4 ms·

How could it ever be considered a protocol? It's a platform - a marketplace for buying opinion.

by flessner 2y ago

RIMR 2y ago
[flagged]
- dang 2y ago
  (...plus came up with Bayesian spam filtering, plus wrote the book on Lisp macros, plus revolutionized startup investing). It's only astounding because your assumptions are false. pg is nothing like a pointy-haired boss. What he is is a highly curious and lazy (in the good sense of the word) hacker who is bored by busywork. How you managed to arrive at the inverse image of that is such a feat of pathfinding that I'd be interested in the steps by which you got there.
  - andrewmcwatters 2y ago
    Because it's a terrible blog post. If you applied this criticism to any other author, it would be valid. But because it's pg it's different? No, it's still a bad post. There are a plethora of other reasons Twitter was a big deal. It being a "protocol" wasn't one of them.
    - dang 2y ago
      Getting from that post to the GP's dramatic assessment of PG as a person would be a disappointing feat of pathfinding.
    - tptacek 2y ago
      I don't think the post has held up, but the "it's not a protocol, it's just HTTP on top of TCP/IP" is a lame argument. It's clearly a protocol. I've been doing protocol engineering work since the mid-1990s, and people have been saying things built on top of HTTP aren't "protocols" since HTTP went mainstream. I was one of them, in the 1990s! That was dumb of me; most new important protocols since then have been built on top of HTTP, and I expect that to continue. The subtext of these "it's not even a protocol" arguments are that Paul Graham doesn't know what a protocol is, which is not a plausible argument. Why make it?
  - hoahmarineman 2y ago
    [dead]
  - hoahmarineman 2y ago
    [flagged]
    - dang 2y ago
      It's true that when I'm fond of somebody, I tend to respond to false attacks on them. Not because of "meal tickets" but just human feeling. It's true that things look different on the outside, though one might add that people who routinely jump to cynical conclusions about others don't make very good Hacker News commenters. But what made you think I know anything about neuroscience?
    - AdieuToLogic 2y ago
      > We get it though; gotta white knight for your meal ticket This is a rather presumptive conclusion. Unless, of course, you have specific knowledge corroborated by others such that this assertion is more than a trite ad hominem. Full disclosure: I am not associated with anyone who owns, administers, or in any way runs this site. Nor am I "white knighting for a meal ticket."
  - lolinder 2y ago
    Dang, I really respect your work here and in general you do a great job, but I think you overstepped here in auto-collapsing this comment thread and replying (edit: To clarify for anyone coming to this later, the GP comment was collapsed by moderator action at the time of writing. It is now flagged by user flags, which I think is entirely appropriate.) PG gets a lot of flack on HN, some comments better-considered than others. Most of us are able to tell the difference and file the mindless attacks appropriately. We don't need you to rush to his defense, and in fact you doing so is likely counterproductive. You've often said that you take a policy of moderating less, not more, when YC is involved. This interaction and moderation action breaks that pattern, which is harmful.
