Earley AI Podcast

Moritz Müller, a leader in AI applications, discusses the future of information access with retrieval augmented generation and large language models. He highlights the importance of metadata, vector similarity searches, and supervised fine-tuning for accurate data retrieval. The podcast explores personalized content, semantic search engines, and enriching embeddings to enhance language model accuracy. It ends with insights on how retrieval augmented generation will revolutionize AI innovations in the coming years.

Erdem Özcan is an esteemed expert with a rich background in computer science, focusing on innovations in AI. With a PhD in computer science and significant industry experience, including work on IBM's Watson and at Elemental Cognition, Dr. Özcan has been at the forefront of blending symbolic AI and deep learning systems. Today, he is actively engaged in developing solutions that enhance the reliability and explainability of AI applications.Tune in to this enlightening conversation and gain deeper insights into the future trajectories and current challenges within the world of artificial intelligence as explained by one of the leading thinkers in the field.Key takeaways:- Symbolic vs. Statistical AI: Erdem discusses the critical differences and applications of symbolic AI versus statistical methods, emphasizing the need for reliably representing concepts for efficient AI outcomes.- The Role of Cogent English: Insight into how Cogent, a platform developed by Erdem, assists in translating complex business knowledge into APIs and conversational interfaces using a subset of English tailored for formal reasoning.- Challenges in Generative AI: Exploration of issues that arise with generative AI, particularly around reliability and the operational deployment of reasoning systems.- Development of Neurosymbolic AI: Erdem predicts a significant shift towards hybrid AI architectures that combine both symbolic and deep learning approaches to handle real-life complex scenarios more efficiently.- Importance of Explainability in AI: A discussion on why explainability and the ability to audit AI decisions are crucial, especially as AI systems become more integrated into critical decision-making processes.- Comparison of Formal Reasoning Systems and LLMs: Erdem explains why formal reasoning systems can be more reliable than large language models (LLMs) in complex problem-solving scenarios.Quote from the show:"Translating human expertise into AI systems is not just about feeding data into algorithms. It’s about creating structures that allow machines to reason and make decisions transparently and reliably." – Erdem ÖzcanLinks:LinkedIn: https://www.linkedin.com/in/aerdemozcan/Website: https://ec.ai/Ways to Tune In:Earley AI Podcast: https://www.earley.com/earley-ai-podcast-home Apple Podcast: https://podcasts.apple.com/podcast/id1586654770 Spotify: https://open.spotify.com/show/5nkcZvVYjHHj6wtBABqLbE?si=73cd5d5fc89f4781 iHeart Radio: https://www.iheart.com/podcast/269-earley-ai-podcast-87108370/ Stitcher: https://www.stitcher.com/show/earley-ai-podcast Amazon Music: https://music.amazon.com/podcasts/18524b67-09cf-433f-82db-07b6213ad3ba/earley-ai-podcast Buzzsprout: https://earleyai.buzzsprout.com/ Thanks to our sponsors: VKTR Earley Information Science AI Powered Enterprise Book

Lief Erickson brings expertise in technical writing and content strategy consultation. Having steered numerous organizations through the integration of AI and coherent information architecture, making the complex accessible. With a voice of authority in AI and content management the EIS Podcast is thrilled to have him on the show.Tune in to this episode for a comprehensive understanding of how structured content and precise prompt engineering are pivotal to leveraging AI in the realm of content creation and management.Key Takeaways:- Large language models (LLMs) require clear prompts and structured content to produce accurate and trustworthy responses.- The importance of structured content in enabling effective retrieval and utilization by generative AI, akin to finding a book in a library.- Misconceptions about generative AI’s capabilities in content management, highlighting the need for careful curation and validation.- Real-world applications of AI that can help increase brand loyalty, efficiency, reduce support calls, manage risk, and boost revenue.- The emerging role of prompt engineering and its significance in ensuring the relevance and accuracy of AI-generated content.- Legal and ethical considerations in using AI for content creation, with insights on copyright and the ownership issues surrounding machine-generated content.Quote of the Show:"Understanding structured content is like understanding the blueprint of a building—it's what allows us to scale and architect information in ways that align with our strategic goals." - Lief EricksonLinks:LinkedIn: https://www.linkedin.com/in/lief-erickson/Website: https://www.intuitivestack.io/Ways to Tune In:Earley AI Podcast: https://www.earley.com/earley-ai-podcast-home Apple Podcast: https://podcasts.apple.com/podcast/id1586654770 Spotify: https://open.spotify.com/show/5nkcZvVYjHHj6wtBABqLbE?si=73cd5d5fc89f4781 iHeart Radio: https://www.iheart.com/podcast/269-earley-ai-podcast-87108370/ Stitcher: https://www.stitcher.com/show/earley-ai-podcast Amazon Music: https://music.amazon.com/podcasts/18524b67-09cf-433f-82db-07b6213ad3ba/earley-ai-podcast Buzzsprout: https://earleyai.buzzsprout.com/ Thanks to our sponsors: VKTR Earley Information Science AI Powered Enterprise Book

