Podcast
The CTO Show with Mehmet Gonullu
Hosted by Mehmet Gonullu · EN
Broadcasting from Dubai, The CTO Show with Mehmet explores the latest trends in technology, startups, and venture funding. Host Mehmet Gonullu leads insightful discussions with thought leaders, innovators, and entrepreneurs from diverse industries. From emerging technologies to startup investment strategies, the show provides a balanced view on navigating the evolving landscape of business and tech, helping listeners understand their profound impact on our world.
mehmet@yassiventures.com
43episodes
Episodes
Newest firstAll episodes
#601 The AI Bottleneck Is No Longer GPUs. It’s Energy and Memory | Eugene Cheah
4d ago00:45:37Tap to summarizeIn this episode of The CTO Show with Mehmet, Mehmet sits down with Eugene Cheah, CEO of Featherless AI. The AI bottleneck is no longer just GPU access. Power, memory, inference cost, and model reliability are becoming the real constraints.Eugene reframes the AI infrastructure debate away from a simple race for bigger models and more chips. The conversation connects energy capacity, HBM shortages, open source model adoption, linear attention architectures, and the enterprise need for predictable AI systems. It also challenges the assumption that the best AI strategy is always to use the largest available model.If you are building, investing in, or operating AI infrastructure, this conversation gives a clearer view of where AI economics, hardware constraints, and production reliability are heading.About the GuestEugene Cheah is the CEO of Featherless AI, an AI startup making open source AI models accessible through a single platform.Featherless AI started from AI research and optimization work around RWKV architecture, with a focus on reducing inference cost and making AI models more accessible. Eugene’s work sits at the intersection of open source AI, model efficiency, GPU infrastructure, HBM constraints, and inference optimization.He is well positioned to frame this shift because Featherless AI works directly on the infrastructure layer between developers, open models, and production inference.LinkedIn: https://www.linkedin.com/in/eugene-cheah-a47791126/Website: https://featherless.aiKey TakeawaysAI infrastructure constraints are shifting from GPU access to power, memory, and inference efficiency.HBM scarcity becomes more serious as models and context windows continue to grow.Bigger models do not solve the enterprise problem of reliable execution.Open source models are becoming strong enough to replace many closed model use cases.Fine-tuned smaller models can outperform frontier models on narrow enterprise tasks.Nvidia’s moat weakens when developers can move workloads across more hardware choices.Linear attention architectures matter because quadratic memory scaling is economically unsustainable.Enterprises value model control when closed providers change, deprecate, or restrict models too often.What You Will LearnThe real infrastructure bottlenecks behind AI deployment beyond GPU availability.How HBM pressure affects model size, context length, and inference economics.Why energy capacity can delay AI infrastructure even when chips are already available.How open source models are changing enterprise AI adoption and deployment control.Why smaller fine-tuned models can beat larger models on specific production tasks.When linear attention architectures reduce memory demand compared with transformer attention.What hardware choice, model portability, and local inference mean for AI infrastructure strategy.Episode Highlights00:00 — AI infrastructure moves beyond the GPU race03:30 — Nvidia, AMD, and Huawei follow different hardware strategies07:30 — Power becomes the first AI infrastructure bottleneck08:30 — HBM pressure exposes the memory constraint12:00 — AI follows the same pluralism as databases15:00 — Developers start with big models, then specialize18:30 — Transformer memory scaling becomes an economic problem23:30 — Hardware choice starts weakening platform lock-in29:30 — Reliability matters more than raw intelligence36:00 — Open source gives enterprises model control41:30 — Small models can now build real applicationsResources MentionedFeatherless AI: https://featherless.aiRWKV architecture: AI architecture referenced by Eugene as part of Featherless AI’s research backgroundListen NowAvailable on all major podcast platforms and YouTube.Connect with the ShowFollow The CTO Show with Mehmet for more conversations at the intersection of technology, startups, and venture capital.
