So What About AI Agents

KeywordsAI projects, hype vs reality, return on investment, success metrics, narrow use casesSummaryIn this conversation, Ammar Bhaisaheb discusses the critical aspects of initiating AI projects, emphasizing the importance of distinguishing between hype and reality. He advocates for a focused approach, where projects are defined by narrow use cases and measurable goals. The conversation highlights the necessity of modeling return on investment before embarking on any AI initiative, ensuring that projects are viable and worth the investment of time and resources.TakeawaysThe very first chat with clients is to separate hype versus reality.AI projects should be approached with a scalpel, not a will.Success in AI requires a narrow, super use case.Clear goals must be measurable for AI projects.Modeling return on investment is crucial before coding.Projects should aim for a three to five X return on investment.If a project can't deliver ROI, it's not worth pursuing.Focus on measurable outcomes to ensure project success.Narrow use cases lead to better-defined AI projects.Invest time only in projects that meet ROI criteria.TitlesNavigating the AI Hype: A Practical ApproachThe Scalpel Method: Precision in AI ProjectsSound bites"Separate the hype versus reality in AI.""Pick a narrow, super use case.""Clear goal that you can actually measure."Chapters00:00 Introduction to AI Agents and Their Impact00:07 Choosing the Right Use Cases for AI Agents00:43 Data: The Foundation of AI Success

John Brazier spent 25 years leading Oracle's technical documentation team, managing 394 books translated into 10+ languages. Today we discuss why AI-generated documentation fails at enterprise scale, the $12M translation problem, and why technical writers are evolving into "Knowledge Orchestrators."🔑 Key Topics:- Why enterprise documentation requires human wisdom, not just AI knowledge- The $500K cost of translating a single 500,000-word manual- How one comma can trigger a $40,000 retranslation- Revenue recognition regulations that make AI documentation illegal- The future role: Knowledge Orchestrator vs Technical Writer📊 Quick Stats:- 394 books in Oracle's documentation library- $12M initial translation cost for 6-product logistics suite- 500,000 translatable words per book- Translation costs: $0.06-$0.30 per word⏱️ Timestamps:00:00 Introduction02:28 The Scale of Documentation at Oracle04:41 Translation Challenges ($12M Problem)09:29 Knowledge vs Wisdom in Technical Writing16:36 The Future of Technical Writers31:39 Why Human Oversight Is Non-Negotiable#TechnicalWriting #AI #Documentation #EnterpriseAI #FutureOfWork

KeywordsAI, government contracts, agentic AI, compliance, certifications, small business, technology, security, federal government, entrepreneurshipSummaryIn this conversation, Philippe Trounev and Cordell Robinson discuss the burgeoning field of agentic AI and the significant investment the US government is making in this area. They explore the steps entrepreneurs need to take to sell AI agents to the government, including navigating the registration process, understanding compliance and certification requirements, and the lifecycle of government contracts. Cordell shares insights on the types of AI solutions in demand, the importance of security, and the advantages of partnering with established firms to access government contracts. The discussion emphasizes the potential for small businesses to thrive in this niche market as the government increasingly seeks innovative AI solutions.TakeawaysThe US government is heavily investing in agentic AI.Entrepreneurs must register their businesses on sam.gov to sell to the government.Compliance with NIST 800-171 is crucial for DoD contracts.The certification process can take several months, so start early.Government contracts often require a competitive bidding process.Selling existing products may involve customization for government needs.Security audits are essential for government contracts.Partnerships can enhance credibility and access to government opportunities.There is a growing demand for AI solutions in government, especially in security.Small businesses have a unique opportunity to enter the government contracting space.TitlesUnlocking Government Contracts for AI SolutionsNavigating the Complex World of Government ContractingSound bites"You have to register your business on sam.gov.""You can bid on contracts for the government.""NIST 800171 is your in-house security audit."Chapters00:00 The Rise of Agentic AI in Government02:56 Navigating Government Contracts: The Basics05:17 Compliance and Certifications for Small Businesses08:17 Bidding on Government Contracts: The Process11:00 Partnerships and Accelerating Government Sales13:43 Financial Aspects of Government Contracts16:43 Security Requirements for Government Products19:29 The Advantages of Selling to the Government22:02 Identifying High-Demand Use Cases24:50 Cost and Competition in the Government Contract Space

