Inside The Silicon Mind

When we look at the explosive market for LLMs and coding assistants, it’s easy to think we are already on the doorstep of AGI. But what if the systems we call "advanced" are actually operating at the level of a preschooler when it comes to the visual world that enterprises actually care about?

In this episode of Inside the Silicon Mind, we sit down with Andrew Dai, co-founder of Elorian, to unpack a fascinating perspective on the true state of artificial intelligence and the legendary culture that birthed the modern AI revolution. Andrew spent 14 years inside Google Brain, working alongside pioneers like Jeff Dean and Geoffrey Hinton. In 2015, he co-authored the seminal paper on language model pre-training and fine-tuning, a breakthrough that created the foundational triangle of modern LLMs. Today, his new research and product lab, Elorian, is taking on the next massive frontier: moving humanity past "baby vision" and building the models that will unlock true visual AGI.

You'll hear how traditional text and code models are rapidly maturing while enterprise AI remains heavily bottlenecked, because current systems still possess the visual spatial awareness of a preschooler. We dive into the legendary culture of early Google Brain, why “career osmosis” inside elite teams matters more than most researchers realise, and how true technological breakthroughs in physical industries depend on AI that can inherently understand complex visual diagrams.

In this episode:

- Traditional text and code models are rapidly maturing, but enterprise AI is heavily bottlenecked because current models possess the visual spatial awareness of a preschooler.

- The legendary culture of early Google Brain thrived on intellectual freedom, a lack of product launch pressure, and the cross-pollination of ideas over micro-kitchen lunches.

- Ambitious researchers heavily underestimate the power of career "osmosis" – the ability to naturally inherit elite problem-solving frameworks by standing in the same room as greatness.

- True breakthroughs in physical industries like aerospace, agriculture, and data center infrastructure cannot be coded; they require AI that inherently understands complex visual diagrams.

Chapters:

0:00 – The Myth of AGI & The "Baby Vision" Problem

0:45 – Google Brain: The Bell Labs of the AI Era

1:20 – Inside the Culture of Freedom and Innovation

2:50 – Andrew’s Journey: From the UK to Google Brain

4:45 – The 2015 Pre-Training Breakthrough That Changed Everything

7:41 – Why Language Modeling is the Core of Understanding

10:00 – The Playbook Behind the Google Brain Residency Program

13:01 – Geoffrey Hinton’s Philosophy: Modeling the Human Brain

17:25 – Psychological Safety and the Power of Being Wrong

18:51 – The Concept of Osmosis in Elite Teams

21:41 – Thinking Bigger: Moving Past Academic Niches

23:28 – Why the Frontier Labs Weren't Built Inside Google

26:00 – Launching Elorian: The Mission for Visual AGI

28:55 – The Nokia vs. iPhone Era of AI Development

32:05 – Final Thoughts

Book recommendation:

About Andrew:

Andrew Dai is the co-founder and CEO of Elorian, a research startup focused on advancing AI visual reasoning. Before founding Elorian, he spent 14 years at Google, most recently serving as a Research Director at Google DeepMind. During his tenure there, he was a key leader in the development of the Gemini and PaLM 2 models, co-leading Gemini data efforts and PaLM 2 pre-training.

Follow Andrew on LinkedIn:

https://www.linkedin.com/in/andrewdai/

About the show

Inside the Silicon Mind takes you behind the scenes with the founders, CEOs, and VCs building the future of technology. Every week, Firas Sozan sits down with the operators and investors rewriting the rules – from deep-tech startups to the security challenges created by AI.

Hosted by Firas Sozan.

Powered by Harrison Clarke.