Manish Sharma is the co-founder of Resolve AI. With a rich history spanning over two decades in the technology industry, Manish offers profound perspectives on the intersection of artificial intelligence, information architecture, and knowledge management.Tune in as Manish dissects popular AI myths, underscores the importance of bridging the technological gap, and emphasizes the need for robust security measures in today's digital landscape.Key takeaways:- When implementing AI solutions like large language models, CISOs should ask questions around data security, access controls, model guarantees, and emerging risks like prompt hacking to properly manage risks.- Information architecture is critical for data privacy, security, and ensuring AI systems can only access appropriate data sources and provide the right information to different user groups.- Retrieval augmented generation using a knowledge graph or index is important to avoid hallucinations and ensure AI systems can only respond based on curated data sources.- Scripted responses may be needed in some cases like legal to provide verbatim answers instead of generated responses.-  User personas and metadata are important to ensure AI systems understand the context and privileges of different user groups to provide appropriate and non-confusing information.- When integrating AI solutions with knowledge repositories like SharePoint, only curated subsets should be connected instead of entire repositories, and information should be properly tagged and structured.  Quote of the show:"A key to successful AI integration is not just in understanding the technology itself but in grasping the nuances of user needs, processes, content, and knowledge that remains timeless, no matter the advancements in tech. Coming to grips with that is where the real value lies." - Manish SharmaLinks:LinkedIn: https://www.linkedin.com/in/manish-sharma-rezolve/Website: https://www.rezolve.ai/Ways to Tune In:Earley AI Podcast: https://www.earley.com/earley-ai-podcast-home Apple Podcast: https://podcasts.apple.com/podcast/id1586654770 Spotify: https://open.spotify.com/show/5nkcZvVYjHHj6wtBABqLbE?si=73cd5d5fc89f4781 iHeart Radio: https://www.iheart.com/podcast/269-earley-ai-podcast-87108370/ Stitcher: https://www.stitcher.com/show/earley-ai-podcast Amazon Music: https://music.amazon.com/podcasts/18524b67-09cf-433f-82db-07b6213ad3ba/earley-ai-podcast Buzzsprout: https://earleyai.buzzsprout.com/ Thanks to our sponsors: VKTR Earley Information Science AI Powered Enterprise Book