    - dang 2y ago
      It seems plainly offtopic (and flamebaity and name-calling) to me, but ok, I've uncollapsed the subthread. As for replying: sorry, but as I said at https://news.ycombinator.com/item?id=42763996 https://news.ycombinator.com/item?id=42763996, when people I'm fond of are maligned, I'm going to respond. That has nothing to do with YC, that has to do with being human. (Another recent example was https://news.ycombinator.com/item?id=42685972 https://news.ycombinator.com/item?id=42685972.)
      - Dalewyn 2y ago
        The problem here is the mixing of administrative powers (collapsing of subthread) and expression of private opinions. One is ideally an unbiased, mechanical action subject to a rigid set of publicized conditions to which the comment(s) concerned are applied. So-and-so comment(s) are moderated so-and-so because they violate so-and-so in the guidelines, for example. The other is (by your own admission) a biased, emotional, personal action subject and liable only to yourself. The two are mutually incompatible when performed together.
        dang 2y ago
        I can't (and don't want to) do this job purely mechanically, and never have. I doubt it's possible, and if it is, I doubt it would make for good moderation. Actually, though, collapsing the GP subthread was just that sort of application of the site guidelines. It's obvious (IMO) that the subthread is flamebaity and well offtopic. I reversed that decision as a courtesy to lolinder and a nod to the "moderate less when YC is involved" principle—even though it was the correct call from the unbiased/mechanical/rigid side of the ledger.
- tayo42 2y ago
  > Twitter was just HTTP over TCP/IP. It was never a protocol. It was a website. I think it was also a text service in the very beginning.
  - sroussey 2y ago
    Which was the reason for the short text length of posts.
- Zak 2y ago
  > He only ever got SSL working on his personal website in 2023. I also added SSL to a site "too late" by hivemind standards. It's static HTML and contains nothing sensitive. I guess maybe a malicious ISP could theoretically inject ads or something. I don't think he believed it was a protocol in a literal sense, but that people were using it like one. It had an open API at the time and both production and consumption of tweets was often automated. It didn't really work out that way longer-term, but it wasn't crazy to guess that it might.
  - likeabatterycar 2y ago
    It wouldn't be HN if it's been six hours without attempting to shame someone with HTTPS demagoguery.
sctb 2y ago
My dictionary says this about "protocol": > [In computing:] a set of rules governing the exchange or transmission of data between devices. In the article, pg says this: > The reason is that it's a new messaging protocol, where you don't specify the recipients. It seems obvious to me what pg is getting at, even though the other protocols he mentioned are all formal while Twitter's is not.
- verzali 2y ago
  It seems like that could apply to many other ways of messaging though, not all of them online. I'm not sure the "new" bit is really correct. Perhaps the potential scale or reach of a message is what matters.
eleveriven 2y ago
Never thought of it this way... It truly is a marketplace where opinions can be "bought"
fergie 2y ago
> a marketplace for buying opinion Never thought of it like that before

