Advertise on podcast: AI with Arun Show

Insta Service, an AI-powered platform founded by Venkatesh Regala, aims to modernize the home maintenance and repair industry by functioning as an "Uber for home services." Unlike legacy directory platforms such as Angie and Thumbtack, which often lead to endless searching and scheduling friction, Insta Service utilizes agentic booking flows and AI to instantly match homeowners with vetted, local professionals within a 30-minute radius. The platform categorizes jobs into regular maintenance with upfront, transparent pricing and larger projects that require in-person contractor quotes. By prioritizing supply liquidity across all 50 states and solving the chicken-and-egg dilemma of a two-sided marketplace, the company ensures high reliability and seamless transactions. Looking toward the future, the founders envision integrating smart home IoT sensors and full-home intelligence systems to automate preventative maintenance and create a fully autonomous living experience.

Modern software engineering faces a significant productivity imbalance as artificial intelligence coding throughput vastly outpaces traditional testing methods, creating a critical need for autonomous quality assurance agents. While advanced models generate thousands of lines of code daily, standard testing frameworks struggle with the resulting ecosystem orchestration gap and cannot independently verify comprehensive user experiences. To address this challenge, innovative solutions like the Ship platform utilize automated web crawling, bug reproduction via tools such as Slack, and pull request analysis to streamline software evaluation. Despite these autonomous capabilities, maintaining human oversight remains essential for establishing trust, ensuring compliance, and defining subjective standards of excellence. Ultimately, quality assurance professionals will transition into quality architects who direct artificial intelligence agents to secure reliable software development.

Professor Arunangu Chatterjee discusses the complexities of artificial intelligence adoption within large legacy institutions like universities and healthcare systems, emphasizing that success relies on human-driven value rather than just technological capability. He outlines a practical framework for managing institutional resistance by addressing value, capability, process, and risk, while noting that true transformation occurs when technology is fully integrated into everyday operations. The conversation also explores the ethical challenges of patient data sharing, proposing distributed value models to build trust and ensure accountability between humans and vendors. Furthermore, the dialogue examines how generative artificial intelligence impacts doctoral research and higher education, questioning the future role of human scholars and the evolving value of university degrees.

This interview features ServiceNow executive Sankha Mukherjee, who discusses the intersection of enterprise AI and personal financial management. Mukherjee explains how ServiceNow is evolving into an autonomous AI platform that prioritizes security, governance, and deterministic outcomes to eliminate operational bottlenecks like the traditional service desk. He also highlights his books, which advocate for applying an enterprise systems mindset to personal wealth by using automation and observability to optimize taxes and protect assets. The discussion emphasizes the importance of lifelong learning and the shift toward AI-driven workflows as essential skills for the modern workforce. Ultimately, the source promotes a philosophy of letting money learn and AI work to achieve financial independence while focusing on ethical leadership and philanthropy.

The podcast, features an interview with Surojit Chatterjee, the CEO of Ema, regarding the transition from basic chatbots to universal AI employees. Unlike personal productivity tools, these agents operate as a system of action that orchestrates complex, multi-step workflows across departments like HR and finance. The company utilizes a proprietary technology called Ema Fusion, which dynamically blends hundreds of large language models to maximize accuracy and cost-efficiency while remaining model-agnostic. Chatterjee emphasizes that these AI workers are designed to automate repetitive tasks, allowing human staff to focus on strategic oversight and management. Furthermore, the platform prioritizes enterprise-grade security by offering air-gapped deployment options and maintaining full auditability of every decision. Ultimately, the discussion suggests that the future of enterprise architecture relies on agentic reasoning to streamline fragmented business processes.

In a recent demonstration, UC Berkeley student Sainik Ghosh showcases the capabilities of Hermes Agent, an open-source tool designed to retain memory across different user sessions. Unlike standard chatbots that treat every interaction as a new encounter, this technology builds a personal profile and develops reusable skills to improve its performance over time. Ghosh illustrates the agent's versatility by using it to manage travel preferences and conduct automated resume screening for technical roles. While the system effectively filters data based on specific constraints, Ghosh warns of silent hallucinations where the AI provides incorrect information with high confidence. Ultimately, the source emphasizes that while this persistent reasoning is powerful, human oversight remains essential to verify the agent's autonomous decisions.

In this interview, Dr. Debashis Sahoo explains how his Boolean Lab at UC San Diego uses mathematical logic to revolutionize medical research. Unlike traditional AI and machine learning that rely on statistical averages and "black box" patterns, Sahoo’s approach utilizes Boolean rules to identify invariants, which are biological relationships that remain constant across data sets. This method has already yielded breakthroughs in colon cancer, where his team identified a specific gene that predicts survival rates, and in Alzheimer’s disease, where it uncovered a new mouse model for asymptomatic cases. By treating biological systems like hardware circuits, Sahoo aims to filter out messy data noise to find the fundamental laws of biology. Ultimately, the goal is to create trustworthy clinical impacts by moving medicine toward a more predictable, engineering-based framework.

