Interesting article in terms of formalizing how taxing a server is. I think a lot of experienced Netrunner players do this sort of thinking already out of habit ("It will cost him X credits to get into this remote, and I have Caprice, so he won't have enough to run twice, etc...").
The Chum example is off to me because I feel like it's exclusively being evaluated in the scenario of the runner already having the breakers installed. That's fair for the purposes of showing how Chum doesn't increase the cost of getting into a server once they have a breaker installed for the ICE after Chum, but I consider Chum to be a positional gear-check ICE. Chum is almost never, ever going to deal the net damage, because runners immediately jack out after letting Chum fire, effectively making it an ETR ICE. In order for a runner to safely run through the ICE after Chum, they have to have a decoder first to be able to safely face-check the following ICE. Then, once they know what that ICE is, they can safely let Chum fire and break the following ICE for cheaper than it would take to break Chum. At that point, Chum's job is over, and should probably be replaced with a different ICE. The corp player should keep this in mind when they think about how to position Chum, and have a plan to replace Chum as well.
The the other thing you could do is never rez the ICE after Chum, which forces them to have to always break Chum, on the off-chance you are just waiting for them to let Chum fire so you can rez your Kitsune or Data Mine or what-have-you. In that case, Chum is sort of like a 2-cost gear check (1 to rez + 1 to install). Whether anyone would actually do this, I don't know, but I mention it for the sake of completeness.
There is another card in D&D that reminds me a lot of Chum: TL;DR. It, too, is a 1-rez 4-str code gate that messes with the ICE that comes after it, but I feel like it does so more effectively because the way the ICE gets changed is based on that ICE itself, not a static buff like Chum. If you only TL;DR a Wall of Static, sure, you only increase the Path Length by 1. But suppose you TL;DR an Eli 1.0? That still only increases the Path Length by 2 using Corroder, but if the runner is breaking with Lady, then it's taxing Lady counters, which could be particularly valuable, and something the runner might consider bad enough to want to break the TL;DR before continuing to run the server.
Once TL;DR comes before an ICE with 3 subs, then we're cooking with gas. TL;DR into a Grail ICE is a fantastic way to make the server more taxing and threatening, because breaking the TL;DR and the 3 extra subs is the same cost to A Gordian Blade, or requires a datasucker for Yog.0. Net-Ready Yog can shut it down pretty hard.
Both Chum and TL;DR can also do well when considering ICE that the runner doesn't ever want to break, anyway, or maybe even isn't ever able to break. For example, Chum or TL;DR into a Gutenberg on R&D seems pretty brutal to me, since few runners have effective ways at breaking a 6-strength sentry, usually opting to just pay the trace, which is great news for TL;DR and Chum.
If you can create ICE situations where not breaking TL;DR or Chum causes much worse things to happen than simply costing the runner an extra credit or two on the following ICE, then they are worthwhile, IMO. But you kind of have to keep that in mind when assembling your ICE suite. You can't just tack these onto any old ICE suite and call it a day, you have to have a real plan for how you intend to use them.