Today’s guest is Gordon Hart, Co-Founder and Head of Product at Kolena. Gordon joins Seth Earley and Chris Featherstone and shares how ​​machine learning algorithms are a challenge from different perspectives. Gordon also discusses the core problem in his company before they turned it around. Be sure to listen to Gordon's advice on how to validate models in order to have a successful product!

Takeaways:

• Gordon noticed that developing algorithms internally or buying from other model vendors has really had a constant unexpected model behavior. It made him feel he couldn’t trust the models to behave sensibly. 
• Gordon started his company because he noticed that time after time, he was getting blindsided. He knew there was a better way to develop models and validate what they were doing. 
• The key challenge that Gordon and his team ran into was that when you have all the data when they were looking at that one number, they were looking at that aggregate metric computed across their entire benchmark.
• Gordon expresses the importance of going through scenarios with your products. He found that when you break down your evaluation into these different scenarios, the test gives you an understanding of how this model improves in the aggregate over previous models and how are the failures distributed.
• Testing data is more critical than training data because your testing data is used to determine if your new model has the correct behaviors.
• Testing the full pipeline from pre-processing through post-processing rather than testing the model component will oftentimes improve the visibility into how your product is actually going to work when you put it out there.

Quote of the Show:

• “Having your evaluation metrics align with the way that your system is going to be evaluated in the field is a key thing that you can do to get a better understanding of ‘is this model better for what I set out to do?’” (22:36)

Links:

• Twitter: https://twitter.com/kolenaIO 
• LinkedIn: https://www.linkedin.com/in/gordon-hart/ 
• Website: https://www.kolena.io/ 

Ways to Tune In:

• Website: 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

Today’s guest is Daniel Faggella, Head of Research and CEO at Emerj Technology Research. Dan joins Seth Earley and Chris Featherstone and shares how martial arts influenced him to get into artificial intelligence. Dan also discusses what his experience was like with surveillance technology creation technology. Dan had a machine that could generate the next 10 slides of your desired moving picture. Be sure to listen in on Dan giving his advice on how you should properly use open AI!

Takeaways:

• Dan got into artificial intelligence by practicing the martial art, Jujitsu. He started a Jujitsu gym which helped support him when he was in school. Jujitsu helped motivate him and keep his mind balanced.
• Dan mentions how generative AI has been starting to bubble up since the spark of ChatGPT. He sees people starting to experiment with social and proposals. 
• With AI in general, people are looking at junctures within the workflow. Identifying junctures where can push a button will lead to streamlined deliverables.
• Generative AI finds the juncture pockets and knows exactly where those settle in.
• Dan speculates that people will evolve their use of ChatGPT and structure different FAQs.
• Dan believes that one day we'll use Generative AI to create a feedback loop allowing humans to say what's wrong and what's right to train AI systems.

Quote of the Show:

• “The dust has yet to settle on the early cluster of those use cases in Generative AI.” (19:06)

Links:

• Twitter: https://twitter.com/danfaggella 
• LinkedIn: https://www.linkedin.com/in/danfaggella/ 
• Website: https://emerj.com/ 
• Podcast: The AI and Business Podcast
• Article: Lotus Eaters and World Eaters

Ways to Tune In:

• Website: 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

Today’s guest is Michelle Zhou, Co-Founder and CEO at Juji, Inc. Michelle joins Seth Earley and Chris Featherstone and dives into what proprietary data is and how it can be used correctly. Michelle also discusses the one lesson she has learned is that you have to build a product that can help people. You want to achieve your customers' outcomes, not your outcomes. Be sure to listen in on Michelle giving her advice on how to pick out the golden nuggets in AI data to show a coherent and meaningful summary!

Takeaways:

• When Michelle first started with computer science, she wasn't fond of it until she attended Michigan State University where two professors changed her perspective on computers. They gave her the opportunity to work on building graphical user interfaces for power management and worked on projects that dealt with AI data storytelling.
• Michelle explains that the AI data storyteller gives a set of data and tasks of the user which then gives the user visual preferences. It also consists of a series of animated data visualization.
• During Michelle’s first 15 years of research, she was working on understanding users in a task context. For example, what their tasks are, what they're looking for, what their visual preferences are, and what their verbal preferences were.
• Michelle has noticed a lot of students will strive for a degree that their family has done in the past. Michelle says that you don’t always have to follow any degree you don’t want. There are so many unique degrees to pick from.
• Michelle believes that transparency drives responsibility and since they have a powerful AI system, she wants to make sure that they use their AI in a responsible way.
• The one lesson Michelle has learned is that you really have to build a product that can help people. Make sure to achieve your customers' outcomes and not yours. You don’t want to waste their time.