How could it ever be considered a protocol? It's a platform - a marketplace for buying opinion.

RIMR 2y ago

[flagged]

dang 2y ago
(...plus came up with Bayesian spam filtering, plus wrote the book on Lisp macros, plus revolutionized startup investing). It's only astounding because your assumptions are false. pg is nothing like a pointy-haired boss. What he is is a highly curious and lazy (in the good sense of the word) hacker who is bored by busywork. How you managed to arrive at the inverse image of that is such a feat of pathfinding that I'd be interested in the steps by which you got there.
- andrewmcwatters 2y ago
  Because it's a terrible blog post. If you applied this criticism to any other author, it would be valid. But because it's pg it's different? No, it's still a bad post. There are a plethora of other reasons Twitter was a big deal. It being a "protocol" wasn't one of them.
  - dang 2y ago
    Getting from that post to the GP's dramatic assessment of PG as a person would be a disappointing feat of pathfinding.
  - tptacek 2y ago
    I don't think the post has held up, but the "it's not a protocol, it's just HTTP on top of TCP/IP" is a lame argument. It's clearly a protocol. I've been doing protocol engineering work since the mid-1990s, and people have been saying things built on top of HTTP aren't "protocols" since HTTP went mainstream. I was one of them, in the 1990s! That was dumb of me; most new important protocols since then have been built on top of HTTP, and I expect that to continue. The subtext of these "it's not even a protocol" arguments are that Paul Graham doesn't know what a protocol is, which is not a plausible argument. Why make it?
- hoahmarineman 2y ago
  [dead]
- hoahmarineman 2y ago
  [flagged]
  - dang 2y ago
    It's true that when I'm fond of somebody, I tend to respond to false attacks on them. Not because of "meal tickets" but just human feeling. It's true that things look different on the outside, though one might add that people who routinely jump to cynical conclusions about others don't make very good Hacker News commenters. But what made you think I know anything about neuroscience?
  - AdieuToLogic 2y ago
    > We get it though; gotta white knight for your meal ticket This is a rather presumptive conclusion. Unless, of course, you have specific knowledge corroborated by others such that this assertion is more than a trite ad hominem. Full disclosure: I am not associated with anyone who owns, administers, or in any way runs this site. Nor am I "white knighting for a meal ticket."
- lolinder 2y ago
  Dang, I really respect your work here and in general you do a great job, but I think you overstepped here in auto-collapsing this comment thread and replying (edit: To clarify for anyone coming to this later, the GP comment was collapsed by moderator action at the time of writing. It is now flagged by user flags, which I think is entirely appropriate.) PG gets a lot of flack on HN, some comments better-considered than others. Most of us are able to tell the difference and file the mindless attacks appropriately. We don't need you to rush to his defense, and in fact you doing so is likely counterproductive. You've often said that you take a policy of moderating less, not more, when YC is involved. This interaction and moderation action breaks that pattern, which is harmful.
  - dang 2y ago
    It seems plainly offtopic (and flamebaity and name-calling) to me, but ok, I've uncollapsed the subthread. As for replying: sorry, but as I said at https://news.ycombinator.com/item?id=42763996 https://news.ycombinator.com/item?id=42763996, when people I'm fond of are maligned, I'm going to respond. That has nothing to do with YC, that has to do with being human. (Another recent example was https://news.ycombinator.com/item?id=42685972 https://news.ycombinator.com/item?id=42685972.)
    - Dalewyn 2y ago
      The problem here is the mixing of administrative powers (collapsing of subthread) and expression of private opinions. One is ideally an unbiased, mechanical action subject to a rigid set of publicized conditions to which the comment(s) concerned are applied. So-and-so comment(s) are moderated so-and-so because they violate so-and-so in the guidelines, for example. The other is (by your own admission) a biased, emotional, personal action subject and liable only to yourself. The two are mutually incompatible when performed together.
      - dang 2y ago
        I can't (and don't want to) do this job purely mechanically, and never have. I doubt it's possible, and if it is, I doubt it would make for good moderation. Actually, though, collapsing the GP subthread was just that sort of application of the site guidelines. It's obvious (IMO) that the subthread is flamebaity and well offtopic. I reversed that decision as a courtesy to lolinder and a nod to the "moderate less when YC is involved" principle—even though it was the correct call from the unbiased/mechanical/rigid side of the ledger.
tayo42 2y ago
> Twitter was just HTTP over TCP/IP. It was never a protocol. It was a website. I think it was also a text service in the very beginning.
- sroussey 2y ago
  Which was the reason for the short text length of posts.
Zak 2y ago
> He only ever got SSL working on his personal website in 2023. I also added SSL to a site "too late" by hivemind standards. It's static HTML and contains nothing sensitive. I guess maybe a malicious ISP could theoretically inject ads or something. I don't think he believed it was a protocol in a literal sense, but that people were using it like one. It had an open API at the time and both production and consumption of tweets was often automated. It didn't really work out that way longer-term, but it wasn't crazy to guess that it might.
- likeabatterycar 2y ago
  It wouldn't be HN if it's been six hours without attempting to shame someone with HTTPS demagoguery.

sctb 2y ago

My dictionary says this about "protocol": > [In computing:] a set of rules governing the exchange or transmission of data between devices. In the article, pg says this: > The reason is that it's a new messaging protocol, where you don't specify the recipients. It seems obvious to me what pg is getting at, even though the other protocols he mentioned are all formal while Twitter's is not.

verzali 2y ago
It seems like that could apply to many other ways of messaging though, not all of them online. I'm not sure the "new" bit is really correct. Perhaps the potential scale or reach of a message is what matters.

eleveriven 2y ago

Never thought of it this way... It truly is a marketplace where opinions can be "bought"

fergie 2y ago

> a marketplace for buying opinion Never thought of it like that before