Kellan Elliott-McCrea writes about software engineering culture, technical debt, social network federation, and the complexity of modern software.
HN links it for essays on software complexity and the craft of coding; commenters cite its series on software's discontents.
I see Twitter getting beaten up a lot for not deleting the spammers faster. Etsy gets beaten up for not deleting the “resellers” faster. Flickr used to get yelled at for not catching the photo stealers or porn spammers faster.
“It’s so fucking easy, they’re right over there, here, let me show them to you, what’s your problem?”
This comes from not understanding the cost benefit ratio of false positives in identifying abuse of a social site at scale.
Imagine you’ve got a near perfect model for detecting spammers on Twitter. Say, Joe’s perfectly reasonable model of “20+ tweets that matched ‘^@[\w]+ http://’” >> https://twitter.com/#!/joestump/status/94845660163088384. Joe is (presumably hyperbolically) claiming 99% accuracy for his model. And for the moment we’ll imagine he is right. Even at 99% accuracy, that means this algorithm is going to be incorrectly flagging roughly 2 million tweets per day as spam that are actually perfectly legitimate.
If you’ve never run a social software site (which Joe of course has, but for the folks who haven’t) let me tell you: these kinds of false positives are expensive.
They’re really expensive. They burn your most precious resources when running a startup: good will, and time. Your support staff has to address the issues (while people are yelling at them), your engineers are in the database mucking about with columns, until they finally break down about build an unbanning tool which inevitably doesn’t scale to really massive attacks, or…
> I think that Berners-Lee and W3C get way too much credit for the web. And this is an incredibly arrogant statement by
Mr. Berners-Lee.
I also think that green is a totally awesome color, and that we
can breed birds with horses to produce pegasii. For more attacks
on the W3C at best tangential to Mr. Berners-Lee's actual essay,
read on.
> Facebook has in-fact done some serious web innovation the last
years and W3C has completely dropped the ball. Facebook and
Twitter have been catalyzing internet adoption and the general
spread of information.
I'll conveniently ignore companies like Flickr, who innovate without compromising data portability: http://laughingmeme.org/2010/05/18/minimal-competence-data-access-data-ownership-and-sharecropping/
> Mr. Berners-Lee's obsession about content-silos shows that there
is a serious disconnect between the current state of the web and
W3C. The web was about content and documents fifteen years ago,
now it's about the flow of data.
Those URI thingamabobs that send you to content or documents?
The web isn't about that anymore. It's about a golden shower of
data, flowing down the firehose. You see, when people look up stuff on
Wikipedia, they don't care about the page they're on, but the
activity data on who's flowing in and out of it.
> I know Berners-Lee is a big Linked Data advocate, but the
approach that's being taken by the W3C is painfully slow and
doesn't take into account the fluidity of information.
See I took "flow of data"…
This sounds right to me:
> Before AI, both camps were doing the same thing every day. Writing code by hand. Using the same editors, the same languages, the same pull request workflows. The craft-lovers and the make-it-go people sat next to each other, shipped the same products, looked indistinguishable. The motivation behind the work was invisible because the process was identical.
Helps explain why some people are delighted to have AI write code for them while others are unhappy that the part they enjoyed so much has been greatly reduced.
Similar note from Kellan (a clear member of the make-it-go group) in https://laughingmeme.org/2026/02/09/code-has-always-been-the-easy-part.html :
> That feeling of loss though can be hard to understand emotionally for people my age who entered tech because we were addicted to feeling of agency it gave us. The web was objectively awful as a technology, and genuinely amazing, and nobody got into it because programming in Perl was somehow aesthetically delightful.
Rachel by the Bay (https://rachelbythebay.com/) has long impressed me as someone who clearly is deep in the actual work of systems, day in and day out, and can write well about it.
Julia Evans has a wonderful approach as well, and has amazing talent for teaching: https://jvns.ca/
Kellan Elliott-McCrea (https://laughingmeme.org/) has given the world some of the better advice on the hardest parts of software scaling, which is of course scaling the human organizations. New grads are virtually always underestimating that part of the work; eventually you realize the hard problems are usually social and not technical.
But when we render the Comment back to you in that same request we need
the ID that the comment will have, but we don't know the ID until we write it out.
Wouldn't something like Snowflake help for this particular case?
Snowflake is a network service for generating unique ID numbers
at high scale with some simple guarantees.
http://github.com/twitter/snowflake
Kellan (from Flickr) has a neat post about Ticket servers: http://laughingmeme.org/2010/02/08/ticket-servers-distributed-unique-primary-keys-on-the-cheap/
Your understanding of what is difficult is wrong because you're looking at it as a detection problem when it's actually a false positive problem.
See: http://laughingmeme.org/2011/07/23/cost-of-false-positives/