Quote of the Show:

“I want to really democratize the use of this cutting-edge technology.” (23:41)

Links:

• Twitter
• LinkedIn
• Website

Ways to Tune In:

• Website 
• Apple Podcast
• Spotify
• iHeart Radio
• Stitcher
• Amazon Music
• Buzzsprout

Thanks to our sponsors:

• VKTR
• Earley Information Science
• AI Powered Enterprise Book

Today’s guest is Juan Sequeda, Principal Scientist at data.world and Co-Host of the Catalog & Cocktails Podcast. Juan joins Seth Earley and Chris Featherstone and shares how to understand the problem that you are trying to solve. Juan also discusses how your company's success should be defined differently. Don’t focus on just on saving money to make money. Focus on solving a problem.  Juan also shares valuable advice on how understanding who you report to helps you speak the same language.  

Takeaways:

• Juan believes the market is immature when it comes to what they want or what they think they want. This is where data catalogs become important so that companies can locate information. 
• From the perspective of the data management world, it’s focused on only technology. The problems that they had been trying to solve 30 years ago continue to be the same problems they’ve been trying to solve.
• If you are on the technical side of your business, it is important to understand who you should be reporting to. Understanding this early on will help you tailor information to meet the correct outcome. 
• Juan’s definition of a knowledge graph is representing real-world concepts and the relationships between those real-world concepts end up forming a graph. The reason why the graph is really valuable is because you can integrate data coming from many diverse sources.

Quote of the Show:

“Keep working on the same vision.” (07:50)

Links:

• Twitter
• LinkedIn
• Website
• Podcast: Catalog & Cocktails presented by data.world
• Juan’s Portfolio
•  Book: Integrating Relational Databases with the Semantic Web

Ways to Tune In:

• Website
• Apple Podcast 
• Spotify
• iHeart Radio
• Stitcher
• Amazon Music
• Buzzsprout

Thanks to our sponsors:

• VKTR
• Earley Information Science
• AI Powered Enterprise Book

In this episode, our guest is Andy Fitzgerald and Information Architecture & Content Strategy Consultant.

Highlights:
1:40 - Getting from Ph.D. in English and Literature in information architecture and knowledge graphs
9:23 - Schema.org
14:30 - How can we get search to be like "Google"?
19:00 - The trouble with self-organizing information
20:40 - The KFC debacle in Germany and case for keeping humans in the loop
22:15 - Knowledge graphs and AI
29:35 - Role of linguistics
33:00 - What happens when you don't apply knowledge graphs to AI projects
37:00 - Boutique knowledge graph - UXMethods.org
48:00 - Value of smaller scale knowledge graphs and simplicity

Links:

• Connect with Andy on Linkedin
• andyfitzgeraldconsulting.com
• UX Methods - a community powered, linked data driven knowledge graph for learning about the techniques of user experience design.
• Jim Hendler's papers
• Gall's Law

Thanks to our sponsors:

• VKTR
• Earley Information Science
• AI Powered Enterprise Book

In this episode, Seth and Chris talk with Peter Voss,  Founder, CEO, and Chief Scientist at AGI Innovations & Aigo.ai.

Highlights:
2:58 "Software is quite dumb"
3:51 "What is reality?"
5:00 Coining the phrase "Artificial General Intelligence" - what it means
9:00 On understanding cognition in the deepest terms
11:10 What is consciousness?
15:20 Difference between "Artificial Intelligence" and "Artificial General Intelligence"
19:00 The 3 waves of AI
29:45 What is cognitive architecture?
34:30 Quality of data vs quantity of data
38:00 Practical applications for building personalization systems
39:20 What can organizations do to prepare for AI driven systems?
46:30 One corporate bot or multiple bots?
53:45 Automation should be able to deliver the superior customer experience, not the cheaper second class option

Links:

• Connect with Peter on Linkedin
• Visit the Aigo.ai website

Thanks to our sponsors:

• VKTR
• Earley Information Science
• AI Powered Enterprise Book

In this episode, Seth and Chris talk with Dan Turchin, CEO & Founder of PeopleReign.