Transcribe →#600 AI Reliability Is a Business Risk. Not Just an Engineering Problem | Helen Gu
1w ago00:37:26Tap to summarizeIn this episode of The CTO Show with Mehmet, Mehmet sits down with Helen Gu, Founder and CEO of InsightFinder AI. Helen brings decades of research in distributed system reliability, anomaly detection, and AI-driven operations. The conversation focuses on why AI reliability is becoming a business risk, not just an engineering issue.The conversation reframes AI observability as a production control layer for enterprises deploying AI agents. Helen explains why traditional DevOps and SRE practices are not enough when systems are probabilistic, model behavior changes, data shifts, prompts evolve, and agents begin taking actions across workflows.If you are building, investing in, operating, or leading AI systems inside enterprise environments, this conversation gives you a practical frame for reliability, drift, runtime monitoring, and accountability.About the GuestHelen Gu is the Founder and CEO of InsightFinder AI, and a professor at North Carolina State University. InsightFinder AI was founded from her research in distributed system reliability using AI technology.Helen has worked on anomaly detection, prediction, diagnosis, and system reliability since the late 1990s. She also spent a sabbatical year at Google evaluating anomaly detection algorithms, which later helped shape the foundation for InsightFinder AI.LinkedIn: https://www.linkedin.com/in/helen-gu-b1aa42b6/Website: https://insightfinder.com/Key TakeawaysAI systems can fail silently while still returning confident answers.AI reliability is becoming a business risk, not only an engineering concern.Multi-agent systems can spread upstream mistakes across business workflows quickly.Traditional SRE practices do not fully cover model behavior, prompts, and data drift.Runtime monitoring matters more once AI moves from sandbox testing to production.Observability alone is not enough without diagnosis, recommendations, and remediation.Model drift can change business outcomes even when infrastructure appears healthy.Human review shifts from doing work to supervising AI decisions and guardrails.What You Will LearnWhy probabilistic AI systems require different reliability practices than software systems.How model drift and data drift change production behavior over time.What silent AI failure looks like inside enterprise workflows.The reason sandbox testing misses real production AI failure cases.How runtime monitoring helps detect hallucinations, bias, leakage, and accuracy issues.Why AI observability must connect infrastructure, data, prompts, models, and business outcomes.What leadership teams need to consider before AI agents begin taking actions.Episode Highlights00:00 — Helen Gu frames AI reliability from research02:30 — AI systems answer confidently even when wrong04:30 — SRE lessons do not fully transfer to AI07:00 — AI reliability needs fine-grained runtime metrics08:30 — Silent failure creates hidden business damage10:00 — Multi-agent mistakes propagate faster than humans12:00 — Model drift changes outcomes without warning15:00 — Sandboxes miss production AI behavior18:00 — Observability must become actionable control21:30 — AI reliability becomes a leadership responsibility24:30 — AI Labs test prompts, models, and datasets28:30 — AI agents become part of enterprise workflows31:30 — Responsible AI starts with accepting failure riskListen NowAvailable on all major podcast platforms and YouTubeConnect with the ShowFollow The CTO Show with Mehmet for more conversations at the intersection of technology, startups, and venture capital.