Subscribe for more deep dives on:

• AI and the future of software

• Venture capital and company building

• Developer tools, infrastructure, and engineering careers

YouTube: https://www.youtube.com/@InsideTheSiliconMind

Apple Podcasts: https://bit.ly/apple-itsm

Spotify: https://bit.ly/spotify-itsm

Follow the host on LinkedIn:

https://www.linkedin.com/in/firassozan/

Website links:

https://www.harrisonclarke.com/

https://www.harrisonclarkeventures.com/

https://thepmfplaybook.com/

#AI #VisualAGI #GoogleBrain #DeepMind #ArtificialIntelligence #InsideTheSiliconMind

When you watch a movie for the second time, what if you could experience it entirely from the villain's perspective without the director ever having to film a single extra frame?

In this episode of Inside the Silicon Mind, we sit down with Alberto, co-founder of Reactor, to explore a massive paradigm shift at the intersection of entertainment, gaming, and software. Reactor is building the fundamental cloud infrastructure and SDK that allows developers to run real-time, interactive, pixel-generated worlds and "world models" at scale with incredibly low latency. Instead of replacing existing media, Alberto's vision is focused on unlocking net-new experiences, moving humanity away from static, pre-rendered content and into a future where pixels are generated on the fly and tailored uniquely to the viewer.

You'll hear how real-time AI video generation solves the massive financial risk and low ROI that killed early choose-your-own-adventure concepts like Netflix's Bandersnatch by driving the marginal cost of content variants down to near zero. We dive into how this tech could rewrite the education system through highly visual, immersive worlds, how it alters the philosophy of artistic authorship, and why Alberto believes that in ten years, static user interfaces will vanish in favor of fluid UIs that adapt to how a human brain naturally thinks.

In this episode:

- Traditional media is fundamentally static, but real-time video generation introduces highly personalized, dynamic content that changes based on who is watching and when.

- The massive financial bottleneck of branching narratives, like filming five different versions of a scene, is eliminated when content variation costs next to nothing.

- Real-time GenAI enforces active human interaction and collaboration on both sides, keeping artists and creators strictly in the loop rather than replacing them with passive batch automation.

- Visual delivery bypasses text-heavy data bottlenecks, allowing humans to absorb tailored knowledge dramatically faster through interactive educational worlds.

Chapters:

0:00 The Future of Interactive Storytelling

1:09 What is Reactor?

2:15 Moving Past Static Video & The Bandersnatch Problem

7:31 How Real-Time Tech Changes the Concept of Media

10:58 Redefining Artistic Authorship & Active Engagement

14:51 Immersive Visual Tech in the Education System

19:13 Under the Hood: Building Cloud Infra for World Models

25:10 How AI Coding Agents Empower Non-Technical Creators

27:20 10 Years From Now: The End of Static User Interfaces

32:00 Final Thoughts & Book Recommendation

Book recommendation:

ENDURANCE: The Greatest Adventure Story Ever Told, by Alfred Lansing

About Alberto:

Alberto Taiuti is the co-founder and CEO of Reactor, a company building the critical infrastructure layer and SDK for real-time generative video and interactive world models. Before launching Reactor, he was an AR/VR engineer at Apple working on the Vision Pro, and co-founded Luma AI, where he served as CTO and helped pioneer the infrastructure behind one of the world's most widely used 3D and video generation platforms.

About the show

Inside the Silicon Mind takes you behind the scenes with the founders, CEOs, and VCs building the future of technology. Every week, Firas Sozan sits down with the operators and investors rewriting the rules – from deep-tech startups to the security challenges created by AI.

Hosted by Firas Sozan.

Powered by Harrison Clarke.

Subscribe for more deep dives on:

• AI and the future of software

• Venture capital and company building

• Developer tools, infrastructure, and engineering careers

YouTube: https://www.youtube.com/@InsideTheSiliconMind

Apple Podcasts: https://bit.ly/apple-itsm

Spotify: https://bit.ly/spotify-itsm

Follow the host on LinkedIn:

https://www.linkedin.com/in/firassozan/

Website links:

https://www.harrisonclarke.com/

https://www.harrisonclarkeventures.com/

https://thepmfplaybook.com/

#AI #GenerativeAI #RealTimeVideo #InteractiveStorytelling #WorldModels #EntertainmentTech

When you drop contracts, strategy docs, or source code into ChatGPT, who actually sees your prompts?