Highlights:
4:30 What drives Dan's mission to impact a billion lives at work?
11:15 Disruption and the future of work
17:50 What will happen when people can have a day a week back from automation?
20:00 Work will change more in the next 30 years than in the previous 300
24:30 AI is really still in its infancy - how can we use it for good as it grows up?
29:00 What are the pre-requisites to success with AI?
37:50 How do you sell the business on the need to address their data?
43:00 Choosing use cases to get started with
51:00 What's next?

Links:

• That episode of his podcast with Seth that Dan mentions
  https://peoplereign.io/2022/09/seth-earley-author-of-the-ai-powered-enterprise-discusses-the-future-of-knowledge-management
• Dan's LinkedIn Profile
  https://www.linkedin.com/in/dturchin/
• PeopleReign
  https://peoplereign.io/

Thanks to our sponsors:

• VKTR
• Earley Information Science
• AI Powered Enterprise Book

In this episode, Seth and Chris talk with Dr. Mark Maybury, former CTO with Stanley, Black & Decker.

Highlights:
3:30 Mark's early influences
9:10 What he does with his "spare" time - it is planned
16:30 His experience at Stanley, Black & Decker - making the elephant dance
22:50 Transitioning from analog to mixed reality (physical and digital)
31:00 AI - doing the early foundational work without existing ML systems
40:00 Early development of sentiment and intent analytics
45:00 Future projects including movie project getting students excited about careers in AI in the service of the public good
48:45 Ready Robotics
51:00 Development of a COVID De-activization protocol

Links:

• LinkedIn Profile
  https://www.linkedin.com/in/dr-mark-maybury-28532/
• Ready Robotics
  https://ready-robotics.com/

Thanks to our sponsors:

• VKTR
• Earley Information Science
• AI Powered Enterprise Book

In this episode, Seth and Chris talk with Steve Stesney, Senior Product and Data Practice Lead at Predictive UX.

Highlights:

3:30 What is predictive user experience and how did Steve get there?
5:30 Grasping disambiguation and the move to graph data
18:00 Building trust in the data
19:26 Explaining taxonomy, ontology and knowledge graph to executives
21:30 Connecting UX and knowledge graphs
28:00 Managing the open floodgates once users discover what knowledge graphs can do
36:30 What is "predictive ux" ?

Links:

• LinkedIn Profile
  https://www.linkedin.com/in/stephenstesney/
• Website
  https://www.predictiveux.com/

Thanks to our sponsors:

• VKTR
• Earley Information Science
• AI Powered Enterprise Book

In this episode, Seth and Chris talk with Scott Taylor, the "Data Whisperer" about telling stories about data management.

Highlights:
2:00 Data Whisperer origin story
9:30 Translating complex dry material into a story that resonates
11:30 Why master data is the most important data and how to help execs understand it
18:15 Bad data + AI = AS (Artificial Stupidity)
22:30 Every system demos perfectly
26:25 Don't say "data quality"
27:30 Definition of digital transformation
32:00 Ugly babies and the reality of bad data
38:30 About the book, "Telling Your Data Story" 99% buzzword free (coupon code in show notes)
47:00 Data management is macro trend agnostic
49:00 What's next - more puppets and dad jokes
52:00 Influencing the next generation of data managers

Links:

• LinkedIn Profile
  https://www.linkedin.com/in/scottmztaylor/
• Website
  www.MetaMetaConsulting.com
• BOOK
  Telling Your Data Story  - Data Storytelling for Data Management
  20% off with code: DATAWHISPERER (publisher site only)
  https://technicspub.com/data-storytelling/
• Top Content
  Too Much Tech Talk?  (A puppet service announcement)
  The Little Red Data Hen - A Cautionary Tale
  Connecting Data Management to the Essence of Your Business
  Data Has Got to Move to Have Value
  The Super Hero Adventures of **Master Data**
  Scott Taylor - The Data Whisperer YouTube Channel

Thanks to our sponsors:

• VKTR
• Earley Information Science
• AI Powered Enterprise Book