Transcribe →#599 AI Agents Are the New Attack Surface. Security Teams Are Already Behind | Jason Remillard
2w ago00:45:43Tap to summarizeIn this episode of The CTO Show with Mehmet, Mehmet sits down with Jason Remillard, Founder of Data443. Jason brings more than 30 years of cybersecurity, data security, infrastructure, and enterprise risk experience. The conversation focuses on the gap between AI adoption speed and the security operating models still built for slower systems.The episode reframes AI security as an execution and visibility problem, not only a model risk problem. Jason argues that security teams lose when they only block users, rely on slow approval workflows, or assume old SOC models can handle AI agents, MCPs, SaaS sprawl, and machine-speed data movement.If you are leading cybersecurity, enterprise IT, AI adoption, or digital infrastructure strategy, this conversation gives you a practical lens for where the real exposure is forming.About the GuestJason Remillard is the Founder of Data443, a data security company focused on securing data across systems, users, and enterprise workflows. His career spans more than 30 years, from early systems operations and ISP infrastructure to enterprise security and regulated environments.Jason has worked across cybersecurity, data protection, ransomware recovery, threat intelligence, DLP, attack surface management, and AI-related security challenges. His perspective is grounded in the operational reality of how users, security teams, and business units behave when controls create friction.LinkedIn: https://www.linkedin.com/in/jremillard/Website: https://data443.com/Key TakeawaysAI agents expand the attack surface faster than security teams can govern with manual workflows.End users bypass controls when security becomes a blocker to legitimate business execution.DLP cannot solve data loss when users can photograph, move, and re-enter information elsewhere.Security teams need to enable safer decisions, not only enforce binary allow-or-deny rules.Inference can reduce AI security costs when models are trained for specific enterprise use cases.Threat intelligence must track agents, connectors, APIs, and machine actions as risk-bearing actors.Post-quantum risk matters because encrypted data can be stored now and decrypted later.Cyber resilience starts with assuming breach, not assuming the perimeter still holds.What You Will LearnThe reason cultural failure still sits behind many enterprise security failures.How AI agents change visibility across SaaS, APIs, Shadow IT, and enterprise data flows.Why traditional exception management breaks when AI decisions happen in milliseconds.How inference can help security teams operate faster without relying only on GPUs.What MCP and agent-to-agent workflows mean for API governance and connector risk.Why post-quantum security is already relevant for long-lived sensitive data.The practical starting point for cyber resilience when attacks cannot be fully prevented.Episode Highlights00:00 — Jason Remillard frames three decades in cybersecurity04:30 — Security failure starts with not-my-job thinking08:30 — DLP breaks when users bypass friction12:00 — AI agents change enterprise visibility13:30 — Approval workflows cannot match AI speed17:30 — Non-human actors create identity risk20:30 — AI defense depends on trained inference27:00 — Multimodal input changes user behavior28:30 — MCP turns APIs into hidden risk31:00 — Attackers gain the same AI velocity35:00 — Quantum risk makes stored data vulnerable39:00 — Resilience starts by assuming breachListen NowAvailable on all major podcast platforms and YouTube.Connect with the ShowFollow The CTO Show with Mehmet for more conversations at the intersection of technology, startups, and venture capital.
Transcribe →#598 AI Pilots Don’t Fail. Enterprise Systems Do | Omid Pakseresht, CEO of Goodfolio
2w ago00:39:40Tap to summarizeIn this episode of The CTO Show with Mehmet, Mehmet sits down with Omid Pakseresht, CEO of Goodfolio. Omid works on enterprise AI systems that move beyond pilots and into real business workflows.The conversation reframes enterprise AI failure as a systems problem, not a model problem. Omid argues that most AI initiatives break because the workflow, ownership model, governance layer, audit trail, and adoption path were never designed properly. The model may work, but the enterprise system around it often does not.If you are building, investing in, or leading enterprise AI adoption, this conversation gives you a clearer way to judge whether an AI initiative is ready for production or stuck as another pilot.About the GuestOmid Pakseresht is the CEO of Goodfolio, a company focused on helping enterprises build and scale AI systems inside real workflows.His background is in product and technology, with a particular focus on finance. He has spent around 10 years building and scaling AI solutions in enterprise environments.Omid is well placed to frame this signal because his work sits at the point where AI models meet workflow design, governance, compliance, and business outcomes.LinkedIn: https://www.linkedin.com/in/omidpakseresht/Website: https://goodfolio.comKey TakeawaysMost enterprise AI fails because the system around the model was never built.A working AI pilot is not proof that the business is ready for production.AI adoption fails when it is treated as a data science project.Workflow owners must be part of the AI design process from the beginning.Human-in-the-loop fails when humans become late-stage QA gates.AI can create new bottlenecks when upstream productivity increases faster than downstream capacity.Regulated AI needs audit trails, governance layers, risk