite number of steps that frequently involves repetition of an operation
>broadly : a step-by-step procedure for solving a problem or accomplishing some end
Right. AI is artificial intelligence. But that's irrelevant; you're answering a question I (explicitly) didn't ask, and not the question I did ask. How is "algorithm" a more descriptive term (your words) than "artificial intelligence"?
> 2. https://www.merriam-webster.com/dictionary/algorithm
Do you have a point here? Whatever you think it is, you're not making it—you're making my case. Social media companies' "algorithms" aren't algorithms.
AI is artificial intelligence, "AI" is artificial but it's not intelligence.
ChatGPT, Gemini, Stable Diffusion, LLAMA, et al. are not AI, they are "AI".
>Do you have a point here? ... Social media companies' "algorithms" aren't algorithms.
Social media run on servers, servers run software, software is quite literally "a procedure for solving a mathematical problem in a finite number of steps that frequently involves repetition of an operation" and "a step-by-step procedure for solving a problem or accomplishing some end"
Please come back when you at least have a basic understanding of how computers work.
> AI is artificial intelligence, "AI" is artificial but it's not intelligence.
And artificial tears aren't tears. Likewise with artificial vanilla, artificial plants, etc. Which is beside the point—because you have yet again attempted to provide a (crummy) response to the question I didn't ask and avoided answering the one I actually did ask.
Please come back when you can demonstrate that you can carry on a conversation without this type of rhetorically dishonest, empty, pigeon-shitting-on-the-chessboard condescending bluster. (And maybe take a course on the theory of computation and the fundamentals of computer science in the meantime.)
K, thanks for the review. It was neither necessary, informative, nor otherwise a response to anything that I asked.
You can make as many bare assertions about being right about the accuracy question as you want. No amount of them is going to transmute these sorts of replies into one that can't be plainly identified as avoiding the question I did ask, but you can keep trying if you want.
>Let's stop with the abuse of the word "algorithm".
I pointed out that it's not an abuse.
If that is not the question you asked, then sincerely: What the fuck did you ask?
That's not a question. For someone with such strong feedback in a discussion about words and their meanings, you sure have a lot of difficulty actually working with them, huh?
* * *
In one double-barreled remark, you managed to pack in four claims—three explicit, and one implicit. The three explicit ones being that, in comparison to "artificial intelligence", "algorithm" is:
- "more accurate"
- "more descriptive"
- a "better"(?) word choice.
Ignoring the dumb false dichotomy that you pulled out of thin air and the ahistoricity of the other claim about "AI" succeeding "algorithm" when it comes to social media, I have repeatedly asked you to defend the second claim. You insist on doing anything but that—mainly by focusing on the first (and then reverting to posting low-effort, overconfident insults like, "Please come back when you at least have a basic understanding of how computers work").
> Without even touching how every other thing you said is wrong (esp. about it being accurate), how is "algorithm" more descriptive than "artificial intelligence"? <https://news.ycombinator.com/item?id=40235433>
> you're answering a question I (explicitly) didn't ask, and not the question I did ask. How is "algorithm" a more descriptive term (your words) than "artificial intelligence"? <https://news.ycombinator.com/item?id=40242056>
I feel like this paragraph sums up the root of the issue. IF your main concern is "I want lots of people to like my work, and by extension like me, and by extension validate my existence" then naturally we gravitate (literally a force of gravity) towards single common places.
Since most people most definitely fall into this camp, common platforms win. If no-ones gonna read my post on Google+, why on earth would I share my profound ideas there? If the worth of my product is based on the number of followers and likes then it's logical to go to the place where those metrics are amplified.
The nostalgia for the "good old days" is ultimately rooted in golden memories (sometimes, admittedly, delusional) of when we felt noticed, and valued, and cared for. For many of us that is "before adulthood" for some definition of "adult".
Put another way, there are a LOT more high-school football stars than NFL stars.
Of course the web of our youth -hasn't gone away- . You are as free as ever to build a blog, write your thoughts, as you ever were. In fact it's easier than it ever was. But how will people like it? Friend you? Follow you?
Search engines ruined by their SEO spam designed "to pull in money via ads and affiliate links", Social networks and their "constant push to satisfy...an algorithm tuned to keeping eyeballs on the platform", newspapers and their "aggressive paywalls and expensive subscription models", the ads, the trackers, the popups, and the chatbots - the enshittification of the internet in all of its forms can be traced back to single source. The root of all evil. Greed is the problem and we're no better equipped to deal with it online than we are offline.
