Analyse Podcast

Fresh out of the studio, Benedict Evans, independent technology analyst and author of AI Eats the World, returns to explore whether the AI model layer is becoming commodity infrastructure. Benedict argues there is no winner-takes-all effect in models yet, drawing parallels to telecoms, cloud, chips and the fiber bubble to ask where durable value actually accrues when everyone runs similar infrastructure on similar tokens. He unpacks why the chatbot remains a poor interface, introduces the "blank screen" and "jagged frontier" problems that keep software companies alive, and explains why large language models inherently give you "the average." Closing the conversation, Benedict reflects on the indicators that would show AI has truly eaten the world — and why the answer is better products, not better models."When you automate away work, you can always see the jobs that are going away because they're right there. And you don't know what the new jobs are going to be. Human needs are infinite. How many people are earning a living from making podcasts now? Imagine predicting that 10 years ago. There's a stage in the evolution of the market where like if you're still arguing about that, you're an idiot. But there's a stage at the beginning where you might have opinions about some of these questions, you're probably not even asking the right questions. That, I think, is where we are with this stuff today." — Benedict EvansEpisode Highlights: [00:00] Quote of the Day by Benedict Evans from AI Eats the World[01:16] The public market test: what are investors buying?[04:21] How far up the stack can models go?[05:30] Models can't build all the apps themselves[06:00] The thesis: models as commodity infrastructure[07:52] "All the value went up the stack"[08:24] Chips and Rock's Law: down to three players[11:23] The 1999 reseller story: one-time sales[13:28] The S-curve framing of technology[16:38] You're probably not asking the right questions on AI[18:02] "If this works, we're competing with a Mac"[20:25] Incumbents make it a feature[22:14] Big tech "killing startups" is overstated[24:39] Cowork as the new spreadsheet[26:01] The blank-screen and jagged-frontier problems[29:00] The hard part isn't writing the code[31:25] "What a good answer would probably look like"[33:38] The job displacement debate[37:38] Jevons paradox and the lump-of-labour fallacy[40:30] LLMs inherently give you the average[42:36] Why you really hire McKinsey[45:33] Punk versus prog rock: outside the training data[49:00] Automating ever-higher human functions[49:55] Why this is unanswerable: no theory of scaling[51:30] Indicators that AI has eaten the world[54:53] The solution isn't a better model[56:39] Where to find Benedict EvansProfile: Benedict Evans, Independent Technology AnalystLinkedIn: https://www.linkedin.com/in/benedictevans/Website: https://www.ben-evans.com/newsletterPodcast Information: Bernard Leong hosts and produces the show. The proper credits for the intro and end music are "Energetic Sports Drive." G. Thomas Craig mixed and edited the episode in both video and audio format.Here are the links to watch or listen to our podcast.Analyse Podcast Main Site: https://analysepodcast.comAnalyse Podcast Spotify: https://open.spotify.com/show/1kkRwzRZa4JCICr2vm0vGl Analyse Podcast Apple Podcasts: https://podcasts.apple.com/us/podcast/analyse-asia-with-bernard-leong/id914868245 Analyse Podcast LinkedIn: https://www.linkedin.com/company/analyse-podcast/Sign Up for Our This Week in Asia Newsletter: https://www.analysepodcast.com/#/portal/signup Subscribe Newsletter on LinkedIn https://www.linkedin.com/build-relation/newsletter-follow?entityUrn=7149559878934540288

Fresh out of the studio, Karen Hao, investigative journalist and author of "Empire of AI" joined us in a conversation to unravel how companies like OpenAI, Anthropic, and xAI have become modern empires reshaping society, labor, and democracy itself. Karen traces her journey from mechanical engineering at MIT to becoming one of the tech industry's most critical voices, sharing how Silicon Valley's innovation ecosystem has distorted toward self-interest rather than the public good. She unpacks the four characteristics that make AI companies mirror colonial empires: resource extraction through data scraping, labor exploitation of annotation workers, knowledge monopolies where most AI researchers are industry-funded, and quasi-religious quests to build an "AI God." Throughout the conversation, Karen reveals OpenAI's governance dysfunction stemming from its