monitoring, and clear decision rights.AI ROI must be tied to business outcomes, not seat counts or software usage.What You Will LearnThe difference between an AI tool and an AI system inside an enterprise workflow.How AI pilots fail after the proof of concept looks successful.Why model quality is rarely the biggest barrier to enterprise AI adoption.How compliance, governance, and auditability shape production AI.What changes when AI becomes embedded into regulated workflows.Why AI can move bottlenecks rather than remove them.How leaders should evaluate AI ROI through outcomes instead of software spend.Episode Highlights00:00 — Enterprise AI failure starts beyond the model02:00 — Proofs of concept became the easy part04:00 — Workflow fit beats model quality in adoption05:30 — AI cannot remain a data science project08:30 — Production AI needs more than a model12:00 — Compliance workflows expose AI bottlenecks17:30 — Human-in-the-loop needs a better framing20:00 — Governance becomes table stakes for enterprise AI24:00 — AI ROI must connect to business outcomes28:00 — AI exposes process gaps before scalingResources MentionedGoodfolio: https://goodfolio.comInspector: Goodfolio tool for compliance review of marketing assets in regulated industriesAI agents: discussed in the context of compliance workflowsModel governance: discussed as a production requirementEvaluation pipelines: discussed as part of production AI systemsPrompt engineering versioning: discussed as part of AI system managementRisk monitoring: discussed as part of regulated AI adoptionData lakes: discussed as a comparison point for large enterprise technology projectsListen NowAvailable on all major podcast platforms and YouTube.Connect with the ShowFollow The CTO Show with Mehmet for more conversations at the intersection of technology, startups, and venture capital.
Transcribe →#597 Dashboards Are Dead. AI Agents Replace the Forecast Call | Laura Fu, GTM Architect at DevRev
3w ago00:44:37Tap to summarizeIn this episode of The CTO Show with Mehmet, Mehmet sits down with Laura Fu, GTM Architect at DevRev. Laura brings a RevOps and sales enablement lens to a question many GTM leaders are now facing: AI does not fix sales by sitting on top of old workflows.The conversation reframes AI in go-to-market as an operating model problem, not a tooling problem. Laura argues that AI-native execution requires new feedback loops, better data capture, agent-readable systems, and a different view of enablement. The strongest claim is that dashboards and forecast calls become less central when agents can surface the signal directly.If you are leading, building, or investing in enterprise sales organizations, this conversation gives you a sharper way to think about AI-native GTM, CRM architecture, RevOps, sales enablement, and pipeline execution.About the GuestLaura Fu is the GTM Architect at DevRev, focused on improving go-to-market efficiency and how revenue organizations operate with AI.She is the author of Designing for Excellence: Sales Enablement in the AI Native World, a book about using AI to make sales enablement and GTM engines more fluid and operational.Laura is the right person to frame this signal because she connects sales enablement, RevOps, CRM systems, data quality, and AI agents into one operating model.LinkedIn: https://www.linkedin.com/in/laurazfu/Key TakeawaysAI does not make broken sales processes better, it exposes where the process was weak.Sales teams still move at human speed, but expectations now move at AI speed.AI-native GTM requires workflow redesign, not summaries copied into old systems.Traditional enablement fails when training is disconnected from the moment of need.CRM becomes more valuable as memory and context, not as a manual reporting database.Dashboards lose power when agents can detect revenue signals directly.Poor data quality breaks trust in AI faster than poor user adoption.RevOps teams will shift from analysts to GTM engineers who build and orchestrate systems.What You Will LearnThe difference between AI adoption and AI-native sales execution.How AI changes sales enablement from a training function into an operating system.Why dashboards become less useful when agents can scan signals directly.The CRM requirements that matter when agents need read and write access.How real-time feedback loops can reshape sales messaging, pricing, and positioning.Why data quality and change management decide whether AI tools get trusted.What an AI-first revenue organization could look like from day one.Episode Highlights00:00 — Laura Fu frames AI-native sales enablement02:30 — Sales teams face AI-speed expectations06:00 — AI adoption does not change execution09:30 — Traditional enablement was already broken12:00 — Enablement becomes a system, not function15:30 — The AI enablement flywheel takes shape20:30 — Change management breaks AI adoption first25:00 — Feedback loops separate messaging from delivery28:00 — Pipeline creation remains the strongest signal30:00 — Dashboards are dead in agent-led RevOps36:30 — AI finds pipeline signals faster39:00 — GTM engineers replace analyst-heavy RevOps42:30 — Laura shares the book and podcastResources MentionedDesigning for Excellence: Sales Enablement in the AI Native World by Laura Fu: Available on Amazon, Barnes & Noble, and bookstores: https://www.amazon.com/dp/B0FVBKGK4ZState of the AI Union: Laura Fu’s podcast on Apple Podcasts: https://podcasts.apple.com/gb/podcast/state-of-the-ai-union/id1851548376 Listen NowAvailable on all major podcast platforms and YouTube.Connect with the ShowFollow The CTO Show with Mehmet for more conversations at the intersection of technology, startups, and venture capital.