It seems like everywhere people act selfishly and try to take as much as they can from everyone else while giving back as little as possible things only get worse for everyone, but you can't stop people from falling into that trap. It's like some kind of mental disorder. A person suffering from greed will knowingly poison themselves and their families. It's caused people to pollute the air they need to breathe and the water they need to drink. We'll never convince someone afflicted with Greed to stay away from our websites.
Any new internet protocol we create, any new platform we develop, any new community we build will be slowly corrupted the moment it gets popular enough to be useful. It will probably start even before then. Speculators will try to get an early foothold to gain an edge over their soon-to-be competitors. The playbook has been written, and now it's just a matter of following the steps. Embrace, Extend, and Extinguish.
We have many advantages compared to the people who created the internet. We have better tech. We have better access to that tech. We have the examples they left us to build from. We also have things they never had to deal with. The technology they used was designed to work for them. All of our technology is designed to be used against us by other people.
They were creating the infrastructure to communicate freely. All of our infrastructure is already captured and monitored. It's designed to get in our way if we aren't using it the way others want us to. Simply hosting a website likely violates your ISPs TOS. Using Tor or a VPN means some webpages never load and those that do harass you to solve riddles to improve the AI they'll use against you.
The creators of the internet never had to deal with AI bots pretending to be normal users spreading astroturf and shilling for their masters. Real content takes time and effort but ads can be spammed by bots which never sleep until those ads drown out everything else.
The only way to get back something like the internet before it was infested with greed is to keep most people off of it. Small communities can be controlled and moderated. In a small community the identities of users can be verified to keep bots away. It doesn't solve the problem of carefully planted shills or paid off members, but it keeps the problem a little more contained and manageable. The problem is that what we'd be left with in that case wouldn't really be the internet. Just another walled garden, one to keep the money away. At best we might be able to pull off going back to the time when you could find a really cool BBS, but the old internet was a once in a lifetime thing.
[0]https://geminiprotocol.net/
I was interested enough to click through to the second page, but stopped reading after a few paragraphs. It's been a long day and my eyeballs are already pretty toasted.
If they want people to use (or even consider) their stuff, they need to make it not look like ass.
Edit:
Consider the non-technical reader who stumbles across this. Likely reaction: "Sounds interesting, but man, that looks like ass."
My experience of geminispace was that it was far less technical (more LJ-like) than I had been expecting.
Looking like ass can be advantageous, if it gets rid of the mundanes.
Different protocols, file formats, programming languages, etc can be useful for different things. If you are only writing such text that is just a simple text rather than all of the fancy stuff like HTML, CSS, etc, then it will work very well OK.
Of course, many people (including myself) do have criticisms, e.g. some people dislike mandatory TLS. (That is (one reason) why I made up my own (which also has many other differences from Gemini (and from WWW)), but that doesn't mean that Gemini cannot be used.)
Note that simple explanations are possible, e.g: Use TLS with port 1965; send the full URL and then carriage return and line feed. (In this case, the URL is "gemini://geminiprotocol.net/")
My initial thoughts are that it seems just as vulnerable to the problem of bots. Popular search engines and aggregators could suffer the same issues as their new web counterparts (even leading to SEO for Geminispace), it still needs certificates, and it's a little concerning that the spec isn't finalized yet, and the whole thing is in the hands of a single person (Solderpunk, I didn't see a real name) who could sell out if Greed gets to him, but overall I like the idea of it.
Andrew Stephens rang a bell. “Save the Web by Being Nice”
My rant: read it or don’t. The way to have the new web is to be the new web. The web is here to talk to folks. I am one of the ones. So are you. Write it like you mean it. Start a blog, join a forum, add talk. It takes time. Do it. Say nice things on other blogs. Post good reviews. And learn to write well.
Check Hemingway’s 5 hints. Write short sentences. Write short first paragraphs. Use vigorous language. Be positive, not negative. Never have just five rules. (As for the fifth, Omar said, “It is my firm belief that it is a mistake to hold firm beliefs.”)
The point being, forget algorithms. Create! Make the new web. Try small words: let large words earn their keep. Be nice to those for whom you write and make the new web. For grins, I wrote my blog for 16 years; I sent 14 to archive. It’s your turn.