Episodes

  • How to Build AI Responsibly | David Adkins Meta AI

    How to Build AI Responsibly | David Adkins Meta AI
    Oct 5 2023

    David Adkins is an experienced senior technology executive who leads engineering teams at Meta AI. David holds an MS in Computer Science from the University at Buffalo focused on machine learning. In this conversation we discuss

    * How to think about AI bias and fairness

    * Why AI Transparency, Explainability and Control are important

    * How Meta de-risked LLaMA

    * How to take AI Research to production and more.

    Listen now on Apple, Spotify and Amazon Music and please subscribe to my YouTube channel.

    ⏰ 2:06 David’s Favorite Restaurant Sammy’s Fishbox

    ⏰ 2:40 Career Stories: David’s Journey

    ⏰ 4:37 What David is working on at Meta AI

    ⏰ 4:50 How to organize AI Teams

    ⏰ 6:56 What is AI Bias and Fairness

    ⏰ 8:52 How to build a Product with AI Fairness in mind

    ⏰ 11:47 Examples of Fairness Mitigations in Meta Products

    ⏰ 13:33 What’s Surprising about working on AI Fairness

    ⏰ 16:05 What causes Bias in AI Products

    ⏰ 19:19 How to Mitigate AI Problems when you’ve already launched

    ⏰ 21:38 What is AI Transparency and Control

    ⏰ 24:12 What is AI Explainability?

    ⏰ 26:10 Why YOU should care about AI Transparency

    ⏰ 31:00 What is surprising about working on AI Transparency

    ⏰ 33:00 AI System Cards

    ⏰ 37:30 Developing LLaMA Responsibly

    ⏰ 39:30 How to de-risk large language models

    ⏰ 43:06 How to do AI Research to production

    ⏰ 46:51 What’s next for David



    This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aistudios.substack.com
    Show More Show Less
    49 mins
  • AI Opportunities in Healthcare with Javier Tordable, Technical Director at Google

    AI Opportunities in Healthcare with Javier Tordable, Technical Director at Google
    Aug 29 2023
    Javier Tordable is a Technical Director at Google where he drives long term technical strategy for Google Cloud at the CTO Office. Recently Javier has been focusing on healthcare, pharma and biotech, and helping organizations use Cloud infrastructure, Machine Learning and generative AI to improve drug discovery and patient care. Javier is also executive sponsor of top Google Cloud customers and advisor to C-suite executives; and helped close over a billion dollars worth of Cloud contracts.In this conversation we discuss * AI opportunities in healthcare, * Why AlphaFold is such a big breakthrough * Longevity (including Javier’s Longevity Protocol) and more.Listen now on Apple, Spotify and Amazon Music and please subscribe to my YouTube channel.⏰ 1:50 Javier’s background ⏰ 5:30 Javier’s role at Google ⏰ 9:05 How Javier approaches Strategy ⏰ 15:17 Tech and AI adoption in Healthcare ⏰ 17:10 Understanding Organizational Incentives in Healthcare ⏰ 18:00 Tools that can overcome Tech Limitations in Healthcare ⏰ 19:00 AI Opportunities and Use Cases in Healthcare ⏰ 23:20 AI Privacy for healthcare ⏰ 29:38 Why Alphafold is a big deal ⏰ 31:16 What Alphafold does ⏰ 32:36 How Alphafold will change healthcare ⏰ 35:55 The big Problem that Alphafold will solve ⏰ 36:43 Longevity ⏰ 44:25 Javier’s Longevity Hacks ⏰ 46:58 How to get into healthcare if you’re a techie and how to get into tech if you’re in healthcare? ⏰ 49:44 Javier’s advice to someone starting out ⏰ 52:12 what’s next for JavierWhere to find Javier • LinkedIn • Web• Javier on TwitterWhere to find Natalia • AI Studios on YouTube • Natalia’s Substack• Natalia on Twitter• LinkedIn This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aistudios.substack.com
    Show More Show Less
    56 mins
  • Leading AI Teams | Sonny Patel Product & Engineering Exec (LivePerson, Amazon, Microsoft)

