I can't believe I didn't know this before, but seeing that Medium runs a partnership program just made me realize why I often associate the site with being low quality. I will encounter it in search results, often around data science topics. The posts are usually poorly explained, with important details omitted and code snippets that don't work. It felt like the author was just padding the content out but now I can understand it; they were trying to meet metrics.
> Earnings will be based on more meaningful metrics.
It seems they've put thought into this so I will assume the best intentions here and hope that this does result in better quality!
It must be tough to decide on this because I can imagine that Substack too has eaten a lot of Medium's lunch, and a typical kneejerk reaction would be to doubledown on attracting views over quality.
There was a time (or maybe still) where some DS etc bootcamps would have their students publish medium articles as a capstone to their experience. I think both as a means of personal branding as well as researching and demonstrating subject matter competence and communication skills.
Which all sounds sensible in a vacuum, but the internet is a big, dumb machine that can’t distinguish competence (or even correctness) from expertise, and you end up with hyper specific articles like “Testing SKLearn Lasso Models inside Flask-SQLAlchemy using Mocked Spark Fixtures in Pytest” that others end up clicking on and it rises to the top of Google under some narrow search terms because of engagement, even if the substance is completely bonkers.
And then you get second order parroting of these ideas because a new grad without the experience to identify cruft actually uses them in public repos, or as the basis to hunt down and answer every SO question remotely related to this, for that sweet karma.
There are also certification authorities who offer CEUs for blogging about the subject matter. So we'll find, say, Cybersecurity bloggers who are writing about some entry-level topic, like basic networking, or types of symmetric block ciphers, and they may blog on a regular basis, because you can rack up CEUs to a certain limit by repeating a particular activity.
We need those little badges like in the 90s (e.g. made for Netscape navigator) that say “certified organic” for posts written by humans. Maybe even add author meta data to page as a signature of validity.
I'd more say "if Medium isn't aggressively stamping out AI articles, I want nothing to do with it" - I have no interest in reading them, let alone paying a fee to do so.
> Earnings will be based on more meaningful metrics.
It seems they've put thought into this so I will assume the best intentions here and hope that this does result in better quality!
It must be tough to decide on this because I can imagine that Substack too has eaten a lot of Medium's lunch, and a typical kneejerk reaction would be to doubledown on attracting views over quality.