In this episode of Inside the Silicon Mind, Firas sits down with Jonathan Mortensen, founder and CEO of Confident Security, to unpack the hidden cost of using AI at work: the fact that what you send to AI tools is not end‑to‑end encrypted and may be retained or even used to train models. Jonathan, a two‑time founder with exits to a cybersecurity firm and Databricks, is now building “you bring the AI, we bring the privacy” – private inference modelled on Apple’s Private Cloud Compute, where even the server cannot see your data.

You’ll hear how today’s AI workflows expose legal documents, trade secrets, and customer data, and how cryptography, anonymisation, and remote attestation can lock that back down. You’ll also hear Jonathan’s lessons from previous startups, why “you’re in the business your customers think you are,” and how AI risks like deepfakes, prompt injection, and cognitive atrophy are reshaping how we think about security and trust.

In this episode:

- AI tools introduce serious privacy risks for sensitive business data.

- Confident Security aims to make AI inference private, modelled on Apple’s cryptographic approach.

- Anonymising traffic and proving what code is running are key to protecting user information.

- The founder mindset behind Confident Security is driven by a desire to create real value and “do things the hard way.”

- Learning from customers – including “you’re in the business your customers think you are” – is essential for product development.

- AI technology raises new authenticity concerns for consumers, from deepfakes to AI‑generated “slop.”

- Agentic browsing, prompt injection, and AI‑generated code present emerging security challenges.

- Iteration and fast feedback loops are vital, whether you’re building AI infra or the fastest plane in history.

- Understanding the market requires constant learning, adaptation, and asking for feedback.

Chapters:

00:00 The privacy dilemma in AI usage

02:46 Confident Security: Apple‑style privacy for AI prompts

05:58 The founder mindset: independence, grit, and doing things the hard way

09:12 “You’re in the business your customers think you are”

12:11 Deepfakes, prompt injection, and consumer AI risks

14:47 Overusing AI and cognitive atrophy

18:20 Skunk Works, iteration, and engineering mindset

Book recommendation:

About Jonathan Mortensen

Jonathan Mortensen is the founder and CEO of Confident Security, a company building private inference for AI workloads. Before Confident Security, he founded Gyroscope, a reinforcement learning startup acquired by a cybersecurity firm, and a serverless Postgres company acquired by Databricks.

About the show

Inside the Silicon Mind takes you behind the scenes with the founders, CEOs, and VCs building the future of technology. Every week, Firas Sozan sits down with the operators and investors rewriting the rules – from deep‑tech startups to the security challenges created by AI.

Hosted by Firas Sozan.

🔔 Subscribe for more deep dives on:

• AI and the future of software

• Venture capital and company building

• Developer tools, infrastructure, and engineering careers

YouTube: https://www.youtube.com/@InsideTheSiliconMind

Apple Podcasts: https://bit.ly/apple-itsm

Spotify: https://bit.ly/spotify-itsm

Follow the host on LinkedIn:

https://www.linkedin.com/in/firassozan/

Website links:

https://www.harrisonclarke.com/

https://www.harrisonclarkeventures.com/

https://thepmfplaybook.com/

#AIsecurity #AIprivacy #ChatGPT #Cybersecurity #DataPrivacy

When AI drives the cost of writing software close to zero, what will still matter in 10 years?

Lenny Pruss - General Partner at Amplify, a venture firm managing over $2.5 billion - sits down with Firas to answer the question every founder, engineer, and VC is wrestling with: if anyone can ship software fast and cheap, what actually makes a software company durable anymore?

His answer isn’t what you’d expect.

In this episode:

- Why lead time has collapsed from years to weeks – and what that means for defensibility

- The point‑solution companies Lenny believes will be “absolutely railroaded” by foundation models

- Why distribution, trust, and ecosystem now matter more than code quality alone

- What Lenny looks for in a company today to decide if it will still matter in 10 years

Timestamps:

00:00 - The existential question: will this matter in 10 years?

