Gen AI in Enterprise, Data P&L, and Event Storming - Dino Scheidt | The Long Game #054

Welcome to Episode #054 of The Long Game, where I have conversations about tech startups, AI, entrepreneurship, and the ideas and principles to make it all happen.

My guest this week is Dino Scheidt, an AI Engineer, former CTO, and Founder who works with Fortune 50s, Start-Ups, and Governments on Data Intelligence and AI Architectures. In this episode, we discuss the evolving landscape of AI with a particular focus on generative AI and its applications. Dino criticizes the concept of 'AI strategies,' arguing that AI should be seen as a tool rather than a strategy. We also delve into the challenges and opportunities posed by non-deterministic systems, the concept of Data P&L, and the complexity of integrating generative AI into existing business operations. Dino emphasizes the importance of lateral thinking and digital representation of business processes to leverage future AI innovations effectively.

Subscribe if you enjoyed our conversation. If you didn’t like it, leave a comment letting me know why.

I drop new episodes every Tuesday. Keep up with The Long Game by following me on my social media platforms linked below.

EPISODE LINKS:

TIMESTAMPS:

00:00:29 Introduction and background
00:01:03 AI and Generative Models
00:04:54 Deterministic vs Non-Deterministic Systems
00:08:36 Unpredictability, Value Chain, and AI Integration
00:13:43 Hype Cycle: Adjective to Noun
00:19:31 Strategic Integration of Generative AI
00:28:53 Value of Generative AI: Consumer vs. Enterprise Perspectives
00:32:29 Event Storming and Tactical AI
00:40:08 Future of AI and Final Thoughts
00:42:53 Closing

CONNECT:

#TheLongGame #GenerativeAI #AIIntegration #DataPnL #podcast #interview

[EasyYoutubeAI]

Explores the misconceptions of AI strategies, the limitations of generative AI, and emphasizes the importance of integrating technology with existing business strategies.
00:05: AI strategies are non-existent; AI is merely a tool.
01:07: Generative AI is gaining attention but has limitations.
02:10: Hallucinations in AI outputs are often misinterpreted.
03:27: Generative AI is currently at the peak of hype.
04:43: The hype cycle varies for different AI research communities.

06:10: Engineers and managers have contrasting perspectives on technology.
08:41: Non-deterministic systems introduce chaos into predictable environments.
10:01: Chatbots are often misapplied in value chains.
12:50: AI should support existing strategies rather than dictate them.
21:50: Generative AI should focus on simple, deterministic outputs.

[DelmaRaySmithJr]

Saw the RE-20 and subscribed to listen to it work for you.

[pedro_nothere]

Hi Elijah and Dino, thank you very much for your podcast, it is really helpful and interesting!
At 21:40-ish, Dino mentioned that it would be useful if we could write prompts that return only "yes or no." For that, you have several tool libraries such as Pydantic and Instruct, which, when integrated with OpenAI or Anthropic chat completions, make it possible to return the output structure we want, such as a JSON structure with different data types and also enums if you only want a option a list of options as a response.

Continue with your amazing podcasts!

[aga]

My screw driver strategy is to keep the relevant screw driver in the easily accessible rack or tool box.

[gokayk]

I agree 90% of whatever Dino said, but the delivery and the calmness of the video makes it hard to watch

[gyurilajos]

There is enough chaos outside the world

[paksabar]

sakjane enak melu maxim luwih bebas timbange sing liyane

[gyurilajos]

Remember data lakes and all the hype cycles of pump and dump The AI winter is nigh again
Made the comment before listening hope to hear voice of reason not rationalizations of patentvinsanity

[cesar]

Saying that you hate a chat bot already speaks of a bias and a childish approach to the subject, which you cannot trust not to color your chain of thoughts.

[INTELLIGENCE_Revolution]

0:32 AI is a general purpose technology. NOT ‘just a tool’. I appreciate your technical knowledge.. but you have a negative view of the obvious strategic benefits. You would be the opposite of what an existing organisation would need right now.

[themultiverse]

First impressions matter: That goes for any conversation of any topic, no matter how difficult. "I hate..." immediately turned me off.
Good luck to the both of you fellows, but I'm not going to be partaking. SMH

[OldManShoutsAtClouds]

A screwdriver doesn't have privacy and data protection implications. It doesn'thave larger implications outside of its narrow use case and therefore doesn't require a strategy. Dude doesn't seem to actually understand business, I'm not sure why he's speaking so confidently on this topic.