Prompting Over Programming

Updated: 2026.01.16 13D ago 15 sources
The post argues the entry‑level skill for software is shifting from traditional CS problem‑solving to directing AI with natural‑language prompts ('vibe‑coding'). As models absorb more implementation detail, many developer roles will revolve around specifying, auditing, and iterating AI outputs rather than writing code from scratch. — This reframes K–12/college curricula and workforce policy toward teaching AI orchestration and verification instead of early CS boilerplate.

Sources

Tyler Cowen's AI Campus
Arnold Kling 2026.01.16 88% relevant
Kling’s rapid creation of a meeting scheduler and a free‑form syllabus interface with Claude is a concrete example of the shift from hand‑coding to natural‑language 'prompt engineering' (vibe‑coding) as the primary productivity skill for many software artifacts.
How “new work” will actually take shape in the age of AI
Zack Kass 2026.01.13 78% relevant
The author emphasizes natural‑language orchestration and tool‑use (asking AI to draft, summarize, triage), which aligns with the 'Prompting Over Programming' idea that entry‑level software and many roles will center on directing models rather than writing low‑level implementations.
Even Linus Torvalds Is Vibe Coding Now
BeauHD 2026.01.13 85% relevant
The article reports Torvalds saying he ‘cut out the middle‑man — me — and just used Google Antigravity to do the audio sample visualiser,’ a concrete instance of the broader claim that entry‑level software work is shifting from hand coding to prompt‑directed orchestration.
AI Links, 1/11/2026
Arnold Kling 2026.01.11 95% relevant
The article centers on 'vibe‑coding' and non‑programmer workflows (Kelsey Piper, Ethan Mollick) showing a shift from hand‑coding to prompt/agent orchestration; that is precisely the existing idea that prompting becomes the primary entry skill and framing of software work.
C# (and C) Grew in Popularity in 2025, Says TIOBE
EditorDavid 2026.01.11 48% relevant
TIOBE’s year‑over‑year shifts (C# largest gain; Rust rising but still modest; TypeScript predicted to break into top 20) suggest the landscape remains language‑centred rather than purely prompt‑centric — a counterpoint to the 'prompting over programming' thesis and a data point to monitor whether language popularity erodes as prompting tools spread.
Friday assorted links
Tyler Cowen 2026.01.09 75% relevant
The coding‑agent item and Cowen’s curation illustrate the broader shift described by 'Prompting Over Programming' — that entry‑level software work is moving toward orchestration of models and prompts rather than traditional hand‑coding.
'The Downside To Using AI for All Those Boring Tasks at Work'
msmash 2026.01.08 72% relevant
The piece illustrates the shift from routinized, low‑attention work to higher‑level, prompt‑driven cognitive tasks once AI takes over mundane tasks — a concrete workplace example of the broader educational and occupational reorientation this idea predicts.
Podcast: When efficiency makes life worse
2026.01.08 50% relevant
Discussion of chess, baseball analytics and computer assistance ties into the existing claim that tooling and prompting change skill sets and encourage efficiency‑first approaches; the podcast raises the normative question about what is lost when tasks become prompt‑driven rather than craftsmanship‑driven.
Actually-existing UATX
Arnold Kling 2026.01.08 80% relevant
Kling’s classroom practice — giving students 'vibe‑coding' assignments and using Claude to iterate student papers — exemplifies the shift toward prompting/AI orchestration as a core student skill rather than traditional hand‑coding, directly reflecting the 'prompting over programming' premise.
AI and Economics Links
Arnold Kling 2026.01.07 72% relevant
Gans’ experiment and the ChatGPT 5.2 test that produced a paper in under 20 minutes are direct examples of the shift from hand‑coding and manual drafting toward 'prompt‑driven' research workflows and entry‑level scholarship, matching the existing idea that prompting is becoming the core entry skill.
I can't stop yelling at Claude Code
Kelsey Piper 2026.01.07 92% relevant
The author describes exactly the transition that 'Prompting Over Programming' predicts: agentic tools (Anthropic’s Claude Code/Opus 4.5) convert much of coding into specifying desired behavior (prompts) while removing the plumbing errors that formerly dominated the workflow, leaving a new skillset centered on prompt design and orchestration.
AI Links, 12/31/2025
Arnold Kling 2025.12.31 95% relevant
Steve Yegge’s insistence that engineers must stop using IDEs and learn how 'agents code' and Karpathy’s claim that programming is being refactored directly exemplify the 'prompting over programming' thesis — a shift from hand‑coding fundamentals to orchestrating LLM agents and prompts as the core developer skill.
When to Graduate from College?
Arnold Kling 2025.12.28 90% relevant
Kling’s core claim—that AI could devalue credentials and shift value toward skillful use of AI tools rather than traditional coursework—directly echoes the 'Prompting Over Programming' idea that entry‑level software work will pivot from hand‑coding to directing models; he uses that shift to argue graduation and curricular timing should change.
3 experts explain your brain’s creativity formula
David Eagleman, Scott Barry Kaufman, Tiago Forte 2025.12.03 82% relevant
The article argues creativity depends on externalizing and recombining knowledge (Tiago Forte’s 'second brain'), which parallels the existing idea that entry‑level software skills shift from low‑level coding to orchestrating tools and prompts; both forecast a redefinition of baseline competencies toward tool orchestration and knowledge‑management.
Some AI Links
Arnold Kling 2025.10.04 100% relevant
Alexandr Wang’s '13‑year‑olds should vibe‑code' and Jensen Huang’s 'natural language is the new programming language,' plus Kling’s claim CS will add less value as AI coding improves.
← Back to All Ideas