Transcribe →#596 AI Agents Need Managers. Not Prompts | Ross Barnes
3w ago00:43:01Tap to summarizeIn this episode of The CTO Show with Mehmet, Mehmet sits down with Ross Barnes, Founder of Galahad Group. Ross brings a rare operator view on AI adoption, shaped by his background as Global CTO at a WPP agency and his current work building AI platforms and adoption frameworks.The conversation reframes agentic AI as a management problem, not a prompting problem. Ross argues that useful AI systems need purpose, boundaries, delegation, accountability, and human judgment. The episode moves away from tool selection and focuses on how companies should structure AI work before shadow systems, weak guardrails, and legacy processes become operational risks.If you are leading AI adoption, building AI-native workflows, investing in enterprise AI, or operating a startup, this conversation gives you a practical lens for separating useful systems from AI theater.About the GuestRoss Barnes is the Founder of Galahad Group, an AI company focused on AI enablement, adoption, and building its own AI platforms. He previously served as Global CTO at a WPP agency and has worked in digital media, marketing, and SEO since 2001.Ross created frameworks including cognitive scaffolding and IKIGAI AI to help companies identify where AI should support human work rather than replace judgment. His work focuses on AI adoption that starts with people, not tools.LinkedIn: https://www.linkedin.com/in/rossbarnes/Website: https://galahadgroup.co.ukKey TakeawaysAI adoption fails when companies start with tools instead of human work.Agentic AI requires management discipline, not better prompt tricks.Shadow AI is already creating invisible data and governance risks inside companies.Good AI agents need narrow tasks, clear boundaries, and permission to fail safely.Startups gain speed because AI compresses the distance between idea and execution.Enterprises still win where trust, liability, safety, and brand matter.AI will expose weak culture faster than it replaces headcount.Future visibility depends on speaking to both humans and machines.What You Will LearnThe difference between cognitive infrastructure and another AI tool.How IKIGAI AI identifies which tasks should involve agents.Why shadow AI is already active inside many organizations.How to manage AI agents like junior team members.When startups gain an AI advantage over enterprises.What enterprises still protect better than AI-native startups.How LLM discovery changes brand visibility and content strategy.Episode Highlights00:00 — Ross Barnes frames AI beyond marketing tools03:30 — Cognitive scaffolding starts with human work06:00 — IKIGAI AI separates human judgment from automation10:00 — Shadow AI is already inside companies11:00 — Agentic workflows work best inside CRM14:00 — AI adoption exposes fear and sunk costs17:00 — Personal AI stacks compound with context19:30 — Marketing shifts from campaigns to systems22:00 — LLM discovery changes brand visibility29:30 — Agents need boundaries like coworkers35:00 — Startups move faster because legacy disappears39:30 — AI-native companies still need accountable cultureResources MentionedGalahad Group: https://galahadgroup.co.ukIKIGAI AI diagnostic: https://galahadgroup.co.uk/ikigaiGRAIL: Galahad Group platform for building authority in LLMsRoss Operating System: Ross Barnes’ personal multi-agent workflow systemIKIGAI AI: Galahad Group diagnostic frameworkCognitive scaffolding framework: Ross Barnes’ framework for AI-supported human work Listen NowAvailable on all major podcast platforms and YouTube.Connect with the ShowFollow The CTO Show with Mehmet for more conversations at the intersection of technology, startups, and venture capital.