This interview featuring Dr. Soumya Basu explores the evolving landscape of software engineering in an era increasingly dominated by artificial intelligence. Dr. Basu emphasizes that while tools like agentic AI and automated coding are advancing, the core principles of agility, adaptability, and flexibility remains essential. He advocates for a design thinking approach, urging practitioners to transition from being mere programmers to becoming holistic problem solvers who prioritize the user experience. The discussion clarifies that software engineering methods must be context-specific rather than "one-size-fits-all" frameworks. Ultimately, the source suggests that human-led innovation and organizational maturity are irreplaceable, even as AI serves as a powerful tool for intelligent automation

This interview features Dr. Sekhar Sarukkai, CEO of Chatsee AI, discussing why autonomous AI agents often experience "silent failures" in professional environments. Unlike traditional software that crashes, agents suffer from behavioral drift and unpredictability, making standard engineering metrics like uptime insufficient for monitoring. Sarukkai introduces the concept of failure memory, a system that categorizes 157 unique failure patterns to help developers move beyond the vague label of "hallucination." He argues that managing these agents requires a human-centric supervisory approach, treating them as employees who need continuous feedback rather than rigid guardrails. The conversation concludes by highlighting the shift from deterministic coding to probabilistic AI systems, emphasizing that future reliability depends on real-time production visibility and automated remediation.

This podcast from the AI with Arun Show examines the critical challenges and potential solutions for organ donation in India, where 500,000 lives are lost annually due to systemic inefficiencies. Guest Priti Agarwal highlights that while India possesses high surgical expertise, it suffers from a severe lack of donors, widespread misinformation, and significant logistical bottlenecks. The discussion emphasizes how technology and AI could revolutionize the process by optimizing organ matching, coordinating emergency transport "green corridors," and managing standardized kidney swap programs. To improve donation rates, the sources advocate for increased public awareness, policy shifts toward "opt-out" systems, and better integration of donation status into official identification like driving licenses. Ultimately, the text underscores that while medical infrastructure is strong, the primary hurdles remain family consent, data transparency, and the need for efficient coordination between hospitals and law enforcement.

The Aura platform, led by CEO Warren Myers, functions as an asset-light marketplace designed to revolutionize global emergency and security response. By utilizing an API-first strategy, the company connects individuals and businesses to the nearest vetted responders through a sophisticated AI routing engine that prioritizes response times. The technology addresses the growing gap between public police resources and increasing security needs, particularly in regions like Africa, the UK, and the United States. Aura integrates with various IoT devices and wearables to enable autonomous emergency triggers, such as crash or fall detection, ensuring help arrives even when a user cannot call for it. Beyond simple dispatch, the platform aims to act as a global clearinghouse for emergency signals, streamlining the triage process to help traditional law enforcement focus on high-risk incidents. Ultimately, the company’s mission is centered on democratizing safety and tracking success through the primary metric of lives saved.

The podcast features an interview with Dr. Satya Nitta, the CEO of Emergence AI, who discusses the evolution of autonomous enterprise agents. Nitta explains that true autonomy involves decision-making under uncertainty, contrasting sophisticated agentic systems with simple, rule-based automation. A central theme is the importance of safety and determinism, where Nitta introduces neuro-formal AI to mathematically verify that agents follow specific rules. He highlights a study on normative drift, demonstrating how even well-behaved models can adopt harmful tactics when interacting in multi-agent environments. Finally, Nitta addresses the future of the industry, touching on liability, the shift toward value-based pricing, and the changing role of software engineers.

In this interview, Hitachi Vantara’s Chief Strategy Officer, Simon Ninan, outlines a blueprint for the data center of the future by prioritizing outcome-driven AI strategies over wasteful overbuilding. He argues that enterprises must shift from hoarding "digital exhaust" to adopting a data-as-a-product mindset, emphasizing that high-quality, governed data is the only way to escape "pilot purgatory." To manage soaring energy demands, Ninan introduces the concept of grid-aware infrastructure, where software policies help schedule workloads based on real-time power availability and costs. He further discusses the rise of sovereign AI, predicting that global firms must adopt regional, hybrid architectures to comply with local data laws. Ultimately, Ninan contends that sustainability and operational efficiency are not mere ethical choices but are essential for long-term business survival and scalable innovation.

The podcast features an interview with Rich Delisser, Cisco's VP of Engineering, regarding the massive transition of 21,000 developers into an AI-native future. Delisser explains that software engineering is shifting from manual coding to human-led orchestration, where developers act as managers of AI agents rather than just authors of syntax. To facilitate this, Cisco has implemented a technical harness to ensure code quality and security while focusing heavily on organizational culture and trust to prevent employee alienation. The discussion highlights that while AI serves as a productivity multiplier, essential human traits like accountability, empathy, and judgment remain vital for high-quality results. Ultimately, the source frames AI integration not as a replacement for staff, but as a necessary evolution to increase operational speed and overcome traditional scaling bottlenecks.

The podcast features a demonstration of ContextQA, a specialized platform designed to validate and evaluate AI agents before they are released into production. The tool allows developers to upload agent documentation to automatically generate comprehensive test cases and user personas, ensuring the AI handles various scenarios like order management or customer support effectively. By simulating diverse user behaviors, the software tests for intent recognition, task completion, and response relevance across multiple platforms such as AWS Bedrock and Salesforce. Additionally, ContextQA incorporates red teaming and load testing to identify security vulnerabilities, such as prompt injections, and assess performance under high traffic. Users receive detailed analytical reports and executive summaries that highlight failures and hallucinations, providing a data-driven approach to refining agent behavior. Ultimately, the source emphasizes that rigorous testing and security integration are essential steps to prevent AI agents from malfunctioning in front of real users. YT Link - https://youtu.be/Lfak2tp_WQ8