TBT Ep56 DEMYSTIFYING AI: No Hype: Simplifying the Buzzwords
Cutting Through the AI Hype – No Jargon, Just Clarity
AI buzzwords are everywhere—LLM, ML, DL, RAG, ChatGPT, Custom GPT, NLP—but do we actually understand what they mean?
In this episode of TechBurst Talks, we strip away the complexity and make AI accessible, practical, and bullshit-free.
I’m joined by Bernard Leong, an AI expert who thrives on theory, while I take a hands-on, real-world approach. Together, we:
✅ Decode AI jargon without the fluff
✅ Explain why these buzzwords actually matter
✅ Break down AI trends in a way that anyone can grasp
If you’re drowning in AI terminology—whether you’re an industry pro or just trying to keep up—this episode is your lifeline.
🎧 Tune in and let’s bring clarity to the chaos.
#AI #NoBS #TechBurstTalks #demystifyingai
Show Notes
Guest Introduction & Background (00:02 - 02:05)
- The guest, Bernard Leong, has a PhD in theoretical physics from Cambridge and has worked in AI-related fields for years.
- His experience spans machine learning in genomics, corporate AI projects, and roles at Airbus, Amazon Web Services (AWS) , and now his AI startup, Dorje.AI.
Defining Artificial Intelligence (03:13 - 04:20)
- AI simulates human intelligence, including learning, reasoning, and self-correction.
- Machine learning (ML) enables AI to predict outcomes, like distinguishing a cat from a non-cat or recognizing traffic lights.
Big Data vs. AI (04:20 - 05:26)
- Big Data involves collecting, processing, and organizing vast amounts of data.
- AI utilizes structured Big Data to generate insights and predictive outcomes.
- Only 10% of companies are truly ready for AI due to data silos.
Machine Learning vs. Deep Learning (07:53 - 09:49)
- Machine learning allows computers to learn patterns and improve over time.
- Deep learning, a subset of ML, uses neural networks to process vast amounts of data quickly (e.g., real-time traffic light recognition).
Generative AI & Large Language Models (09:49 - 13:21)
- Generative AI (Gen AI) creates content (text, images, audio, video) using large language models (LLMs).
- The breakthrough came from Google’s 2017 paper, Attention Is All You Need.
- Tools like ChatGPT have democratised AI, making it accessible beyond academia and tech companies.
Real-World AI Applications (15:28 - 20:49)
- AI tools like ChatGPT, Microsoft Copilot, Notion, and Canva improve productivity.
- AI-driven video and audio editing (e.g., Descript) automate tasks like transcription and filler-word removal.
- AI-powered legal assistants and enterprise AI tools are reshaping industries.
Custom GPTs & Retrieval-Augmented Generation (RAG) (24:07 - 29:40)
- Custom GPTs let businesses build AI assistants tailored to their knowledge base (FAQs, policies, legal docs).
- RAG enhances AI by pulling accurate, real-time data from structured sources, reducing errors (e.g., Air Canada chatbot fiasco).
Risks & Ethical Concerns in AI (35:09 - 38:50)
- AI is often a "black box," making it hard to interpret its decision-making.
- Federated learning allows AI models to be trained on user data while maintaining privacy.
- Ethical concerns include misinformation, bias, and over-reliance on AI-driven decision-making.
AI’s Impact on Jobs & Future Trends (42:36 - 53:52)
- AI is shifting work from repetitive tasks to creative problem-solving.
- Software development is becoming commoditised due to AI-powered coding.
- Companies are exploring smaller AI models that run on personal devices rather than cloud-based models.
Enterprise AI & ERP Transformation (51:40 - 57:27)
- AI can streamline Enterprise Resource Planning (ERP) by breaking down data silos and automating workflows.
- AI’s role in mission-critical systems (e.g., Airbus’ autonomous landing systems) highlights the need for human oversight.
- The discussion wraps up with reflections on making AI more accessible to general audiences.