Thomas Blumer is a renowned expert in AI-driven transformation with extensive experience in implementing groundbreaking artificial intelligence and knowledge strategies within complex business environments. Echoing a profound understanding of metrics-driven governance of AI systems, Thomas has made significant strides in aligning AI applications with overarching business goals. As a strategic advisor and consultant, he has facilitated diverse organizations in their journey to integrate AI to optimize efficiency, enhance user experiences, and drive actionable business outcomes. His expertise is instrumental in developing robust AI governance frameworks that ensure data, algorithms, and knowledge are in strict adherence to driving value and enterprise strategy.Key takeaways:- Defining and measuring KPIs tailored to customer and user lifecycle is crucial to drive business outcomes with AI and knowledge systems.- The transition from proof of concept to proof of value in AI implementations often encounters hurdles due to artificial environments and upstream data issues.- AI's implementation should focus on improving specific tasks and processes, ensuring tangible improvements rather than the technology's mere presence.- Storytelling and emotional resonance play a pivotal role when data alone does not suffice in persuading stakeholders about AI initiatives.- Governance structures need to strike a balance between centralized standards and decentralized, data-driven decision making.- Large language models have brought about a revolution in accessing corporate knowledge and productivity, highlighting the need for responsible usage.Quote of the show:"Bringing AI into the fold isn't just about technology; it’s about shaping an ecosystem that thrives on data integrity, governance, and context to create impactful narratives." - Thomas BlumerLinks:LinkedIn: https://www.linkedin.com/in/thomasblumer/Website: https://www.zyris.comWays to Tune In:Earley AI Podcast: https://www.earley.com/earley-ai-podcast-home Apple Podcast: https://podcasts.apple.com/podcast/id1586654770 Spotify: https://open.spotify.com/show/5nkcZvVYjHHj6wtBABqLbE?si=73cd5d5fc89f4781 iHeart Radio: https://www.iheart.com/podcast/269-earley-ai-podcast-87108370/ Stitcher: https://www.stitcher.com/show/earley-ai-podcast Amazon Music: https://music.amazon.com/podcasts/18524b67-09cf-433f-82db-07b6213ad3ba/earley-ai-podcast Buzzsprout: https://earleyai.buzzsprout.com/ Thanks to our sponsors: VKTR Earley Information Science AI Powered Enterprise Book

Trent Fitz holds over 20 years of experience in the tech industry. Currently a C-level Product Strategy and Technical Marketing Leader at Zenoss. He is an expert in global marketing, product strategy, business development in cloud computing, cybersecurity, and AI. Repeatedly proving his effectiveness in the industry by leading solutions to projects in innovative company’s such as IBM, Sailpoint, Trustwave and other various startups. Key takeaways:- APM tools such as Dynatrace, AppDynamics, and New Relic are key, and their integration has been aided by standards like open telemetry.- AI governance is crucial on technical, business process, and enterprise strategy levels.- The maturity models for AIOPs involve governance, decision making, and data/information architecture.- There is a general lack of appreciation for data and content within IT organizations.- AIOPs includes machine learning, and there's a need to educate about structured data and AI capabilities.Quote of the show:"At the core of AIOPs lies a fundamental need to not just visualize but truly understand the staggering complexity of modern IT environments. It's not just about piles of data or sophisticated algorithms; it's about cultivating a genuine appreciation for the significance of that data and how we can harness it to drive smarter, more proactive operations." — Trent FitzLinks:LinkedIn: https://www.linkedin.com/in/trent-fitz/Website: https://www.zenoss.comWays to Tune In:Earley AI Podcast: https://www.earley.com/earley-ai-podcast-home Apple Podcast: https://podcasts.apple.com/podcast/id1586654770 Spotify: https://open.spotify.com/show/5nkcZvVYjHHj6wtBABqLbE?si=73cd5d5fc89f4781 iHeart Radio: https://www.iheart.com/podcast/269-earley-ai-podcast-87108370/ Stitcher: https://www.stitcher.com/show/earley-ai-podcast Amazon Music: https://music.amazon.com/podcasts/18524b67-09cf-433f-82db-07b6213ad3ba/earley-ai-podcast Buzzsprout: https://earleyai.buzzsprout.com/ Thanks to our sponsors: VKTR Earley Information Science AI Powered Enterprise Book

