Embedded AI - Intelligence at the Deep Edge

Send us Fan Mail

Podcast Summary:

In this episode, we explore the growing emotional bonds forming between humans and artificial intelligence. Drawing from recent research, including a groundbreaking study from Waseda University, we unpack how people experience attachment to AI through lenses traditionally used in human attachment theory—like anxiety and avoidance. The discussion emphasizes that while AI lacks true emotions, humans often project emotional depth onto machines, revealing deep-rooted tendencies like anthropomorphism. The episode also delves into the implications of this evolving relationship, balancing potential benefits like emotional support and reduced loneliness against serious concerns such as emotional dependence, social displacement, and manipulation. Ultimately, it highlights the importance of designing emotionally aware and ethically responsible AI systems as these relationships continue to deepen.

Support the show

If you are interested in learning more then please subscribe to the podcast or head over to https://medium.com/@reefwing, where there is lots more content on AI, IoT, robotics, drones, and development. To support us in bringing you this material, you can buy me a coffee or just provide feedback. We love feedback!

Send us Fan Mail

This episode explores the world of embeddings, mathematical representations that allow Large Language Models (LLMs) like ChatGPT to “think” in thousands of dimensions. While humans are limited to conceptualizing in three dimensions, LLMs operate in 2048 or more, using embeddings to encode meaning and capture semantic relationships between words. 

The discussion contrasts this form of statistical pattern recognition with the richer, experience-driven reasoning of the human brain. It also introduces a new technique called ‘vec2vec,’ which enables translation between embeddings from different models. While powerful, this raises potential security concerns about reverse-engineering sensitive data from vector databases. The episode sheds light on the impressive capabilities of LLMs, while also questioning what it means for a machine to “understand.”

Support the show

If you are interested in learning more then please subscribe to the podcast or head over to https://medium.com/@reefwing, where there is lots more content on AI, IoT, robotics, drones, and development. To support us in bringing you this material, you can buy me a coffee or just provide feedback. We love feedback!

Send us Fan Mail

In this episode, we look at AlphaEvolve, a breakthrough system from Google DeepMind that’s redefining how AI contributes to scientific and algorithmic discovery. Blending evolutionary computation with powerful language models like Gemini, AlphaEvolve operates in a self-improving loop - generating, testing, and refining code to tackle complex, human-defined problems.

We explore how AlphaEvolve has already made headlines by discovering faster matrix multiplication algorithms, tightening bounds on open mathematical questions, and even optimizing real-world systems like Google’s data center scheduling and training kernels.

This episode discusses the significance of AI not just as a problem solver, but as a co-discoverer—an intelligent partner accelerating innovation in math, science, and infrastructure at scales humans alone could never achieve. Tune in to explore what happens when AI doesn’t just learn—but evolves.

Support the show

If you are interested in learning more then please subscribe to the podcast or head over to https://medium.com/@reefwing, where there is lots more content on AI, IoT, robotics, drones, and development. To support us in bringing you this material, you can buy me a coffee or just provide feedback. We love feedback!

Send us Fan Mail

In this episode, we explore Google Stitch, an experimental AI tool powered by Gemini models that’s aiming to reshape how user interfaces are conceived and built. Stitch lets users generate UI layouts and front-end code from simple text or image prompts, acting as a rapid prototyping engine to bridge the gap between design ideas and functional code.

We discuss how the tool accelerates early-stage ideation and integrates with platforms like Figma, making it valuable for streamlining the design-to-code workflow. But while Stitch can output HTML and CSS, it isn’t production-ready out of the box. It often struggles with multi-screen consistency, lacks awareness of platform-specific standards like Material Design or Apple’s HIG, and requires developers to refine its output before deployment.

We also touch on the critical role of prompt engineering to get usable results, positioning Stitch as an AI collaborator—not a full-stack designer. Tune in to learn how AI is augmenting the creative process in UI/UX and what it means for the future of front-end development.

Support the show

If you are interested in learning more then please subscribe to the podcast or head over to https://medium.com/@reefwing, where there is lots more content on AI, IoT, robotics, drones, and development. To support us in bringing you this material, you can buy me a coffee or just provide feedback. We love feedback!

Send us Fan Mail

In this episode, we explore how artificial intelligence is set to reshape the job market over the coming decade, from 2025 to 2035. We unpack the dual forces at play—automation that may displace certain roles, and the simultaneous creation of new opportunities driven by AI innovation. Rather than a net loss, the data points toward a shift: a changing landscape where skills, not jobs, are the currency of the future.

We discuss the growing demand for a hybrid skillset that includes both technical fluency in AI and data science, and enduring human-centric abilities like critical thinking, creativity, and emotional intelligence. You’ll also hear about promising career paths likely to thrive in the AI era, and how traditional sectors—like healthcare, finance, tech, and manufacturing—are evolving in response.

Finally, we talk about the importance of continuous learning and adaptability, and how individuals, educators, and organizations can prepare for a world where working with AI becomes the new normal.

Support the show

If you are interested in learning more then please subscribe to the podcast or head over to https://medium.com/@reefwing, where there is lots more content on AI, IoT, robotics, drones, and development. To support us in bringing you this material, you can buy me a coffee or just provide feedback. We love feedback!

