Holy moly, one of my cousins (8 weeks into undergraduate computer science) asked me for help on a university assignment. I was surprised because he's very switched on.
It's a group assignment and it was extremely obvious that every other student has logged into a parent's corporate LLM and written all the code via prompt with no understanding.
When I asked how they're getting away with it, he said that the lecturers have just given up on all policing. (University is RMIT in Melbourne.)
It was extremely obvious from some questions that they're learning absolutely nothing, the faculty have totally phoned it in (they've actually had NO teaching staff for two weeks -- the lecturer quit mid-semester with no replacement), and it's all just a fucking disaster.
On the bright side, I guess my work is safe indefinitely.
Are there any recent publicly available curricula for people who value the lessons l & skills of compsci and don't give a shit about college degree prestige or a chasing current popular programming language meta? Or outdated ones, with a load of pre LLM popularity textbooks and exercises?
A "learn and know the fundamentals of this before going near this" flowchart would help, without it everything feels as demotivating as a guessing game without even any "hotter!/colder!" feedback.
@barcode @ludicity Looking for "recent" might be part of the problem. There are many old computer science, software management, and software management texts that are relevant regardless of recent popular programming language or process. e.g., "Algorithms + Data Structures = Programs" by Niklaus Wirth 1975, "Peopleware" by DeMarco and Lister 1987, "Applied Software Measurement" by Capers Jones 2008, Sally Shlaer and Stephen Mellors books from 1998 and 1991, etc.
@barcode @ludicity FWIW, I 5think software has been backsliding in terms of applying known good engineering principles for a long time. The popularity contest culture combined with the favoring of the terms, "art" and "craft" over "science" and "engineering" helped dumb down the culture before LLMs (AI) and (as a positive) has dumbed down the learning potential of the LLMs.