01:17 - Why traditional technical moats are collapsing

04:07 - The classes of companies that get railroaded by AI

06:27 - What a real moat looks like in 2026

09:00 - Trust, maintenance, and ecosystems as the ultimate software value

17:52 - What happens to engineers when agents write the code

18:11 - Tokens, models, and the trillion‑dollar question

Book recommendation:

The Road to Serfdom, by Friedrich A. Hayek

About Lenny Pruss:

Lenny Pruss is a General Partner at Amplify, a venture capital firm with over $2.5 billion under management. He focuses on early‑stage software and infrastructure investments, backing companies at the intersection of developer tools, data, and AI.

About the show:

Inside the Silicon Mind takes you behind the scenes with the founders, CEOs, and VCs building the future of technology. Every week, Firas Sozan sits down with the operators and investors rewriting the rules – from zero‑to‑one startup journeys to the decisions that define industries.

Hosted by Firas Sozan.

Powered by Harrison Clarke.

🔔 Subscribe for more deep dives on:

• AI and the future of software

• Venture capital and company building

• Developer tools, infrastructure, and engineering careers

YouTube: https://www.youtube.com/@InsideTheSiliconMind

Apple Podcasts: https://bit.ly/apple-itsm

Spotify: https://bit.ly/spotify-itsm

Follow the host on LinkedIn:

https://www.linkedin.com/in/firassozan/

Website links:

https://www.harrisonclarke.com/

https://www.harrisonclarkeventures.com/

https://thepmfplaybook.com/

#VentureCapital #AI #Startup #SoftwareEngineering #AmplifyPartners

Coding used to require years of training and a full engineering team. That era is over.

In this episode, Eli Chen, Principal Technical Architect at AI Fund, explains how and why the cost and effort of coding is quickly dropping toward zero - and what that means for early-stage founders, engineers, and anyone building in the agentic AI era.

Eli has spent 25 years at the frontier of software engineering, from early machine learning to today’s AI-native startup landscape. His central argument is simple: when coding becomes automated, the human skills - architecture, judgment, taste, and customer understanding - become even more important.

In this episode:

Who this episode is for:

Timestamps:

Book recommendation:

About Eli Chen:

Follow Eli on LinkedIn:

https://www.linkedin.com/in/elichen/

About the podcast:

Inside the Silicon Mind goes behind the scenes with the founders, CEOs, and venture capitalists building the future of technology. New episodes every week.

Listen and subscribe:

Follow the host on LinkedIn:

https://www.linkedin.com/in/firassozan/

Website links:

Prevention is dead. The companies still building walls are already losing - and a $70M cybersecurity founder is ready to prove it.

In this episode of Inside the Silicon Mind, Firas sits down with Shachar Hirshberg, co-founder of Artemis - an AI-native cybersecurity platform helping enterprises detect and contain attacks before they cause irreversible damage.

Shachar paints a stark picture: we’re entering an era of fully autonomous AI-driven cyberattacks that can move from initial access to full data exfiltration in mere seconds. Zero-day exploits that once struck every few months will soon number in the hundreds per day. Nation-state-level sophistication that once cost millions is now available to cybercrime gangs for the price of a ChatGPT subscription.

But defenders aren’t powerless. Shachar’s core argument: attackers will never know your business the way you do. The future of cybersecurity isn’t prevention – it’s intelligent, contextual containment that stops damage without stopping your operations.

In this episode, you’ll learn:

Who this episode is for:

CISOs, security leaders, founders, and technical operators who are ready to move beyond prevention and want a battle-tested playbook for detecting, containing, and surviving AI-driven attacks.

Book recommendation:

The Mom Test, by Rob Fitzpatrick

About Shachar Hirshberg:

Shachar Hirshberg is a cybersecurity and AI product leader who’s built and scaled world class enterprise security products at companies like AWS and Demisto. He led product management for Amazon GuardDuty and key cloud security initiatives at AWS, and previously helped define the SOAR category as an early engineering leader at Demisto, acquired by Palo Alto Networks for 560M. With a background in computer science, economics, and an MBA from Harvard Business School, Shachar focuses on building new security categories that solve the most critical challenges for global customers.