    Leading AI Teams | Sonny Patel Product & Engineering Exec (LivePerson, Amazon, Microsoft)
    Jun 20 2023
    Sonny Patel is a tech executive who ran a team of 250 product managers and engineers at LivePerson. Before that, she ran a cross-functional org of designers, product managers, technical program managers and developers at Amazon. Sonny was also at Microsoft where she grew from an entry level product manager to a leader of leaders.Here's a written summary of our conversation. You can also listen to the audio version via SubStack, Apple and Spotify. Or you can check out the convo on YouTube.  Rising to and Operating at Executive LevelCareer Breakthroughs I asked Sonny what was the pivotal moment that put her on the executive path. Sonny recounted how a VP at Amazon took a bet on her by entrusting her with a cross functional team. This expanded her scope significantly and set her up for bigger opportunities. Sonny attributes this breakthrough to both leadership support and a fortuitous situation. However, to position herself Sonny offered the following playbook. Playbook📖 Develop your knowledge and skills over years. No shortcuts here.🏆 Champion - You need a champion to take a bet on you. Your knowledge and skills will enable your champion to stand behind you.📍Situational Awareness - you need to recognize and capitalize on a unique opportunity to distinguish yourself. In Sonny’s case it was about saying yes to managing an engineering team in addition to product.🎗Support System - You can’t make it without a professional support system.Value of Good and Bad Managers in early careerSonny believes that there is value in having one good and one bad manager early on.  Watching and learning from other people’s management mistakes is a good way to build empathy for future reports. A good manager is able to provide psychological safety for their reports. This improves performance and sets people up to do their very best work. Sonny’s own experience with a bad manager taught her to have empathy and cultivate patience towards her reports. When people are new to product management or the AI space, managers should be especially patient. How Sonny creates Psychological Safety for her Teams🙋🏻‍♀️ Encourage and support people to ask silly questions 🙊 Allow people to make mistakes and learn without fear. ✔️ Sonny used regular check-ins and reporting mechanisms to monitor team progress and identify issues earlyBuilding AI Products and Running AI TeamsHow AI products are different 🪩 People tend to get enamored with the latest shiny technology. Sonny emphasized the importance of focusing on usefulness and not just the "cool" factor. AI Products should solve real problems for users in meaningful ways.🔐 Privacy, Transparency and Control are critical. Users are willing to share data when they see a benefit and feel in control. Apply the idea of a privacy transaction when building products - if a product collects users data, the user should get something in return. Users should feel in control and everything should be done with their consent. Provide user control options in a coherent way that all fits together. Why most AI products fail AI products often fail due to edge cases that were not considered during design and testing. User expectations are often higher than what the technology can reliably deliver.What makes Amazon an efficient execution machineBefore building a tech product, start with the customer and work backwards by understanding their problem. Amazon believes in the power of writing down things. Write a Press Release to imagine what your product unveiling may look like including all the related messaging. A Press Release is a one-pager that anybody should be able to read and understand. Some of the questions that a Press Release addresses are:* What is the customer problem?* Who is the target customer?* Why is the idea big enough?* Why now?* What does the product development team say to customers? * What would your customers say after using that product or feature?* How is this overall fitting into your existing product strategy? Furthermore a Press Release includes how the customers can get started, what they need to do, any associated costs, configuration experience, etc. After this, the team starts to dive deeper in terms of thinking about the product design aspect. Sonny’s favorite aspects of the process are two things, Tenets and Rude Questions FAQ. The Power of Tenets Tenets are a set of principles around decision-making criteria. Having a clear set of tenets is useful for breaking debates during product design. Tenets define what is important in terms of trade offs. For example, sacrificing complex additional functionality in favor of simple and intuitive design for a non-tech audience. This is a potential debate that the product team could have. If the team was to make a trade-off, which side would they pick over the other? That's a great tenet. Definition of tenets requires a lot of thought. Why you Need a Rude Questions FAQ for your ProductA...
    Show More Show Less
    1 hr and 1 min
  • Amy Karle | Exploring how Technology is Changing Humanity

    Amy Karle | Exploring how Technology is Changing Humanity
    May 27 2023
    Groundbreaking artist and visionary futurist Amy Karle specializes in the transformative impact of emerging technologies on humanity, including AI and biotech. Her work examines how interventions could alter the trajectory of the future and how technology could be utilized to support and enhance our future. In this episode we discuss AI from an artists perspective, how AI will change the way we think and function, Amy’s vision for future with AI and feast our eyes on Amy’s work. Listen now on Apple and Spotify.⏰ 00:34 Amy's Journey ⏰ 02:17 AI from an Artists Perspective ⏰ 03:19 How AI is changing the way we Think and Function ⏰ 7:44 AI Dangers ⏰ 11:53 AI Opportunities ⏰ 13:12 AI and Human Mortality ⏰ 15:58 Biofeedback Work ⏰ 21:55 Amy's Approach to exploring Technology's Impact on Humanity ⏰ 24:15 How AI will Change what it means to be human ⏰ 26:15 How AI will enhance our lives ⏰ 27:18 Explorations of AI and Humans Merging ⏰ 29:47 Tools that Amy Uses to create her art ⏰ 30:36 Regenerative Reliquary Piece ⏰ 33:36 The Heart of Evolution Piece ⏰ 36:47 Feasibility of Organ Plug and Play ⏰ 38:13 What's next for AmyWhere to find Amy* Amy’s Website* Instagram* Twitter* Facebook* LinkedIn* Discord* WikipediaWhere to find Natalia* 🆕 Maven Generative AI for Business Workshop* YouTube* Substack* Twitter* LinkedIn This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aistudios.substack.com
    Show More Show Less
    39 mins
  • How a Google Product Manager became a full time AI Creator with Bilawal Sidhu