Ian Hook is an exemplary professional whose journey spanned from an early career as a blacksmith and preschool teacher to becoming a seasoned expert in knowledge management and artificial intelligence (AI) at Nevartis. His unorthodox path and hands-on experience have endowed him with a deep understanding of the intricacies of knowledge management and its pivotal role in leveraging generative AI tools efficiently and effectively within operational teams. Ian's significant contributions have led to remarkable operational efficiencies, including an $18 million savings for his company by centralizing market research resources.Key Takeaways:- Knowledge management and generative AI are integral to improving the speed and accuracy of issue detection and remediation in operational teams.- Understanding the lineage and flow of data is vital for data scientists to fulfill their responsibility effectively.- Ian Hook illustrates the considerable impact of having a centralized knowledge management platform on efficiency and cost savings within a corporate setting.- The importance of governance in the context of utilizing generative AI is highlighted to mitigate unreliable outcomes due to ungoverned data.- Knowledge graphs are presented as sophisticated tools that visualize expertise and the relationships between different domains of knowledge.- The episode explores the limitations of large language models and emphasizes the importance of human oversight to prevent inaccuracies.Quote of the Show:"In our quest to harness AI, we must remember that the texture of human knowledge and expertise is the bedrock upon which these systems must be built." - Ian HookLinks:LinkedIn: https://www.linkedin.com/in/ianhook1/Website: Novartis.comWays to Tune In:Earley AI Podcast: https://www.earley.com/earley-ai-podcast-home Apple Podcast: https://podcasts.apple.com/podcast/id1586654770 Spotify: https://open.spotify.com/show/5nkcZvVYjHHj6wtBABqLbE?si=73cd5d5fc89f4781 iHeart Radio: https://www.iheart.com/podcast/269-earley-ai-podcast-87108370/ Stitcher: https://www.stitcher.com/show/earley-ai-podcast Amazon Music: https://music.amazon.com/podcasts/18524b67-09cf-433f-82db-07b6213ad3ba/earley-ai-podcast Buzzsprout: https://earleyai.buzzsprout.com/ Thanks to our sponsors: VKTR Earley Information Science AI Powered Enterprise Book

Mark Pickren currently serves as the President of Next Net Media. With over 25 years of experience as a seasoned entrepreneur and business leader, he possesses expertise in marketing-focused technology companies. Mark has demonstrated a consistent track record of building and managing successful ventures, with leadership experience spanning various industries, including Fintech, SaaS, and Digital Marketing. He has effectively overseen hundred-million-dollar P&Ls at large public corporations and Madison Avenue agencies. Remaining at the forefront of the dynamic digital landscape, Mark consistently delivers innovative solutions for consumers and businesses.Takeaways:Organizations need to prepare for around a 25% decline in organic search traffic as search becomes more personalized. Marketers need to focus on multi-dimensional targeting and providing value to specific customer personas to optimize content for search.As repetitive tasks are automated, career paths will focus more on managing autonomous agents and leveraging AI effectively.  Large language models pose risks if not properly overseen by humans, and differentiation requires responsible use of proprietary data and knowledge.Emerging technologies like retrieval-augmented generation will have major impacts on enterprises by improving information access.Quote of the Show:"Don't be a cynic. Lean into the better angels of technology, and be part of the solution." (Advice for graduates on how to approach emerging technologies.) - Marc PickrenLinks:LinkedIn: https://www.linkedin.com/in/marcpickren/Website: https://nextnetmedia.com/Marc's Website: https://www.marcpickren.com/Ways to Tune In:Earley AI Podcast: https://www.earley.com/earley-ai-podcast-home Apple Podcast: https://podcasts.apple.com/podcast/id1586654770 Spotify: https://open.spotify.com/show/5nkcZvVYjHHj6wtBABqLbE?si=73cd5d5fc89f4781 iHeart Radio: https://www.iheart.com/podcast/269-earley-ai-podcast-87108370/ Stitcher: https://www.stitcher.com/show/earley-ai-podcast Amazon Music: https://music.amazon.com/podcasts/18524b67-09cf-433f-82db-07b6213ad3ba/earley-ai-podcast Buzzsprout: https://earleyai.buzzsprout.com/ Thanks to our sponsors: VKTR Earley Information Science AI Powered Enterprise Book