KeywordsAI, video analytics, security, Vault AI, incident detection, privacy, technology, company culture, future of AI, Igor OlteanuSummaryIn this episode, Igor Olteanu, co-founder of Vault AI, discusses the transformative impact of AI-powered video analytics on physical security. He shares his journey from military service to tech innovation, the technology behind Vault AI's solutions, and the importance of real-time incident detection. Igor addresses privacy concerns, highlights the effectiveness of their technology, and shares notable success stories. He also reflects on the cultural insights gained from his time at Google and how they influence the company culture at Vault AI. Finally, Igor offers his thoughts on the future of AI in security and encourages listeners to embrace innovation.TakeawaysIgor transitioned from military to tech due to a passion for innovation.Vault AI focuses on real-time incident detection in educational spaces.The technology uses geolocation and 3D mapping for accuracy.AI validation is complemented by human oversight to reduce false positives.Privacy concerns are addressed by clarifying existing camera usage.The effectiveness of AI allows for 100% camera monitoring.Notable incidents include detecting medical emergencies and bullying.Company culture emphasizes direct communication and innovation.Igor's military background aids in understanding security needs.The future of AI in security will see increased adoption and trust.TitlesTransforming Security with AI: Igor Olteanu's JourneyReal-Time Threat Detection: The Future of SafetyChapters00:00 Introduction to Vault AI and Igor's Journey02:49 The Technology Behind Vault AI's Video Analytics05:12 Real-Time Incident Detection and Response08:02 Privacy Concerns and Customer Objections10:50 Effectiveness of AI in Security13:41 Notable Incidents and Success Stories16:47 Cultural Insights from Google to Vault AI19:24 Building a Strong Company Culture22:00 The Future of AI in Security24:47 Final Thoughts and Encouragement

KeywordsAI agents, agentic society, culture, ethics, workforce, education, technology, future generations, automation, human-AI interactionSummaryIn this conversation, Philippe Trounev and Duri Chitayat explore the transformative impact of AI agents on society, culture, and the workforce. Duri shares his insights on how AI can be integrated into company culture, the challenges of teaching AI about human culture, and the implications for future generations growing up in an AI-driven world. They discuss the ethical responsibilities of AI developers and the need for a robust educational framework to prepare for a future where AI plays a central role in our lives. The conversation emphasizes the importance of trust, ethics, and adaptability in navigating the evolving landscape of technology and society.TakeawaysDuri Chitayat emphasizes the need for AI to adapt to human culture.AI agents can fundamentally change business models and operations.Understanding cultural nuances is crucial for effective AI integration.The future workforce will require new skills and adaptability due to AI.Ethics in AI development is essential to prevent negative societal impacts.Trust will become a key factor in human-AI interactions.Education systems must evolve to prepare children for an AI-driven world.Diversity of skills will be important for future job security.Cognitive decline can occur if we rely too heavily on AI.A balance between technology use and human interaction is necessary.TitlesNavigating the Agentic RevolutionAI Agents: Redefining Business and CultureSound bites"Agents must adapt to our culture""We might lose autonomy with AI tools""Prepare for a genetic society"Chapters00:00 Introduction to Agentic Society02:50 Cultural Integration of AI Agents05:25 The Role of AI in Corporate Culture08:12 AI and the Next Generation10:37 Future Workforce and Education13:36 Ethics in AI Development16:04 Preparing for a Genetic Society

