4,867 karma · joined March 20, 2007
CEO at Medium
In general though, we love being subscription driven rather than ad driven and I don't expect that to change.
Where we are today is 30% rolled out but not even announced to our authors. That'll happen in a week and is when I think author behavior will change.
That "X is dead" trope will disappear from your recs both because we'll have better things to show and because the new incentives make it not worth writing. An article like that is very sensitive to incentives because there is no authentic reason to have written it. Probably without Medium's payment, it never would have been occurred to the person.
The subscription actually does work and could (will?) be enough to sustain us. But we didn't roll it out in a way that's congruent with anything else we were doing.
Most notably, the subscription implies that we have some premium content to share. But instead of incentivizing premium content we spent the subscription revenue to pay people to flood us with low quality click bait and content mill articles. Of course, that's not what we set out to do. But it's effectively what happened. I'm partly through reversing it.
It's sad and entirely our fault. We didn't fail but we did lose our way. Here's how I see it:
1. Lost our way on recommendations. When I showed up the company was convinced that engagement equals quality. That's not true and it gets even more pronounced if you pay people to game your recommendation system. I think we were boosting articles that made people think we were a site for clickbait. The canonical example for HN is "Why NodeJS is dead" by a new programmer with zero experience or context. Readers noticed this, but worse, so did authors. And so we lost the incentive for a lot of the best and most interesting authors to bother because they were getting swamped by content-mill type authors. As of December, about 30% of our recommendations are generated by a new system that is picking much higher quality articles that have been vetted for substance over clickbait. This is getting a lot better, rapidly.
2. Got lost thinking about the creator economy, when we should have kept thinking about doers. Distribution was our winning value proposition (on top of simple free tools). We were built to find and boost individual articles and that meant that anyone with something great to say had a chance to get their story boosted, often by a lot. This is my original background in publishing: working at O'Reilly helping them publish programming books that were written by programmers. For a lot of topics, personal experience trumps everything. Not to knock creators, but by definition full time content creation gets in the way of having personal experiences that are worth writing about. We are partly through fixing this and #1.
Those are the two most obvious ones. But then there's a longer list. We competed with our platform publishers by starting our own in house publications. Those are shut down now. We started but didn't finish a number of redesigns and so the tools didn't get better for a couple of years. We're past that now and are putting out table stakes features again and some innovations too.
What I told our investors was that there was a huge pile of shit to dig out of, but that it would be worthwhile eventually. And I still believe both that there is a lot more to do and also that it'll be worthwhile.
Eliud Kipchoge, WR in marathon, seems to only be doing one hard faster than race pace workout each week. Call that the HIT. The others are slow or race pace. What actually stands out most to me is the specificity of running close to a race pace marathon every week. https://www.sweatelite.co/eliud-kipchoge-a-typical-week-of-t...
Then on the other end of the spectrum, the Ingebrigsten family of runners seems to put an emphasis on quality. But even though they do more faster running they keep a lid on the intensity to avoid training to exhaustion. It actually reminds me more of the "Grease the Groove" concept where the faster running may be more for neurological reasons. https://www.outsideonline.com/health/running/culture-running...
Anyway, neither of these two training models seem to have much in common with HIT. The bulk of the workouts are easy intensity and they are both seemingly ware of the dangers of high intensity. That's where injuries come from and so there can't be too much of it.
Mastodon represents a short form corner of the fediverse. To participate in this corner via a medium-length tool would mean participation from people who are blindly syndicating at worst, and having a disconnected experience at best. It'd be disrespectful to the community there.
Instead, Medium authors should be fully present as short form participants, even if all they want to do is share a link.
To speak in the town square you need to be present in the town square. In the case of Mastodon, that means reading and writing from a short-form app.
The point of adding ActivityPub (both ways) would be to be part of a much, much smaller blog-length corner of the fediverse. It's interesting, but is fundamentally a different goal.
Definitely a possibility. Where we ended up landing is that Mastodon is culturally short form. From prior experience just syndicating into a short form ecosystem is kind of lame even if the ActivityPub protocol supports it. So if we have Medium authors on this part of the fediverse then we want to encourage them to be fully there.
The medium-length fediverse, i.e. blogs and articles, is still emerging and for us is also a much more complicated product question. But definitely open to it.
Chinese kids needed a 68, which could mean they got just one B. White kids needed a 62, black and latino even lower. So I totally get why Chinese families were screaming. The standards for them were way higher than for other races.
But also just as jarring, the middle schools that fed this system were incredibly biased. My middle school, Hoover, had a segregation system that split the school in half. The honors half had at most one black kid per grade. We even had a different lunch period.
The honors half was being groomed for schools like Lowell or at least to look at High School as an on ramp to college.
It's not like we didn't have a lot of black students. We actually bussed in two, maybe four busloads of black kids from another neighborhood. But only one of those was ever admitted to honors. I just can't believe that there were no kids on those busses that were smart. I knew a lot of them through sports and they were plenty smart, together, hard working.
If this level of segregation can happen in San Francisco, I can't imagine how bad it is in the rest of the country.
It's complicated, and you can have your own takeaway, but mine is that merit often defined wrong. I'm on the lookout for people who did more with less, or even did equal or near to it with less. To me, that is more a measure of merit than top line measures like test scores.
I am a programmer, I've written a programming book, I used to work at a tech publisher (O'Reilly). From O'Reilly, I know a lot about how to build a system around programmers that puts out important and accurate information. Medium started in the ballpark of that and then went in the wrong direction.
I think we can do a lot to change that. For me, change is less about changing our reputation and instead just getting the utility right.
We have syntax highlighting coming out in a week or so and I'll do a bigger post then about what is changing for technical readers and authors.