Charlie and Oz reconnect after a year to tackle the question of our times: now that coding agents actually work, does learning computer science still matter — and is the agent an asset to your growth or a liability? They get into RLVR and why agents crossed the threshold, Amdahl's law as a metaphor for automated productivity, why Oz still wants everyone to write a little assembly, and the creeping fear that agents sap "the will to learn" the same way social media saps attention.

Then the constructive turn: using agents to build your own challenges within your zone of proximal development, why you should read the back of the textbook chapter first, and Oz's homework for Charlie — figure out what you earnestly want to learn, not what you feel you should.

Shownotes:

• Steve Krouse of Val Town's post on why coding is still worth learning
• Amdahl's law
• Vygotsky's zone of proximal development
• Bloom's 2 sigma problem
• Oz's blog
• Charlie's blog

Brian "Beej Jorgensen" Hall joins Oz and Charlie to dive into the joy and art of learning computer science. It's all here - achieving flow, "How to Solve It" by Polya, self-publishing guides/books on the web, Beej's take on AI and coding, and, of course, the origin of the "Beej Jorgensen" moniker.

Shownotes:

• Beej's website: https://beej.us/
• Beej's guide to learning computer science: https://beej.us/guide/bglcs/
• Hacker News post on Charlie's short story about a giant diamond asteroid

Oz and Charlie vibe-chat their way through respective Mathacademy experiences + reviews, leading to a discussion of "binge learning" vs more-structured, consistent learning / time-tracking. Also, Oz shares that he's currently reading a math textbook with this incredible sentence in its introduction: "I consider the mathematical treatment of these problems to be among the chief glories of Western civilization, and I hope you agree."

Shownotes:

• mathacademy.com
• csprimer.com