Follow Shachar:

https://www.linkedin.com/in/shachar-hirshberg/

About the podcast:

Inside the Silicon Mind goes behind the scenes with the Founders, CEOs, and Venture Capitalists building the future of technology. New episodes every week.

Subscribe to the podcast and never miss an episode:

YouTube: https://www.youtube.com/@InsideTheSiliconMind

Apple Podcasts: https://bit.ly/apple-itsm

Spotify: https://bit.ly/spotify-itsm

Follow the host, Firas Sozan:

https://www.linkedin.com/in/firassozan/

More from Inside the Silicon Mind:

https://www.harrisonclarke.com/

https://www.harrisonclarkeventures.com/

https://thepmfplaybook.com/

Keywords:

AI hacking, AI cyberattacks, autonomous cyberattacks, AI security, zero day exploit, zero day exploit explained, cybersecurity startup, future of cybersecurity, cybersecurity founder interview, AI threat detection, CISO strategy, CISO, cybersecurity 2025, cyber threats AI, AI attacks, cybersecurity playbook, startup founder interview, Shachar Hirshberg, Artemis cybersecurity, AI driven attacks, nation state cyberattacks, inevitable breach, Inside the Silicon Mind

When Russia invaded Ukraine, 80% of Cyberhaven’s engineering team was in Kyiv. Employees were stopped at the border. Families were being separated. And one CEO had to decide what kind of leader he was going to be.

In this episode, Howard Ting - CEO of Cyberhaven - shares the moment he thought it was the end of the company, and the bold decisions that saved it. You’ll hear how he led through a live-fire crisis, why he believes most companies fail because of people rather than product, and what “people-first leadership” really looks like when everything is on the line.

In this episode, you’ll learn:

Chapters:

00:00 Navigating Crisis: The Impact of War on Business

11:46 The People Factor: Why Companies Fail

15:01 Leadership and Personal Growth: A Shift in Perspective

20:02 Work-Life Integration: Balancing Ambition and Family

22:49 Redefining Wealth: Beyond Economic Success

Book recommendation: The 5 Types of Wealth, by Sahil Bloom

Time charting link mentioned: https://www.instagram.com/p/DFu4TTnxt66/

About Howard Ting

Howard Ting is the CEO of Cyberhaven, a data security company protecting the world’s most sensitive information. He has led teams through some of the most extreme operational challenges in startup history and has spent his career building and scaling go-to-market and product teams in high-growth enterprise companies.

About Inside the Silicon Mind

Inside the Silicon Mind goes behind the scenes with the Founders, CEOs, and Venture Capitalists building the future of technology. New episodes every week.

Subscribe to the podcast and never miss an episode.

Video version: https://www.youtube.com/@InsideTheSiliconMind

Website: https://insidethesiliconmind.com/

Follow & Subscribe:

Follow the host on LinkedIn: https://www.linkedin.com/in/firassozan/

Don’t forget to subscribe to the podcast for more conversations with founders, operators, and investors shaping the future of technology.

Keywords: Howard Ting, Cyberhaven, CEO, startup, Silicon Valley, entrepreneurship, crisis leadership, cybersecurity, wartime CEO, why startups fail, people-first leadership, Ukraine war

You Don’t Realise What You’re Giving Up When You Become a Founder | Cosmin Nicolaescu

Most people talk about the upside of being a founder.

Cosmin Nicolaescu (Microsoft → Stripe → Brex → founder & CEO of Accrual) talks about the cost – losing your optionality, carrying the weight of every hire, and living with a handful of decisions that can define your entire career.

In this episode, we dive into:

- Why becoming a founder means giving up the “I’ll just go do something else” optionality.

- The loneliness of the role and why even the best co‑founders can’t fully remove it.