Transcribe →#595 Engineers Optimize Code. They Mismanage Money | Stanley Leong, Author of Engineering Your Finances
4w ago00:44:12Tap to summarizeIn this episode of The CTO Show with Mehmet, Mehmet sits down with Stanley Leong, private wealth advisor and author of Engineering Your Finances. The core tension is simple: technical people often apply logic to money, but still make emotional financial decisions.The conversation reframes wealth planning for engineers, founders, and senior tech professionals as a risk management problem rather than a returns problem. Stanley explains why concentrated employer stock, overexposure to technology stocks, late retirement planning, and AI-generated financial advice can create hidden fragility for high earners.If you are building, investing in, or leading in enterprise technology, this conversation gives you a sharper way to think about personal wealth, equity compensation, and risk before it becomes expensive.About the GuestStanley Leong is a private wealth advisor and the author of Engineering Your Finances. He holds a master’s degree in electrical engineering from Cornell, previously designed computer chips at IBM, and later moved into financial advisory after being laid off during the tech downturn.His work focuses on helping technology professionals think through retirement planning, concentrated stock risk, tax-aware savings, behavioral finance, and long-term financial security.LinkedIn: https://www.linkedin.com/in/stanleycleong/Website: https://engineeringyourfinancesbook.comKey TakeawaysHigh income can hide poor financial structure until a job loss or market shock exposes it.Engineers often underestimate how emotional their financial decisions become under stress.Employer stock can create wealth, but it can also quietly dominate net worth.Diversification is not owning several tech stocks if the entire portfolio depends on one sector.Retirement planning changes for tech professionals because career durability is not guaranteed.AI can answer financial questions, but outdated or incomplete advice can still create real damage.Founders carry concentrated risk even when their company is growing and well funded.Good investing starts with risk first, return second.What You Will LearnThe most common financial mistake Stanley sees among technology professionals.How concentrated employer stock becomes a hidden risk over time.Why engineers can rationalize emotional money decisions better than most people.When high income stops being an advantage and becomes a planning trap.How the seven key areas of financial planning create a more systematic approach.Why after-tax 401k plans and mega backdoor Roth strategies matter for high earners in the US.What separates gambling from investing when evaluating financial decisions.Episode Highlights00:00:00: Why an engineer became a wealth advisor00:05:30: Tech portfolios often carry hidden risk00:08:30: Finance overwhelms analytical people fast00:11:30: Gambling mindset follows engineers into investing00:17:30: Seven areas make planning more systematic00:21:30: Logic can disguise emotional money decisions00:26:30: Stock options create concentrated financial exposure00:32:30: Late savers need structure before returns00:37:30: Founders carry risk they cannot diversify00:40:30: Investing starts with asking what failsResources MentionedEngineering Your Finances by Stanley Leong: https://engineeringyourfinancesbook.comFIRE: Financial independence, retire early401k, Roth IRA, after-tax 401k, mega backdoor Roth: Retirement and tax planning structures discussedListen NowAvailable on all major podcast platforms and YouTube.Connect with the ShowFollow The CTO Show with Mehmet for more conversations at the intersection of technology, startups, and venture capital.