SummaryIn this episode, Philippe Trounev and Grant McCracken discuss the evolving landscape of cybersecurity, particularly focusing on the role of agentics and AI in enhancing security measures. They explore the implications of new attack vectors like prompt injection, the potential for agent-to-agent attacks, and the innovative approach of agentic pen testing. Grant shares insights on how AI can improve the efficiency and effectiveness of security testing while emphasizing the importance of foundational cybersecurity practices. The conversation highlights the balance between leveraging advanced technologies and maintaining robust security protocols.TakeawaysCybersecurity is evolving with the introduction of agentics.Fundamentals of cybersecurity remain crucial despite new technologies.68% of data breaches originate from human error.Prompt injection is a significant new attack vector.AI can identify existing vulnerabilities but not create new ones.Agentic pen testing offers continuous, fatigue-free testing.Contextual vulnerabilities are better identified by AI.AI pen testing could reduce costs compared to human testers.Dark Horse aims to make cybersecurity solutions accessible.AI compute costs are high but expected to decrease over time.TitlesThe Future of Cybersecurity: Agentics and AIUnderstanding Cybersecurity in the Age of AISound bites"68% of data breaches start at the human layer.""Prompt injection is a new attack vector.""Agentic pen testing can run 24/7 without fatigue."Chapters00:00 Introduction to Cybersecurity and AI Agents01:42 Changing Attack Vectors in Cybersecurity03:44 Human Layer Vulnerabilities and AI's Role07:13 Prompt Injection and Offensive Strategies11:26 Agentic Systems and Security Risks12:58 Agent-to-Agent Attacks and Communication Risks17:03 Agentic Pen Testing: A New Approach21:33 Cost Efficiency of Agentic Pen Testing24:48 Exciting Developments in AI Pen Testing27:53 Future of Agentic Pen Testing and Market Potential

SummaryIn this episode, Philippe Trounev interviews Bobbie Chen, a product manager at Stytch, about the challenges and solutions related to AI agents, particularly in the context of security and fraud prevention. They discuss the concept of agentic experience, the importance of effective authentication methods, and the need for businesses to protect themselves from bad bots. Bobby shares insights on implementing agentic solutions, navigating compliance issues, and the future of fraud prevention in an increasingly automated world.TakeawaysAI agents have unique needs that differ from human users.Agentic experience is crucial for effective interaction with AI.Security measures must evolve to protect against bad bots.Device fingerprinting is a key tool in identifying fraudulent activity.CAPTCHAs are becoming less effective against sophisticated attacks.Shadow banning can be an effective strategy against spammers.Integrating agentic solutions can be quick and efficient.Compliance with regulations like GDPR is essential for businesses.The future of fraud prevention will require constant adaptation.Understanding attack vectors is vital for effective security measures.TitlesNavigating the World of AI AgentsThe Importance of Agentic ExperienceChapters00:00 Introduction to AI Agents and Security Challenges02:51 Understanding Agent Experience05:24 Agent Authentication and Security Risks08:03 Identifying Good and Bad Agents10:38 Defending Against Malicious Bots13:22 Real-World Impacts of Agentic Attacks16:07 Implementing Agentic Solutions18:52 Future of Fraud Prevention and Agentic Technology21:47 Navigating Privacy and Compliance Issues24:29 Final Thoughts on AI Agents and Security

KeywordsAI, drug discovery, Pauling.AI, language models, FDA approvals, automation, inhibitors, in silico experiments, agentic prompts, healthcareSummaryIn this conversation, Javier Tordable from Pauling.AI discusses the innovative approach to drug discovery using AI and language models. He explains the mission behind the company, the challenges of FDA approvals, and the automation of literature reviews and simulations. The discussion also covers the importance of inhibitors in drug discovery, the differences between in silico and wet lab experiments, and the need for adaptability in AI models. Tordable emphasizes the significance of the mission in improving healthcare and the potential for partnerships in the drug discovery process.TakeawaysPauling.AI aims to shorten drug discovery time significantly.AI-generated drugs follow the same FDA approval process as traditional drugs.Language models excel at reviewing and summarizing prior research.Automation of initial simulations can save weeks of work.Most drugs developed are inhibitors, which block specific biological processes.In silico experiments are a focus for Pauling.AI, differentiating from wet lab experiments.The risk of hallucinations in AI requires careful management in drug discovery.Adapting to rapid changes in AI models is crucial for success.Human-initiated interactions guide the AI agents' processes.The mission of drug discovery is to improve lives, not just profit.Sound bites"Most drugs are inhibitors.""We focus on in silico experiments.""Human-initiated interaction is key."Chapters00:00 Introduction to Agentic Drug Discovery02:15 The Role of AI in Drug Discovery04:55 Current State of AI-Driven Drug Development07:25 Challenges in Drug Discovery and AI Integration10:04 Optimizing AI Agents for Drug Discovery12:53 Human-AI Collaboration in Drug Discovery15:39 Future of AI in Drug Discovery18:13 Insights and Best Practices for Building AI Agents21:10 The Economics of Drug Discovery23:41 Conclusion and Future Directions