contradictory non-profit-for-profit structure and shares the inspiring story of Chilean water activists who successfully blocked Google's data center from draining their community's freshwater resources. She explains how Sam Altman's plans for 250 gigawatts of data center capacity—equivalent to four dozen New York Cities—would be environmentally catastrophic, while demonstrating how China's export restrictions paradoxically spurred more efficient AI innovation. Last but not least, she argues that empathy-driven journalism remains irreplaceable and calls for global citizens to hold these companies accountable to the broader public interest."These empires are amassing extraordinary amounts of resources by dispossessing a majority of the world. That includes like the data that they're extracting from people by just scraping it from online or intellectual property that they're taking from artists and creators. Most AI researchers now work for the AI industry and/or are funded in part by the AI industry. Even academics that have stayed within universities are often funded by the AI industry, and the effect that that has had on knowledge production is akin to the effect we would imagine if most climate scientists were bankrolled by the fossil fuel industry. I cannot stress enough how much they genuinely believe that they are on the path to creating something akin to an AI god, and that this is going to have cataclysmic shifts on civilization." - Karen Hao, Author of Empire of AIEpisode Highlights:[00:00] Quote of the Day by Karen Hao[00:47] Introduction: Karen Hao, Author of "Empire of AI"[01:44] From MIT engineering to investigating AI journalism[02:51] Silicon Valley distorts innovation toward self-benefit[04:12] AI companies as modern empires of power[06:00] Four traits of Empire: extraction, exploitation, monopolies, ideology[09:01] Quasi-religious movements driving Silicon Valley AI development[10:04] AGI believers speak specialized fanatical vocabulary[11:16] OpenAI founding: nonprofit facade, profit ambitions[13:53] Sam Altman firing: board's failed governance attempt[17:13] Fragmentation: every billionaire building their own AI[19:06] China's export controls sparked efficient AI innovation[21:57] Silicon Valley lacks American democratic values entirely[25:06] Chilean activists successfully blocked Google's water extraction[28:51] Sam Altman's 250 gigawatts: four dozen New York cities[31:21] Scaling continues despite base model asymptote reached[32:53] Benchmarks faulty: training data unknown, results unreliable[39:11] Success: sparking conversation about AI's human costs[39:40] ClosingProfile: Karen Hao, Author of Empire of AI and Investigative Journalist LinkedIn: https://www.linkedin.com/in/karendhao/Personal Site: https://karendhao.com/Podcast Information: Bernard Leong hosts and produces the show. The proper credits for the intro and end music are "Energetic Sports Drive." G. Thomas Craig mixed and edited the episode in both video and audio format.

Fresh out of Oracle AI World 2025, Chris Chelliah, Senior Vice President of Technology and Customer Strategy for Japan and Asia Pacific at Oracle, joins us to unpack how Oracle is positioning itself as the definitive enterprise AI platform across the region. He shares his career journey from a computer science geek working on distributed databases to leading technology strategy across a market representing two-thirds of the world's population. Chris explains Oracle's comprehensive four-tier AI stack—infrastructure, data platform, applications, and agentic orchestration—emphasizing how this unique full-stack ownership enables enterprises to consume AI out of the box and extend seamlessly without ripping and replacing existing systems. He highlights compelling use cases from financial fraud detection and healthcare automation to precision agriculture and energy grid optimization. Closing the conversation, Chris shares his vision for what great Oracle will look like in Asia Pacific, continuing its 50-year legacy as the behind-the-scenes platform provider powering everything from OpenAI and TikTok to global banking infrastructure. What's been consistent for Oracle is to be a platform provider that helps organizations unlock full value of their data. Today it is all about AI and unlocking the value of your data in AI, and cloud is a mandatory enabler. With AI and agentic AI, an agent is effectively an employee—it's an automated employee, a process, a workflow. You want your employees to be within your ecosystem, within your firewall. AI thrives at the edge