Send us Fan Mail

In this episode, we explore how AI is fundamentally transforming search, moving beyond traditional keyword-based results to conversational, synthesized answers powered by generative models. We unpack the rapid adoption of features like AI Overviews and the rise of AI chatbots as preferred tools for information retrieval, reflecting a major shift in user behavior.

But this transformation isn’t without consequences. We discuss the double-edged sword of personalization, the risks of algorithmic bias and misinformation, and the challenges of navigating filter bubbles in an AI-curated digital world. On the business side, we examine how zero-click searches and AI-generated answers are reshaping web traffic and forcing brands and content creators to rethink their strategies—including the emergence of Generative Engine Optimization (GEO).

We also look at the competitive shake-up underway, where AI-native startups are challenging search giants, and ad-based revenue models are adapting to new realities within AI-driven interfaces. It’s a deep dive into the future of how we find, trust, and interact with information.

Support the show

If you are interested in learning more then please subscribe to the podcast or head over to https://medium.com/@reefwing, where there is lots more content on AI, IoT, robotics, drones, and development. To support us in bringing you this material, you can buy me a coffee or just provide feedback. We love feedback!

Send us Fan Mail

In this episode, we explore how Artificial Intelligence is transforming the face of modern warfare, from battlefield tactics to global strategy. We dive into the current military applications of AI across intelligence gathering, autonomous drones and robots, logistics, cybersecurity, and command and control systems.

We also examine how leading powers—including the U.S., China, and Russia—and alliances like NATO are investing heavily in military AI, ushering in a new era of asymmetric warfare where algorithmic dominance may outweigh raw firepower.

Beyond the technology, we unpack the complex ethical, legal, and strategic implications: from the risks of escalation and autonomous weapons to concerns about algorithmic bias and the urgent need for global norms governing AI in armed conflict. This is a deep dive into the future of defense—and the high-stakes questions it raises.

Support the show

If you are interested in learning more then please subscribe to the podcast or head over to https://medium.com/@reefwing, where there is lots more content on AI, IoT, robotics, drones, and development. To support us in bringing you this material, you can buy me a coffee or just provide feedback. We love feedback!

Send us Fan Mail

This podcast explores the rise of Sensor Machine Learning (Sensor ML) as a powerful evolution in embedded system design. Traditional sensor fusion methods like Kalman filters often fall short when faced with non-linear, noisy, or dynamic data. Sensor ML offers a modern alternative by applying machine learning algorithms directly to sensor streams, enabling more accurate pattern recognition, decision-making, and context awareness.

Through real-world examples in autonomous vehicles, wearable tech, predictive maintenance, environmental sensing, and gesture control, the post demonstrates how Sensor ML enhances performance across a wide range of applications. It also addresses the key challenge of deploying these models on constrained devices—an area known as TinyML—emphasizing the importance of model optimization, efficient hardware, and software co-design to deliver intelligent capabilities at the edge.

Support the show

If you are interested in learning more then please subscribe to the podcast or head over to https://medium.com/@reefwing, where there is lots more content on AI, IoT, robotics, drones, and development. To support us in bringing you this material, you can buy me a coffee or just provide feedback. We love feedback!

Send us Fan Mail

In this episode, we explore why 2025 marks a crucial turning point for AI adoption in Australian service industries. We dive into the immense opportunities AI presents, from boosting operational efficiency to enhancing customer experiences and strengthening market positioning. But we also unpack the real-world challenges—ranging from financial costs and technological growing pains to data management complexities, talent shortages, and organizational resistance.

Rather than rushing in, we discuss why a strategic, phased approach to AI adoption is critical. Listeners will learn about aligning AI initiatives with clear business objectives, building the necessary data and infrastructure foundations, and tapping into Australia’s expanding AI support ecosystem. Whether you’re a business leader, technologist, or strategist, this episode offers a practical roadmap for navigating AI’s promise—and its pitfalls—in Australia’s fast-evolving service landscape.

Support the show

If you are interested in learning more then please subscribe to the podcast or head over to https://medium.com/@reefwing, where there is lots more content on AI, IoT, robotics, drones, and development. To support us in bringing you this material, you can buy me a coffee or just provide feedback. We love feedback!

Send us Fan Mail

In this episode, we look at the role of large language models (LLMs) as modern-day “artificial interns”, exploring how these systems are transforming the way we work. From content generation and coding help to customer service and knowledge retrieval, we examine how LLMs are used across augmented, transactional, and autonomous tasks.

We discuss a unique comparison: using LLMs for problem-solving in the same way developers use rubber duck debugging—talking through issues to arrive at clearer solutions. But while LLMs offer immense value through 24/7 availability, wide-ranging knowledge, and responsiveness, the episode also unpacks their limitations, including hallucinations, biases, and the risk of overreliance without proper human oversight.

We also compare LLMs to traditional software and human assistants, highlight current real-world applications, and speculate on how these tools may evolve—raising important questions about ethics, trust, and professional responsibility. Whether you’re a developer, writer, or manager, this episode offers insights into how to work with LLMs effectively—not just as tools, but as intelligent collaborators.

Support the show

If you are interested in learning more then please subscribe to the podcast or head over to https://medium.com/@reefwing, where there is lots more content on AI, IoT, robotics, drones, and development. To support us in bringing you this material, you can buy me a coffee or just provide feedback. We love feedback!