KeywordsAI news, investment, governance, AI companions, market dynamics, technology trends, AI deployment, valuations, regulatory landscape, business intelligenceSummaryIn this episode of So What About AI, Philippe Trounev discusses the latest developments in AI news, focusing on significant investments, the challenges of AI deployment, the rise of AI companions, and the shifting focus of venture capitalists from artificial general intelligence (AGI) to return on investment (ROI). The conversation highlights the complexities of AI governance, the societal implications of AI companionship, and the regulatory landscape affecting AI technologies.TakeawaysCognition AI raised $400 million at a $10.2 billion valuation.95% of enterprise AI pilots fail, indicating a significant challenge in deployment.97% of AI deployments globally lack proper governance.AI companions are becoming a concern for governments due to their increasing prevalence.Venture capitalists are shifting their focus from AGI to ROI, emphasizing profitability.Companies are increasingly buying AI solutions rather than building them in-house.The AI market is experiencing a peak hype phase, with significant investments flowing in.OpenAI's job platform aims to teach humans to work alongside AI.Mistral's $1.3 billion investment reflects the growing intersection of hardware and AI.The regulatory landscape is becoming more complex, particularly in the EU.Sound bites"OpenAI launches a job platform.""95% of enterprise AI pilots fail.""AI agents are at peak hype."Chapters00:00 AI News Overview04:50 Governance and Regulation in AI09:32 Data Breaches and Security Concerns

keywordsAI agents, enterprise implementation, pharmaceutical sector, SAP, ChatGPT, employee training, AI adoption, technology consulting, Infosys, automationsummaryIn this conversation, Philippe Trounev interviews Patrick Rayes, a senior principal at Infosys Consulting, about the implementation of AI agents in enterprise settings, particularly in the pharmaceutical sector. They discuss the challenges faced during AI adoption, the importance of human intervention, and the approval processes required in regulated industries. Patrick shares insights on employee training and the top-down approach necessary for successful AI integration, as well as managing expectations and misconceptions about AI capabilities.takeawaysAI implementation in enterprises requires a top-down approach.Human intervention is still necessary in AI processes.Approval processes in regulated industries are crucial for AI adoption.Employee training and education are key to successful AI integration.AI agents can simplify complex workflows and processes.Managing expectations about AI capabilities is essential for stakeholders.The pharmaceutical sector has specific requirements for AI implementation.A centralized interface enhances user experience with AI tools.Consulting firms play a vital role in guiding AI adoption.Companies must adopt AI with the same urgency as the internet.titlesNavigating AI Implementation in EnterprisesThe Role of Human Intervention in AISound Bites"AI is getting big right now.""Human intervention is still required.""AI agents simplify complex processes."Chapters00:00Introduction to AI Agents in Enterprises02:19AI Implementation in Pharmaceuticals and Banking04:43Challenges and Lessons Learned in AI Adoption07:01Human Intervention in AI Processes09:37Real-World Examples of AI Implementation11:02Approval Processes for AI in Regulated Industries12:45Comparing AI Tools: SAP Jewel, ChatGPT, and Gemini15:21Employee Adoption and Training for AI Systems18:03Impact of AI on Workflows and Efficiency20:41Managing Expectations Around AI Capabilities23:24Consulting Strategies for AI Transformation26:10Safeguarding Proprietary Knowledge in AI28:56Key Takeaways for Successful AI Implementation