In computer science we occasionally – and not often enough in the past – refer to how well an algorithm scales, which means asking how does the algorithm react across the range of data, as measured by size, it encounters? Does it crash? Does it grind to a halt? Does it limp along with performance in the toilet (sometimes known as the Nx problem, meaning the algorithm’s performance, as related to N, the amount of data, is roughly Nx, where x > 1), or is it good (x == 1), or even very good (x < 1)? One of the difficulties in calculating the potential scaling is communications. Are you talking to another server or a database that’s not next door, but in Paris/Shanghai/Seattle/all three? How does that affect scaling?
The reason I bring this up is an observation by Richard Hanania:
People go into academia and journalism for generally idealistic reasons. Some conservatives might be turned away from these professions for political reasons, which poses a “chicken or egg” problem. In my experience though, a smart young person going into journalism is probably better off going into conservative media than they are liberal media, which is already saturated with people with elite degrees who cannot find stable employment. There’s a great deal of demand for conservative journalism among the general public, but few competent conservatives who want to be journalists given what the profession pays relative to what else smart people can be doing. Thus conservative media tends to see the rise of completely incompetent outlets like OANN, which posts fake COVID cures when it’s not arguing the whole thing is fake.
While he doesn’t cite any studies, intuitively it seems correct. To my eye, this is an example of scaling gone mad. The data, in this case, is the audience size; the algorithm is the journalists.
And the Internet is the communications channel that replaced broadcast TV and local newspapers.
If the algorithm is efficient then fewer resources – fewer journalists – need to be dedicated to it.
But the result is not better journalism to more people. While, yes, you can get WaPo or The New York Times – both of which some argue are falling on hard times, quality-wise – even better than before, and these were quality national newspapers prior to the Internet, you can also get sucked into such propaganda outlets as RT, Fox News, OANN, and an absolute myriad of scam sites that deliver false news or malicious analysis of news.
In a sense, nothing new here. I was simply struck by the parallels between my world and the journalism world as observed by Hanania. But analogies occasionally end up being more than illustrative: they can produce solutions that are not obvious, which is known, as I understand the terms, as isomorphic.
So maybe someone will see a solution in this analogy.