    How a Google Product Manager became a full time AI Creator with Bilawal Sidhu
    May 15 2023

    How did a Google Product Manager decide to leave his full time job to become an AI Creator? Bilawal Sidhu most recently worked as a Senior PM at Google Maps, where he led Immersive View and was responsible for new technology innovation. Over time he grew his hobby into a new business. In this conversation he explains how he did it, how he approaches content creation, creator tools, mistakes, how AI is helping creators, pros and cons of working in a big company and more.

    Listen now on Apple and Spotify.

    ⏰ 0:37 Bilawal's Journey

    ⏰ 3:25 How Bilawal managed to both work at Google and be a Creator, what kept him motivated and Tips.

    ⏰ 4:57 Approach to Creation

    ⏰ 8:08 Creation Strategy

    ⏰ 9:44 How knowledge of creator tools can helped in day job

    ⏰ 12:00 Mistakes and Advice to tech professionals

    ⏰ 14:53 That Creator Life

    ⏰ 20:30 AI Tools Landscape - Industries, verticals, and Tools (Autodesk, Adobe, Cinema 4d, Blender.org)

    ⏰ 26:52 How AI is helping Creators

    ⏰ 29:45 Advent of the AI Creators

    ⏰ 36:24 What AI will do for Content Consumption

    ⏰ 43:13 Pros and Cons of working on AI at Google and Meta AI

    ⏰ 1:03 Bilawal's favorite AI Tools and what's next

    Referenced

    * Corridor Digital

    * Marques Brownlee

    * Freddie Wong

    * Riley

    * Nathan Lands

    * Joma Tech

    * Midjourney

    * Controlnet

    Where to find Bilawal

    * TikTok

    * YouTube

    * Twitter

    * LinkedIn

    Where to find Natalia

    * Twitter

    * LinkedIn



    This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aistudios.substack.com
    Show More Show Less
    59 mins
  • 👩🏻‍🔬How to do User Research for AI Products with Lauren M. Kaplan, PhD

    👩🏻‍🔬How to do User Research for AI Products with Lauren M. Kaplan, PhD
    May 5 2023

    What is User Research and why is it useful for AI product development? Lauren Kaplan is a mixed methods researcher passionate about inclusion, leveraging technology for social good, and learning. At Meta, she led research on Privacy Preserving Machine Learning (PPML) and PyTorch (an Open Source AI framework) advocating for people centric AI.

    ⏰ 0:49 About Lauren

    ⏰ 1:11 What is Mixed Methods UXR

    ⏰ 1:27 What is User Research

    ⏰ 2:10 How to match User Research with Product Development

    ⏰ 3:13 What are the benefits of User Research for AI Products

    ⏰ 4:00 What's the difference between User Research and User Feedback

    ⏰ 6:45 Challenges of doing User Research for AI

    ⏰ 9:05 How to approach User Research for Generative AI

    ⏰ 10:20 Privacy Preserving ML User Research

    ⏰ 12:23 Synthetic Users

    ⏰ 16:05 How to get into AI User Research

    ⏰ 17:22 How Lauren stays on top of AI News and Advancements

    ⏰ 19:10 How to do User Research for Open Source AI

    ⏰ 21:47 Working with AI Researchers and bridging the discipline gap

    ⏰ 23:06 How should AI Researchers ensure they're people centric

    ⏰ 26:45 What stood out about AI Privacy vs other AI

    ⏰ 29:15 What was it like to work on PyTorch

    ⏰ 31:30 What AI Lauren is excited about next

    Referenced

    * Mapping Strategic, Iterative, and Evaluative Research to Product: Matt's UXR Process & FAQ

    * Google PAIR resources: People + AI Research - Chapters

    * “How can companies help people understand privacy-enhancing technologies like on-device learning?” 