Transcribe →#594 AI Is Being Deployed Without Control. Security Is Playing Catch Up | Tim Freestone
4w ago00:37:01Tap to summarizeIn this episode of The CTO Show with Mehmet, Mehmet sits down with Tim Freestone, Chief Strategy Officer at Kiteworks. AI is already inside the enterprise, but control is not keeping pace.The conversation reframes AI security as a data control problem rather than a tooling problem. Tim argues that agents are not just another interface. They act, call tools, move data, and introduce a new identity layer that most enterprise security architectures were not designed to govern.If you are leading, securing, building, or investing in enterprise AI systems, this conversation clarifies where the real risk sits: data access, agent identity, sovereignty, and governance.About the GuestTim Freestone is the Chief Strategy Officer at Kiteworks, a company focused on secure content communication and data protection. His background includes roles at Contrast, Fortinet, NetApp, and over 10 years running his own business supporting technology and cybersecurity companies.Tim brings more than 22 years of experience across cybersecurity, strategy, go-to-market, and enterprise security. His perspective is grounded in how enterprises are actually deploying AI, where governance is lagging, and why data layer control is becoming central to AI security.LinkedIn: https://www.linkedin.com/in/freestone/Website: https://www.kiteworks.comKey TakeawaysAI adoption is no longer waiting for enterprise readiness or formal governance.Employees are already creating shadow AI risk through uncontrolled tool usage.AI agents introduce a new identity layer that security teams must govern.Data protection becomes harder when agents can access information at machine speed.Sovereignty is no longer just about where data is stored.Frontier AI models force enterprises to choose between control and capability.Security architectures built around infrastructure need stronger data layer controls.AI-powered vulnerability discovery changes the speed and scale of cyber risk.What You Will LearnThe difference between chatbots, copilots, and agents in enterprise environments.How uncontrolled AI usage creates hidden exposure inside organizations.Why agent identity needs to be governed like human identity.The reason data security becomes the starting point for AI governance.How sovereignty changes when enterprise data moves through external models.What CTOs and CISOs should prioritize when AI enters production.Why AI-specific security roles are becoming necessary inside enterprises.Episode Highlights00:00 - Why AI security starts with enterprise readiness02:30 - AI is being deployed before governance catches up04:30 - Agents act differently from chatbots and copilots07:00 - Shadow AI creates a new enterprise exposure layer10:30 - AI agents become new actors inside security architecture13:30 - Data layer control becomes the security priority15:00 - Sovereignty becomes harder when AI moves data18:30 - On-prem interest returns as control concerns rise24:00 - AI models change the vulnerability discovery equation29:30 - Agent native security starts with controlled dataListen NowAvailable on all major podcast platforms and YouTube.Connect with the ShowFollow The CTO Show with Mehmet for more conversations at the intersection of technology, startups, and venture capital.
Transcribe →#593 AI Agents Are Becoming the Workforce for Events | Ritesh Patel, CEO & Co-Founder, Ticket Fairy
Apr 2700:33:26Tap to summarizeIn this episode of The CTO Show with Mehmet, Mehmet sits down with Ritesh Patel, CEO and Co-Founder of Ticket Fairy. He has built a full-stack operating system for the global events industry, spanning ticketing, payments, marketing, and AI.The conversation reframes event technology as an infrastructure problem, not a commerce problem. Ticketing looks simple on the surface, but hides deep system complexity, fragile scaling layers, and continuous engineering trade-offs. AI is not simplifying this stack. It is expanding both capability and risk, especially in fraud, automation, and operational control.If you are building or investing in AI infrastructure, marketplaces, or vertical SaaS, this conversation sharpens how complexity, defensibility, and automation actually play out in production systems.⸻About the GuestRitesh Patel is the CEO and Co-Founder of Ticket Fairy, a platform that provides a full operating system for the independent events industry, including ticketing, CRM, marketing technology, fintech, and AI. He has spent more than a decade producing over 500 events and building systems that address the operational and financial constraints of the industry. His perspective comes from running both sides of the system, event production and infrastructure, which shapes how he approaches automation, fraud, and scalability.LinkedIn: https://www.linkedin.com/in/riteshdpatel/⸻Key TakeawaysTicketing systems look simple, but operate as highly complex distributed infrastructure.AI agents make fraud more effective by mimicking real user behavior at scale.Event platforms require continuous engineering cycles, often running close to 24 hours a day.Defensibility in event tech comes from relationships and capital layers, not software features.Most events are not profitable for years, mirroring early-stage startup dynamics.Centralized systems can solve fraud problems more effectively than blockchain approaches.Real-time data at micro-level granularity drives marketing and conversion performance.Vertical SaaS fails when it tries to serve everyone instead of owning a specific segment.