because that's where inference happens. With AI and agentic AI, an agent is effectively an employee—it's an automated employee, a process, a workflow. You want your employees to be within your ecosystem, within your firewall. AI thrives at the edge because that's where inference happens. - Chris ChelliahProfile: Chris Chelliah, Senior Vice President of Technology and Customer Strategy for Japan and Asia Pacific at Oracle https://www.linkedin.com/in/chrischelliah/Episode Highlights: [00:00] Quote of the day by Chris Chelliah[02:10] Chris's journey from computer science to enterprise tech[03:13] Technology tinkering and Oracle's innovation culture explained[04:17] Two-thirds world population drives APJ market potential[05:06] Career advice: Find passion, own your brand[06:54] Oracle's mission: Unlocking data value for enterprises[07:58] 47,000 customers, 44% yearly consumption growth in JPAC[08:51] Oracle AI World 2025: AI changes everything announcement[09:12] Four-tier stack: Infrastructure, data, applications, agents[11:25] AI Data Platform enables production-grade AI systems[14:14] AI Agent Studio and Marketplace solve scaling challenges[15:12] Agents as higher-level abstraction for enterprise automation[16:27] Real-world AI use cases across industries shared[18:49] Multi-cloud strategy accelerates enterprise AI adoption[21:16] Partners enable scale with 100 marketplace solutions[23:01] Convergent AI: Consume applications then extend capabilities[26:51] Multi-cloud and multi-model future requires strong governance[27:31] Four-tier security isolation from infrastructure to applications[29:57] AI agents need enterprise-level data residency controls[31:02] Using AI to accelerate cloud migration skills[[33:08] Design thinking to working prototype in days[36:10] Success metrics: Beat your personal best daily[39:55] Why Oracle differs: Only four-tier stack player[43:01] What great looks like for Oracle in the Asia Pacific[45:51] Closing Podcast Information: Bernard Leong hosts and produces the show. The proper credits for the intro and end music are "Energetic Sports Drive." G. Thomas Craig mixed and edited the episode in both video and audio format.

Fresh out of the studio, Arnaud Frade, Managing Partner at Mesh Advisory and author of the upcoming book The Sentient Startup, joins us to explore how AI is fundamentally transforming entrepreneurship through the radical concept of AI as a true co-founder rather than merely a tool. He distinguishes between startups that are merely "AI-onboard" versus true sentient startups where AI operates as part of how the business is run, and introducing the concept of Machine Resources (MR) as a new organizational function alongside Human Resources to manage evolving AI models and maintain knowledge integrity. He emphasizes the co-intelligence model over the sterile human-versus-machine debate, highlighting how the 95% AI project failure rate stems from organizational issues rather than technology problems—a reality perfectly captured by the Banani framework (Brittle, Anxious, Non-linear, Incomprehensible) that explains why companies struggle with AI readiness. Closing the conversation, Arnaud shares what great looks like for The Sentient Startup: being part of the essential conversation about AI governance, accessibility across income levels, child safety, and building better regulations."The entire debate on human versus machine is extremely in a way sterile because it very quickly veers towards the science fiction, the movies, the images we have, and it forgets the point about what can we gain. but what can we gain in the context of having accountability and integrity with it? I think this is where the debate becomes more complex. So for me that's the bigger point is this is not about having less people. It's about having people doing better things and using their talent to really scale what they can do in that model of co-intelligence. " - Arnaud FradeEpisode Highlights:[00:00] Quote of the Day by Arnaud Frade[01:00] Introduction: Arnaud Frade [02:49] Arnaud's career journey [05:27] Advice: stick to values and volunteer for opportunities [07:21] Arnaud's story on his first book about algorithmic marketing [09::32] Sam Altman's solopreneur unicorn comment was the epiphany [10:04] Sentient startup: AI as co-founder, enhanced by co-intelligence [12:15] AI onboard versus AI as true co-founder distinction[16:45] Barriers to entry reducing with AI and agentic models[23:40] MIT report: 95% of AI pilots fail organizationally[28:15] Need overlay systems for effective agentic AI[30:45] Culture: retain humanity when engaging with AI[36:10] Governance frameworks for AI decision-making transparency [40:10] Banani framework: brittle, anxious, non-linear, incomprehensible[42:16] Co-intelligence model versus human versus machine debate[43:19] Goal: be part of the conversation on AI[44:15] Focus on regulations, access equity, and child safety[45:00] ClosingProfile: Arnaud Frade, author of "The Sentient Startup" which will launch on May 19, 2026 via Penguin Random House. Register at thesentientstartup.com for community updatesLinkedIn Profile: https://www.linkedin.com/in/arnaudfrade

