AI coding tools can generate huge amounts of functional code in seconds, but does that mean developers no longer need to learn how to code? In this episode, Matt and Mike discuss how aspiring and junior developers can use AI coding agents as both builders and tutors - following their curiosity, learning the concepts behind generated code, and deciding how much manual coding knowledge they actually need to retain. They also explore the growing challenge of balancing traditional programming skills with agentic coding skills, especially when preparing for developer jobs and technical interviews.

Show Notes: https://www.htmlallthethings.com/podcast/dont-just-vibe-code-learn-while-you-code

Use our Scrimba affiliate link (https://scrimba.com/?via=htmlallthethings) for a 20% discount!! Full details in show notes.

GPT-6 Astra is here, bringing another huge jump in AI capabilities - and once again raising the question: are we approaching AGI?

But defining artificial general intelligence gets messy fast. If AGI means matching or exceeding humans across virtually any cognitive task, how do you actually test that? And as new benchmarks appear, are we discovering better ways to measure intelligence or simply moving the goalposts? In this edition of the Web News, Matt and Mike discuss what AGI means from a regular user's perspective, why benchmark scores don't necessarily tell the whole story, and whether the label even matters if AI can already handle increasingly complex real-world work. For most of us, the question may be much simpler: does the AI actually do what we need it to do?

‍Show Notes: https://www.htmlallthethings.com/podcast/does-agi-actually-matter-to-you

Most of us repeatedly teach our AI coding agents the same lessons: ask questions before coding, write a specification, break the work into tickets, test everything, and review the result. Agent skills let us package those instructions into reusable workflows that an agent can load whenever they become relevant. In this episode, Matt and Mike explain what agent skills are, how they differ from prompts, project rules, tools, MCP servers, and custom GPTs, and why they can make AI-assisted development more consistent. They also explore Matt Pocock’s engineering skills - including grill-with-docs, wayfinder, to-spec, to-tickets, implement, code-review, and handoff - and discuss why installing too many overlapping skills can make an agent worse instead of better.

Show Notes: https://www.htmlallthethings.com/podcast/ai-agent-skills-explained-what-they-are-and-which-ones-to-use

Use our Scrimba affiliate link (https://scrimba.com/?via=htmlallthethings) for a 20% discount!! Full details in show notes.