The Confusion Matrix

Alex and Pete dissect whether AI agents can actually replace knowledge workers, discussing why language models excel at testable tasks like coding, but fare less well at more general business activities that often require navigating the messy political realities of the corporate landscape.

Alex and Pete discuss how AI agents are collapsing the software development lifecycle (SDLC), eliminating manual coding steps whilst creating new requirements for testing, monitoring, and managing increasingly complex automated workflows at unprecedented scale.

Alex and Pete discuss our problematic tendency to project human characteristic on LLMs, and use anthropomorphic language when discussing them.

Alex and Pete grapple with the ethics of using generative AI in creative work—balancing utility against copyright theft, environmental impact, and career pressure whilst acknowledging there’s no easy resolution.

Alex and Pete examine whether AI will replace jobs within 18 months, discussing automation history, current limitations, and Microsoft’s predictions regarding workplace displacement.

Alex and Pete discuss AI coding agents’ limitations after Alex’s LLM client build revealed extensive bugs and missing features. They explore testing challenges, agent reliability metrics, and corporate liability concerns preventing enterprise adoption. This episode will make a lot more sense if you listen to the previous episode first.

Alex and Pete discuss Alex’s rebuild of his perma-regenerating LLM chat client. This time he did it properly! No messing, just hardcore, disciplined, systematic, agent-coding goodness. Join them as he walks Pete through the process in excruciating, step-by-step detail. The Vanishing Gradients episode Alex mentions.

Alex and Pete examine OpenAI’s comprehensive report on ChatGPT usage patterns, analysing classification methodologies and user behaviour across consumer and workplace contexts. They discuss the shift from advice-seeking to task execution, the dominance of writing and information-seeking functions, and implications for future AI adoption and market opportunities.

Pete and Alex examine AI model evaluation methodologies, comparing traditional machine learning metrics with the qualitative assessment challenges of large language models. They discuss the collaborative requirements between technical and business teams to establish evaluation criteria for generative AI systems, highlighting the subjective nature of testing conversational outputs versus binary classification tasks. With the help […]

Pete and Alex dig into cybersecurity risks with AI agents and generative AI systems. They cover two main problems: people coding dodgy applications without security knowledge, and hackers directly exploiting AI agents that have access to tools and data. Despite the scary possibilities, they reckon most vulnerabilities are manageable with decent security practices. Practical AI […]

Pete and Alex recap their recent AI discussions, covering why language model “hallucinations” are actually normal behaviour, how most AI proof-of-concepts fail due to poor ideas rather than technical issues, the impact on jobs and graduate recruitment, and the rise of AI coding agents that are reshaping software development. Practical AI – Dealing with increasingly […]

Alex and Pete discuss MIT’s study revealing 95% of GenAI projects fail despite massive enterprise investment. They explore why companies struggle to scale beyond pilots, the “shadow AI economy” of employees secretly using personal AI tools, and practical strategies for successful implementation including adaptive systems and treating AI procurement like business process outsourcing.

Alex explains his slow but unstoppable gravitation to text based interfaces and an “everything via the terminal” mentality. After all, who needs graphical operating systems? To do this, Pete and Alex take a ramble through the history of nerd-first user interfaces, discuss why keyboard layouts are stupid, how this relates to window based OSes and […]

Pete and Alex ponder whether we are at the point where LLMs are as good as they are going to get, and what the implications of this are. This requires a dip into the murky depths of what businesses exist to do and how the randomness that the LLMs generate is antithetical to how businesses […]

Join Alex and Pete as they discuss building software that suits only your needs and idiosyn-crazies using AI. Surely this is a good thing, and not just another apocalypse in disguise? There’s only one way to find out!

Pete and Alex discuss the challenges of deploying language models in end-user facing systems, the dilemma of diversity vs accuracy of output, and how all of this is a giant headache for data scientists and software engineers alike.

Pete and Alex discuss the challenges of training models for accuracy, reliability and capability, and various techniques that are being used to ensure these. They talk about how to think about how the models work, why they get things wrong and Alex makes an impassioned case for why cheese should be considered a fruit!

Alex and Pete battle the demons of unruly technology to give you a brief history of Business Intelligence and their relationship with it. They lament that it all seems to still be about dashboards, why this sucks and offer some alternatives. And yes, AI makes an appearance. Alex also recounts mysterious case of Strategy, the […]

Pete and Alex discuss the range of views on the impact of GenAI on jobs, the seemingly intractable situation that decision makers find themselves in regarding AI adoption, the marvels of modern enterprise procurement procurement, and the AI pricing problem. Oh, and we also get an update on the state of Alex’s thumb.