⸻What You Will LearnThe hidden system complexity behind seemingly simple ticketing platformsHow AI agents bypass traditional bot detection and fraud controlsWhy feature flags and modular architecture are critical in vertical SaaSThe economics of event businesses and why profitability is delayedHow real-time behavioral data improves conversion and marketing outcomesWhy blockchain fails to solve most real-world ticketing problemsThe role of AI agents as operational workforce in resource-constrained industries⸻Episode Highlights00:00 — Why simple products hide extreme system complexity03:00 — Event infrastructure complexity most people underestimate05:30 — AI agents make fraud harder to detect08:30 — Trust layer challenges in event platforms11:00 — How to architect systems that survive demand spikes13:00 — Real-time data as a competitive advantage15:00 — Why most events fail financially early17:00 — Pricing models shift cost to the consumer19:00 — Defensibility comes from relationships not software27:00 — AI agents as workforce for event operations⸻Resources MentionedTicket Fairy: https://ticketfairy.comRedis: In-memory data store used for session managementWordPress: Website framework mentioned in comparisonBlockchain and NFT communities: Used for token-gated access use cases⸻Listen NowAvailable on all major podcast platforms and YouTube⸻Connect with the ShowFollow The CTO Show with Mehmet for more conversations at the intersection of technology, startups, and venture capital
Transcribe →#592 Stop Treating AI Like Software. It Is Workforce Infrastructure | Karl Simon, CTO, Subatomic AI
Apr 2400:41:38Tap to summarizeIn this episode of The CTO Show with Mehmet, Mehmet sits down with Karl Simon, Co-Founder and CTO at Subatomic AI. Karl is building orchestration infrastructure for AI agents and enterprise workflows, focused on turning AI into operational capacity rather than isolated tools.AI adoption is often framed as a model problem. This conversation reframes it as a systems problem. The gap is not model capability but data quality, workflow design, and orchestration. The discussion breaks down why AI agents perform well in demos but fail in production, and why observability and context are now core requirements for enterprise AI.If you are building, operating, or investing in enterprise AI systems, this conversation clarifies where value is created and where most implementations fail.⸻About the GuestKarl Simon is the Co-Founder and CTO at Subatomic AI, a company focused on orchestration layers for enterprise AI workflows. His work centers on agentic systems, data integration, and operationalizing AI across business functions.He has spent decades helping companies modernize across data, cloud, and AI systems, with a focus on automation, optimization, and enterprise-scale transformation.He is building infrastructure that treats AI as a workforce layer, not a software feature.LinkedIn: https://www.linkedin.com/in/karlsimon⸻Key TakeawaysAI failures in enterprises are driven by data and workflow gaps, not model limitationsAI agents succeed only when guided by structured workflows and bounded contextData quality issues scale faster with AI, amplifying errors across systemsObservability is required to trust and operate AI in production environmentsEnterprise AI requires orchestration across multiple systems, not isolated toolsAI should be treated as workforce capacity, not a software deploymentSOPs and workflows must evolve continuously or AI will reinforce inefficienciesROI from AI comes from time reallocation and revenue expansion, not just cost reduction⸻What You Will LearnWhy AI models are not the primary bottleneck in enterprise adoptionHow data quality and context directly impact AI output reliabilityThe difference between automation, integration, and orchestration in AI systemsWhat causes AI agents to fail when moving from demo to productionHow observability frameworks enable trust and auditability in AI workflowsThe concept of AI coworkers and how they fit into enterprise operationsWhat CTOs should prioritize first to achieve early ROI from AI⸻Episode Highlights00:00 — AI models are not the real problem02:00 — Orchestration is the missing layer in enterprise AI04:00 — Why AI fails without context and trained data06:30 — Data quality issues break AI systems at scale09:00 — Orchestration vs automation and integration explained12:00 — Trust, auditability, and observability in AI systems16:00 — AI as workforce infrastructure, not software20:00 — Can AI optimize broken enterprise workflows27:00 — AI in regulated industries and compliance requirements29:00 — Where to start for real AI ROI35:00 — What changes in the next 12 to 18 months⸻Resources MentionedSubatomic AI: https://getsubatomic.aiDeep Lens: Observability framework for AI workflowsNIST: Security and compliance frameworkOWASP: Application security frameworkISO 27001: Information security standard⸻Listen NowAvailable on all major podcast platforms and YouTube⸻Connect with the ShowFollow The CTO Show with Mehmet for more conversations at the intersection of technology, startups, and venture capital
Transcribe →