Fresh out of the studio and we hit our 500th episode milestone, guest host Yana Fry from Yana TV turns the tables on Bernard Leong, CEO of Dorje AI and host of Analyse Asia, in a special ask-me-anything format. We start with Bernard's journey from finding his first guest to navigating 11 years of podcasting, revealing his 12-word life philosophy: "Learn from everyone, follow no one, observe the patterns, work like hell." Following on, Bernard shares how his theoretical physics background provides the tools on everything from digital transformation to building Dorje AI's vision of reimagining ERP systems. The conversation dives deep into Bernard's pragmatic idealist worldview, product management philosophy and focus on the future. Bernard announces a major rebrand: Analyse Asia is dropping "Asia" to become the Analyse Podcast as it expands to a global audience, marking a new chapter in the show's evolution."For Dorje AI, what great looks like is being able to solve the ERP problem for businesses. It could take five years, ten years, or even two decades — because every technology adoption cycle takes time. We’re at the beginning of a massive shift, but many still cling to the old ways of doing things. The one thing I’ve learned about digital transformation is this: everyone loves transformation, but they hate to change. Everything that people say will happen in two years usually takes five.When I think about Analyse Asia, greatness for me is being able to do an interview without looking at a set of questions — to tease out a guest’s story authentically, without prejudice, without being a fanboy. Just getting the story out. If I can do that, that’s great. Of course, hitting a million subscribers would be fantastic — that’s the next milestone I’m chasing. But for me, it’s always: ‘I’ve reached this milestone — what’s next?’When you think about frugality at the highest level, it’s not about resources — it’s about time. The real measure is how much time you can spend doing what truly matters. That’s what great looks like for me: asking, what’s the minimum amount of time I can make the maximum impact? Maybe I’ll never fully get there. But if we can say we lived this life without regret — that’s enough." - Bernard LeongEpisode Highlights:[00:00] Quote of the Day by Bernard Leong[01:00] Introduction: Yana Fry (Guest Host) and Bernard Leong[02:05] Early Days of Analyse Asia Finding first guest was biggest initial challenge[04:30] Looking at Everything from a corporate executive like a theoretical physicist[06:23] Podcast interviews are combinations of science and art[08:19] Pragmatic idealist philosophy shapes Bernard's worldview[10:17] 12 Word Advice: Learn from Everyone, Follow No One, Observe the Patterns & Work like Hell[13:00] Dorje AI solving fundamental ERP ledger problems[16:15] Attacking competitors' strongest strength - Lessons of history from Kublai Khan[21:34] New is Easy and Right is Hard for Product Management[24:00] Audio to video was hardest podcast pivot[29:25] Japanese craftsman approach keeps Bernard going[33:09] Analyse Asia rebranding to Analyse Podcast globally[39:55] Father's 50-year loyalty shaped Bernard's management philosophy[44:00] Asia is diverse cities, not monolithic continent[46:00] Most problems aren't AI problems after questioning[51:00] What does Great Look Like for Dorje AI, Analyse Asia and Bernard Leong?[52:10] ClosingProfile: - Bernard Leong, Host of Analyse Asia Podcast, CEO of Dorje AI https://dorje.ai, Adjunct Associate Professor from NUS Business School & Institute of Systems Science. - Guest Host: Yana Fry from @yanatvsg which we highly recommend and subscribe to: https://www.youtube.com/@yanatvsgLinkedIn: https://www.linkedin.com/in/yanafry/Podcast Information: Bernard Leong hosts and produces the show. The proper credits for the intro and end music are "Energetic Sports Drive." G. Thomas Craig mixed and edited the episode in both video and audio format.