PERL: The Smart AI Workflow

PERL: The Smart AI Workflow

Still throwing one prompt at an AI agent and hoping for the best? In this episode of SEEK Bytes – we unpack PERL: Plan, Execute, Review and Learn — a smarter way to work with AI that helps teams get better outputs, catch mistakes earlier and turn trial-and-error into a repeatable workflow. This is the episode for anyone curious about how AI tools are really being used inside modern tech teams — not just to generate answers, but to plan better, work faster and learn as a team. In this episode you’ll learn: • Why one-shot prompting is holding people back – and how separating planning from execution can lead to better results, fewer misses and more confidence in what your AI tools are doing. • How “review and learn” takes AI from personal hack to team advantage – by spotting what went wrong, improving the next run and feeding those lessons back into shared skills, rules and documentation. • The hidden traps of advanced AI workflows – from runaway token usage and looping behaviour to context compaction and hallucinations, plus why mastering these tools is becoming a real skill in tech. 🔔 Follow the SEEK Bytes podcast so you never miss a new episode

Vibe Coding with AI: Non-Developers building apps (with Helen Giapitzakis)

What happens when a designer with almost no coding experience picks up an AI tool and starts building a working product prototype – and then a staff engineer has to make sense of it all? In this episode of SEEK Bytes, Elliott Millar, Seamus, Arnav and Staff Engineer Helen Giapitzakis dive into “vibe coding”: using AI agents, context and prompts to build software by feel instead of by hand. This episode's special guest: Helen Giapitzakis (SEEK Staff Engineer) In this episode, we explore: • How a SEEK designer used AI to go from idea to clickable prototype, put it in front of real users, and only pulled in engineers when things started to wobble • Why “vibe coding” changes who can build software – and what that means for engineers, from assisted coding and context engineering to owning standards, readability and long-lived code • The surprising ways LLMs misunderstand intent and rules (especially negatives like “never use any”), and how teams can use positive, shared guidelines and repo-level “AI rulebooks” to steer them Whether you’re in software engineering, UX, product, data, mobile, platform or IT leadership, this episode will get you thinking differently about who codes, how AI fits into real teams, and what “good” code even means in an AI-first world. 👍 Follow the SEEK Bytes podcast so you never miss a new episode

New AI Tools, Same Old Bugs? The crew react to the Next Wave

Drowning in new AI tools and wondering which ones actually help you ship better software? In this episode of SEEK Bytes, Elliott, Seamus and Will react to the latest wave of AI‑powered dev tools, from coding assistants and test generators to agents that promise to refactor your whole codebase while you sleep. They dig into what these tools are really good at (speeding up boilerplate, exploring APIs, generating tests), where they fall over (context, architecture, edge‑cases), and how they’re already changing code review, onboarding and day‑to‑day developer experience at SEEK. In this episode, we explore: • Which new AI tools are genuinely useful for everyday coding – and which ones are more hype than help • How AI is reshaping pairing, code review, debugging and tests, and where human judgment still absolutely has to stay in the loop • Practical ways to experiment with AI tools on your team without burning trust, wrecking your codebase or slowing everyone down If you’re a software engineer, tech lead or engineering manager trying to make sense of the exploding AI tool ecosystem, this episode is for you. 🔔 Follow the SEEK Bytes podcast so you never miss a new episode

OpenClaw: Build your own AI Agent workforce (without getting burnt)

What if you could spin up a whole AI workforce on your own machine – planners, “engineers”, researchers and content creators – and run it all from a simple chat window? In this episode of SEEK Bytes, we explore OpenClaw, the agent platform that lets you act like the CEO while AI agents do the grunt work in the background. In this episode you’ll learn: • How OpenClaw turns AI into a digital team – using one “front-of-house” bot as your main interface, then spinning up specialised agents behind the scenes to research, draft content, update tickets and even prepare changes for real systems. • What it really costs to run this at home – from beefy GPUs and Mac minis to cloud API bills that can quietly blow past a few hundred dollars a month, plus why some “cheap and easy” YouTube setups gloss over the real price tag. • The security & scam risks you need to know about – including AI “app” stores where around 1 in 5 skills are malicious, how the wrong plugin can leak your data or crypto, and why doing your homework before you plug AI into your systems is non-negotiable. Drop a comment if you’ve tried building your own AI stack (or are tempted to), and hit Subscribe for more real-world AI, security and engineering stories from the SEEK Bytes team. 🔔 Follow SEEK Bytes podcast so you never miss a new episode

NX Supply Chain Attack Explained (with Trevor Kilvington)

Imagine updating a trusted monorepo tool… and the next time you open your terminal it asks for your password, then tries to shut your machine down. In this episode of SEEK Bytes, Will, Seamus and SEEK Staff Engineer Trevor Kilvington unpack the NX supply chain attacks – how they unfolded, why they were so scary, and what every IT team can learn from them. This episode's special guest: Trevor Kilvington (SEEK Staff Engineer) In this episode, we explore: • How a popular open-source tool ended up shipping malicious code to developers’ laptops and CI pipelines – and how SEEK engineers helped spot it first • Why attackers leaned on AI prompts and worm-like behaviour to hunt for secrets and quietly spread through NPM packages • The uncomfortable question of whether keeping everything “always up to date” is still the safest choice – and what this means for CI/CD ownership and developer responsibilities Whether you’re in software engineering, DevOps, SRE, security, platform or IT leadership, this episode will change how you think about package updates, build pipelines and the tools you trust every day. 👍 Follow the SEEK Bytes podcast so you never miss a new episode