    * The Future of AI is People-Centered

    * Mapping qualitative and quantitative methods Comparing UX Research Methods

    * Synthetic Users: [2209.06899] Out of One, Many: Using Language Models to Simulate Human Samples

    Where to find Lauren & her work

    * LinkedIn

    * Twitter

    Where to find Natalia

    * Twitter

    * LinkedIn



    This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aistudios.substack.com
    Show More Show Less
    34 mins
  • 📈How Esteban Constante grew Leonardo.Ai to over a Million Subscribers

    📈How Esteban Constante grew Leonardo.Ai to over a Million Subscribers
    Apr 25 2023

    Esteban Constante is the Chief Marketing Officer and entrepreneur who is the mastermind behind Leonardo.Ai’s growth. In this conversation Esteban explains how to build and grow an AI startup that's billed by some as the Midjourney killer. Esteban discusses his background, Leonardo.Ai use cases, growth and taking the latest cutting edge research to production. He also covers some interesting startup challenges, his approach to growth and advice for startup founders looking to grow their audience.

    Listen now on Apple and Spotify.

    Detailed Breakdown

    ⏰ 0:00 How Esteban landed an exec role at one of the hottest Generative AI Startups

    ⏰ 4:57 How to Grow a Startup to over a Million Users

    ⏰ 6:52 Why Leonardo.Ai may be the Midjourney Killer

    ⏰ 10:20 What is an AI Artist, a new breed of creative

    ⏰ 13:12 Leonardo.Ai Use Cases

    ⏰ 18:00 Original Thesis behind Leonardo.Ai

    ⏰ 21:45 The Team behind Leonardo.Ai

    ⏰ 26:00 Taking AI Research to Production

    ⏰ 27:30 Challenges tied to Fast Growth

    ⏰ 30:00 Growth Advice for Startup Founders

    ⏰ 32:35 How to Position your Product

    ⏰ 36:20 Who Inspires Esteban

    Referenced

    * Leonardo.Ai

    * Breakthrough Advertising by Eugene M. Schwartz

    Where to find Esteban & his work

    * Esteban on LinkedIn

    * Twitter

    Where to find Natalia

    * Twitter

    * LinkedIn



    This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aistudios.substack.com
    Show More Show Less
    43 mins
  • Framework for Generative AI Use Cases with Barak Turovsky, Executive in Residence at Scale Venture Partners, ex-Head of Product, Google Languages AI

    Framework for Generative AI Use Cases with Barak Turovsky, Executive in Residence at Scale Venture Partners, ex-Head of Product, Google Languages AI
    Apr 14 2023

    Barak Turovsky is an Executive in Residence at Scale Venture Partners. Previously Barak was a Director of Product at Google AI where he spent 10 years leading product management and user experience for Languages AI and Google Translate. Most recently, Barak was  a CPO of Trax, a global retail tech leader that leverages Computer Vision AI for omnichannel shopping experiences. In this episode we discuss whether the current AI cycle is hype or real, Barak’s framework for evaluating generative AI use cases, how to think about foundational and fine tuned models, the business of AI and more!

    Listen now on Apple and Spotify.

    Detailed Breakdown

    ⏰ 2:05 Barak’s AI Product Career

    ⏰ 4:13 AI Cycle - Hype or Reality

    ⏰ 5:44 Barak’s Framework for Evaluating Generative AI Use Cases

    ⏰ 13:54 Foundational vs Fine Tuned Models

    ⏰ 18:53 Limitations of Large Language Models

    ⏰ 22:00 Business of Generative AI

    ⏰ 25:00 What Generative AI means for Content Generation

    ⏰ 27:00 Key Considerations for building AI Products

    ⏰ 32:00 Will AI mean the end of some jobs?

    ⏰ 38:00 What could make an AI Product and Business defensible?

    ⏰ 42:00 What Barak is excited about next?

    Referenced

    * Barak's Framework for Evaluating Generative AI Use Cases (Written Form) 

    * The Great AI awakening story from New York Times (covering the first-ever productization of deep neural networks with Google Translate), is one of best written stories about ML/AI history and challenges

    * Barak’s presentation to Indian Prime Minister Modi when he visited Google campus

    * Google Translate vs. La Bamba (Google Translate team playing with visual, AR-based translation feature)

    Where to find Barak & his work

    * Barak on LinkedIn

    * Twitter

    * Barak's Interview with GLG Experts Network about Generative AI

    * Barak's Keynote at CES 2021 about latest and greatest in Languages AI 

    Where to find Natalia

    * Twitter

    * LinkedIn

    * Instagram



    This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aistudios.substack.com
    Show More Show Less
    45 mins