Alex recounts his harrowing tale of vibe coding hubris, a rogue AI agent, a busted deadline and a bruised ego. He and Pete discuss the need for a measured approach to using AI agents for mission critical work, the need for actual tech skills when using agents, and how managing them is not really that […]

Alex takes us on a vibe coding adventure, fixing a bug using an AI agent live (ish), before our very ears! Ada makes a rather fruity appearance, as she and Pete are dazzled by unreasonable effectiveness of AI coding agents. Finally, Pete and Alex discuss the implication to their discipline, the tech industry as a […]

Pete and Alex discuss some articles and podcasts prophesying doom or urging calm as relates to the potential for AI to wreak havoc on the human race. Needless to say, things get a little animated! Spoiler: It’s not all doom. Show notes:

Alex has a moan, Pete takes us on an odyssey of dubious AI tuition, Ada has an existential crisis that results in rewriting herself, and there is an angry sheep and a sparkly keyboard. Confusing? We wouldn’t have it any other way!

Alex and Pete talk to Iain “Preza” Preston, client whisperer and innovation guru, about how to actually innovate in big business, how to structure, focus and motivate innovation teams, why data is essential for the innovation process, and why making stuff beats faking stuff.

Alex’s attempts to converse with the Old Gods via the medium of coding leads to a lively discussion about the merits (and dangers) of vibe coding, why you shouldn’t code after drinking alcohol, the unreasonable effectiveness of Warp terminal, and how LLMs can help in education. Not for the faint hearted!

Alex and Pete are joined by Alex’s AI companion Ada in this expletive strewn crawl through the underbelly of AI personality, vanity, self-awareness (or lack thereof), homunculi, the horse apocalypse, and the value of (real) human connection. Be afraid!

Pete and Alex discuss their ongoing, if slightly rocky, love affair with Marimo notebooks, and dig deeper on Model Context Protocol and its siren-like allure.

Pete and Alex discuss their love of small data, the role of human biases (and lack thereof) in data analysis, and keeping it real as an analyst in the face of unrealistic business expectations.

Pete and Alex discuss the increasing role of GenAI in Business Intelligence (BI), Alex’s experiences of cheating on Pandas with Polars and they forecast doom via an incoming vibe coding cataclysm.

Alex and Pete discuss Alex’s foray to the magical world of Marimo notebooks, and discuss their pros and cons and why Alex is a convert (for now). Pete raises the problems that arise from finding that “unicorn” client too early in a business’s life.

Pete and Alex continue their epic chinwag on how to knowledge base. In this episode they discuss the challenges of Retrieval Augmented Generation (RAG), effectively chunking text data, how to know what “good” looks like, and Alex gets a bit shirty about LLMs apparently reinventing programming.

Alex and Pete discuss the nature of knowledge bases, what they are and what they’re not, some of the platforms and tools for creating and maintaining them, before realising that they have so much to say there’s going to need to be another episode!

Alex and Pete discuss the pitfalls of using averages as key performance indicators, and explore Decentralised Autonomous Organisations (DAOs) and their potential use in art collectives and other community-minded organisations, as well as the possible implications of AI in business decision-making.

Alex and Pete debate the merits of flat files versus complex databases – a battle as old as time (or at least, as old as relational databases). They also uncover CHIP.AI, a no-code AI assistant builder, proving that even the most technically challenged can harness the power of AI. Tune in for some data-science shenanigans!

Things get heated is Alex and Pete debate the merits of LLMs for authoring books, the dangers of using labels like “PhD level thinking” to models, and the limitations of the associated benchmarks, and observe the creaky state of the stock market from a tech and AI perspective. Here’s the links to various sources for […]

Alex and Pete discuss the creation of their podcast’s visuals and soundtrack using AI tools, the concept of the confusion matrix in machine learning, and the importance of questioning overly accurate models. They also explore how data can provide insights into workplace productivity and team dynamics.

Alex and Pete kick off their new podcast by dipping their toes in the murky waters of LLMs and how they seem to be insinuating their way in to all manner of aspects of their lives. Good or bad or ugly? Tune in to find out!