- How Cosmin’s time at Microsoft, Stripe, and Brex prepared him – and what still shocked him as a founder.

- Why growth problems are the “best problems” to have, and the hidden downside of losing small‑team camaraderie.

- How he thinks about keeping Accrual small and focused while still scaling fast.

- The idea that only a handful of decisions truly change a company’s trajectory – and how he decides which ones to sweat.

- What he optimises for now: impact, learning, and building from within rather than hiring “free agents”.

Chapters

00:00 Introduction – “Locked In” and the Long‑Term Cost of Founding

01:56 The Loneliness and Constant Trade‑Offs of Being a Founder

04:10 Cosmin’s Journey: From Microsoft to Stripe and Brex

06:54 Leaving Stripe for Brex – Using the Regret Minimisation Framework

10:01 Starting Accrual – Deep Domain Knowledge and Customer Impact

14:12 Choosing an Industry – Why Accounting and Financial Infrastructure

16:13 Solving for Impact – Automating Mundane Work for Accountants

19:07 Customer Obsession – Acquisition, Retention, and Expansion

23:05 Growth Problems, Change Management, and Staying Small While Scaling

25:48 From Operator to Founder – What Actually Changes

27:56 Books, Parenting, and How Cosmin Thinks About Learning

Book recommendations

- Project Hail Mary, by Andy Weir

- The Anxious Generation, by Jonathan Haidt

Connect with Cosmin

LinkedIn: https://www.linkedin.com/in/cosminn/

About Inside the Silicon Mind

Inside the Silicon Mind is a podcast powered by Harrison Clarke, exploring how builders, founders, and technical leaders use AI and frontier tech to change industries from the inside out.

Subscribe for more Silicon Valley operator insights

YouTube: https://www.youtube.com/@InsideTheSiliconMind

Apple Podcasts: https://bit.ly/apple-itsm

Spotify: https://bit.ly/spotify-itsm

Follow the host on LinkedIn

https://www.linkedin.com/in/firassozan/

More from Harrison Clarke

https://www.harrisonclarke.com/

https://www.harrisonclarkeventures.com/

https://thepmfplaybook.com/

#founder #startups #entrepreneurship #leadership #InsideTheSiliconMind

AI has just lowered the bar to hacking. In this episode, cybersecurity founder Peyton Smith explains how script‑kiddie teenagers, China, and Russia are all using AI to probe critical infrastructure - and what defenders have to do next.

Peyton spent years on CrowdStrike’s red team, paid by Fortune 1000 companies to break into “secure” environments and show them where their defences really failed. Now he’s building an AI‑powered platform to continuously stress‑test legacy infrastructure before attackers find the holes.

In this episode, you’ll learn:

Timestamps:

Book mentioned:

The Hard Thing About Hard Things, by Ben Horowitz

Listen if you work in tech, security, or you’re just trying to understand how AI is reshaping the offence-defence balance in cyberspace.

Video version: https://www.youtube.com/@InsideTheSiliconMind

Follow the host on LinkedIn: https://www.linkedin.com/in/firassozan/

#ai #CyberSecurity #Hacking #CriticalInfrastructure #InsideTheSiliconMind

In this conversation, Neil Patel breaks down how AI is quietly transforming healthcare delivery, improving workflows, reducing administrative waste, and changing the future of work in healthcare. We explore what this means for doctors, patients, families, and the broader healthcare system.

We cover:

Neil also shares why he believes AI could help raise the quality of care, reduce variation across providers, and unlock a more accessible healthcare system for more people..

Key moments:

Book recommendation: Sapiens: A Brief History of Humankind, by Yuval Noah Harari

Video version: https://www.youtube.com/@InsideTheSiliconMind

Website: https://insidethesiliconmind.com/

Follow & Subscribe:

Follow the host on LinkedIn: https://www.linkedin.com/in/firassozan/

Don’t forget to subscribe to the podcast for more conversations with founders, operators, and investors shaping the future of technology.