How Companies and Data Scientists Can Capitalize on the AI Boom (Vin Vashishta) - KNN Ep. 165
Aug 30, 2023
auto_awesome
Vin Vashishta, author of 'From Data To Profit' and founder of V-Squared, discusses the AI hype, the impact of AI on job roles, the importance of intuition in coding, potential dangers of AI models in marketplaces, and the need for continuous education in data science and AI.
Generative AI should be understood in a nuanced way, considering the need for rewriting and avoiding obsolescence in business and writing.
Generative AI serves as an interface to access capabilities and knowledge bases, enhancing customer experiences.
Unintentional destabilization can arise from models with accuracy and reliability, emphasizing the importance of preparing for disruptions caused by individuals with access to powerful models.
Deep dives
Implications of Generative AI in Business and Writing
Generative AI presents both opportunities and challenges in business and writing. The podcast episode discusses how the author of a book had to rewrite it with generative AI in mind, considering the fast pace of progress and the need to avoid obsolescence upon release. It also emphasizes the importance of differentiating ourselves from routine and mundane tasks that can be automated. The focus should shift towards the value of data and our role as creators and generators of high-quality data. Additionally, the conversation touches on how generative AI is an enabling technology, acting as an interface that provides access to various functionality, knowledge graphs, and complex task flows. It highlights the need for businesses to adopt a maturity model approach, delivering incremental value and utilizing cheaper technologies before deploying more expensive ones.
The Nuance of Generative AI and its Impact on Jobs and Decision Making
Generative AI, such as chat GPT and large language models, has attracted significant hype but needs to be understood in a more nuanced way. While it has its benefits in augmenting certain job roles, particularly those requiring higher-end skills that justify the cost, it does not replace the need for human intuition and decision-making. The podcast episode explores how generative AI serves as an interface to access capabilities and knowledge bases. It has the potential to simplify complex application suites and enable natural language interactions. However, the true value lies in coupling generative AI with decision-focused datasets to enhance customer experiences and empower non-technical users with more advanced self-service tools. It also highlights the importance of individuals and domain experts who possess irreplaceable knowledge and engage in complex decision-making, emphasizing their significance in surviving disruption.
The Dangers of Unintentional Data-Driven Destabilization
Beyond the technological advancements facilitated by generative AI, the podcast episode delves into potential risks and threats. It discusses that unintentional destabilization can arise from models that possess accuracy and reliability without malicious intent. Such models, when adopted by individuals or small groups, can inadvertently disrupt marketplaces, finance, and societal systems. The episode emphasizes the vulnerability of these complex systems to models that grasp the rules governing them. It warns about the potential upheavals caused by sufficient knowledge and community support, outlining examples like Wall Street Bets and their impact on hedge funds as a glimpse into market destabilization possibilities. It calls for vigilance and highlights the importance of preparing for disruptions caused by individuals with access to powerful models, even if not intentionally malicious.
The Growing Importance of Data in AI Development
The podcast episode discusses the crucial role of high-quality data in the development and deployment of AI models. It emphasizes that starting with a better dataset leads to smaller and more efficient models that require fewer resources to train and achieve the same level of reliability. It highlights the shift from companies attempting to build their own models from scratch to leveraging foundational models provided by other companies. The competitive advantage lies in the unique access to data and the ability to curate high-quality data sets that lead to the creation of models that are unmatched by competitors.
The Challenges and Liabilities of Outsourcing to Large Language Models
The episode explores the drawbacks of outsourcing to large language models, particularly open-source models, and the potential legal liabilities associated with them. It recognizes the value of open-source models and the need to continue advancing open-source AI. However, it also highlights the importance of protecting businesses from potential liabilities arising from using these models. It raises concerns about the lack of liability protection for both open-source creators and companies integrating these models into their products. The discussion underscores the need for frameworks that balance open-source innovation while safeguarding companies and individuals from legal repercussions.
Today I had the pleasure of interviewing Vin Vashishta. Vin is the author of the Wiley book, ‘From Data To Profit.’ It’s the playbook for monetizing data and AI. Vin is the Founder of V-Squared and built the business from client 1 to one of the world’s oldest data and AI consulting firms. His background combines over 25 years in strategy, leadership, software engineering, and applied machine learning. Today Vin gives his take on the AI hype, what makes us uniquely human, and how companies can with with strategic changes around new large language models. This is a great conversation that can benefit data leaders and data scientist a like. Enjoy!
Podcast Sponsors, Affiliates, and Partners: - Pathrise - http://pathrise.com/KenJee | Career mentorship for job applicants (Free till you land a job) - Taro - http://jointaro.com/r/kenj308 (20% discount) | Career mentorship if you already have a job - 365 Data Science (57% discount) - https://365datascience.pxf.io/P0jbBY | Learn data science today - Interview Query (10% discount) - https://www.interviewquery.com/?ref=kenjee | Interview prep questions