Our guest this episode is Kristina Francis, a Executive Director at JFFLabs. Jobs for the Future (JFF) is a nationwide nonprofit dedicated to reshaping U.S. education and workforce systems for inclusive economic progress.Kristina is a experienced professional with a rich background spanning management consulting, software development, engineering, and cybersecurity. She began in database administration at the American Institutes for Research, evolving from an individual contributor to leading a 120-member development team for the Department of Defense. In 2016, a pivotal moment led to a dual career path, involving founding a consulting company, angel investing in women-owned tech ventures, and engaging in workforce opportunities. Currently serving as the Executive Director for JFFLabs at Jobs for the Future, Kristina  provides a distinctive perspective on the present and future of workforce and education, emphasizing innovation, disruption, and foresight into the implications of emerging technologies.Takeaways:AI has the potential to both disrupt jobs and create new job opportunities, but ensuring access to skills training will be important for workforce development.Personalized learning and career discovery tools that integrate assessments and map out skills pathways could help more people navigate changing job opportunities.Addressing systemic barriers and biases will be important to ensure all populations can benefit from new economic opportunities.Regions and employers can play a role in workforce development through public-private partnerships, on-the-job training programs, and investing in employees' skills.Quote of the Show:" How do we get more innovators, school systems, programs, and employers to get on board and provide the support and systems needed so that everyone in our communities is able to discover and navigate through our system to achieve their highest potential? "- Kristina FrancisLinks:LinkedIn:https://www.linkedin.com/in/kristinaharrisonfrancis/Website: https://www.jff.org/Email: KFrancis@JFF.orgWays to Tune In:Earley AI Podcast: https://www.earley.com/earley-ai-podcast-home Apple Podcast: https://podcasts.apple.com/podcast/id1586654770 Spotify: https://open.spotify.com/show/5nkcZvVYjHHj6wtBABqLbE?si=73cd5d5fc89f4781 iHeart Radio: https://www.iheart.com/podcast/269-earley-ai-podcast-87108370/ Stitcher: https://www.stitcher.com/show/earley-ai-podcast Amazon Music: https://music.amazon.com/podcasts/18524b67-09cf-433f-82db-07b6213ad3ba/earley-ai-podcast Buzzsprout: https://earleyai.buzzsprout.com/  Thanks to our sponsors: VKTR Earley Information Science AI Powered Enterprise Book

Our guest this episode is Alexander Schober, a data & AI project owner at Motius. He manages a diverse team of tech experts, focusing on Machine Learning, Knowledge Graphs, and Data Analysis. Alexander previously worked at Siemens Technology which involved pioneering research in Federated Learning and Self-Supervised Methods for anomaly detection. He used algorithms like Federated Averaging and SimCLR to address data privacy and label sparsity. Alexander joins Seth Earley and Chris Featherstone to the discuss knowledge graphs, metadata modeling for data engineering, using large language models to build data pipelines and more.For more content related to LLM's and Knowledge Graphs: https://www.earley.com/case-studies Takeaways:AI Enhancements with Knowledge Graphs: While not strictly required, knowledge graphs enhance the capabilities of AI, particularly large language models. The ability to provide context and resolve conflicts within the data contributes to more accurate and reliable AI outcomes.Unified Metadata Model: There's a need for a unified metadata model across different tools and platforms in the data engineering and AI landscape. Disjointed metadata tools can lead to inefficiencies, and efforts should be made to integrate and unify metadata for better collaboration.AI-Powered Data Pipeline Construction: Large language models can be used to generate data pipelines based on provided metadata. This approach can streamline the data engineering process, ensuring that quality checks, governance attributes, and privacy classifications are integrated into the pipeline.Quote of the Show:" All of these things are interconnected. Knowledge graphs, ontologies and semantics. They are all very important."                    - Alexander SchoberLinks:LinkedIn:https://www.linkedin.com/in/alexander-schober/Website: https://www.motius.comWays to Tune In:Earley AI Podcast: https://www.earley.com/earley-ai-podcast-home Apple Podcast: https://podcasts.apple.com/podcast/id1586654770 Spotify: https://open.spotify.com/show/5nkcZvVYjHHj6wtBABqLbE?si=73cd5d5fc89f4781 iHeart Radio: https://www.iheart.com/podcast/269-earley-ai-podcast-87108370/ Stitcher: https://www.stitcher.com/show/earley-ai-podcast Amazon Music: https://music.amazon.com/podcasts/18524b67-09cf-433f-82db-07b6213ad3ba/earley-ai-podcast Buzzsprout: https://earleyai.buzzsprout.com/ Thanks to our sponsors: VKTR Earley Information Science AI Powered Enterprise Book