Chris Kanich
banner
kaytwo.org
Chris Kanich
@kaytwo.org
I get unreasonably excited about Chicago, cybersecurity, education, and local impact.

https://kaytwo.org
For a very physical representation, look up at the beginning of Terminal A at MDW: www.flickr.com/photos/artfa...
The Body Of Lake Michigan
Todd Slaughter, 2002, Chicago Midway International Airport, Garfield Ridge, Chicago, Illinois, USA, sculpture
www.flickr.com
July 31, 2025 at 12:25 AM
A corollary to this is that the mental difficulty of WRITING an N page paper might end up being lower overall than generating and then fact-checking every last word, even if getting something that LOOKS complete can be done in seconds. That's what's so insidious about it
June 2, 2025 at 3:45 PM
Students would probably generate the screenshot text, then you're back at square one with respect to Brandolini's law. Call me crotchety, but this goes back to how socially unacceptable lying should be - I don't care if you used AI every step of the way, UNTIL you pass off a hallucination as true.
June 2, 2025 at 3:44 PM
Zed for sure. There's still AI stuff to turn off, the debugger isn't in yet and it doesn't have the python notebook functionality that vscode has, but I like it.
May 29, 2025 at 1:42 AM
You could also take the "Metroidvania approach" where you introduce the students at a superficial level to the fully correct built out solution, then take it all away and build back up to that spot. That's roughly what I do in my secure web dev class.
May 20, 2025 at 5:39 PM
I agree in principle - to accomplish it we should make “Chicagoland” more prominent an identifier. They do this well in Northern California with “Bay Area.”
May 9, 2025 at 11:12 AM
Reposted by Chris Kanich
Even accepting the premise that AI produces useful writing (which no one should), using AI in education is like using a forklift at the gym. The weights do not actually need to be moved from place to place. That is not the work. The work is what happens within you.
April 15, 2025 at 2:56 AM
For novices, I often describe "writing more than 2-3 lines of code without testing it" as "Wile E. Coyote running off the side of a canyon." You think you're making progress, until you look down. LLMs crank this phenomenon up to 11. Hard to see a non-abstinence way to learn in this environment.
April 23, 2025 at 7:48 PM
I may need to take a Metra trip just to see the busway!
April 18, 2025 at 5:22 AM
nope, I think it's just a garden variety "that administrator made it a priority, so it happened" type situation.
April 17, 2025 at 4:13 PM
Overall cheating, not yet; we did draft a genAI policy for graduate theses however. I had to advocate pretty hard to prevent any prescriptive requirements, and focus on academic integrity as an agreement with the advisor/program on how the thesis is completed.
Thesis | Graduate College | University of Illinois Chicago University of Illinois ChicagoUniversity of Illinois ChicagoExpand Admissions menuExpand Fun...
grad.uic.edu
April 17, 2025 at 4:05 PM
Agreed - my question is, if the traditional path is to learn/practice a, then b, c, d, ... then z to build actual expertise, will the path in the presence of AI be to do everything the same, but don't get distracted by the answer-generation-machine, or something else?
April 16, 2025 at 9:05 PM
Reposted by Chris Kanich
The advice I'm hearing is that students on visas should be checking SEVIS daily. Ideally, students' colleges & universities could do that for them - if they have the capacity.
April 7, 2025 at 3:17 PM