No Doubt, Let's Go!
On This Episode of No Doubt, Let’s Go! Coach Reggie [https://substack.com/profile/71158847-coach-reggie] interviews Oliver King Smith, founder of Smartr.ai [http://smartr.ai], a consultancy specializing in private AI models for enterprise clients. Oliver discusses how he pivoted from a rejected AI division proposal at his previous company in 2018 to founding Smarter AI, currently working on a major enterprise-wide grant with a manufacturing client. We explored Oliver’s approach to implementing AI in organizations, emphasizing the importance of starting with low-risk, high-value projects like handling legacy data or expense reports to build trust among skeptical stakeholders. Oliver shared his unique working setup reclining on a couch with a laptop on his chest, his use of mental walks to solve complex problems, and his preference for working in burst-like sessions rather than long coding marathons. We also discussed the challenges of managing AI development in an era of rapidly evolving technology models and the balance between planning and adaptability in running a business. Summary Oliver’s AI Entrepreneurship Journey Oliver and Coach discussed Oliver’s journey as an AI pioneer who founded Smarter AI after his previous company rejected his 2018 proposal to start an AI division. Oliver explained that the rejection initially felt like a setback, but it ultimately became an ally moment that led to the creation of his own consultancy. The conversation was cut off mid-sentence, so the specific details about how the rejection specifically helped his current venture were not covered in the provided transcript segment. AI Implementation in Manufacturing Oliver discussed his experience working with AI projects for clients and his determination to implement AI technology in product development. He explained the challenges of introducing AI in legacy manufacturing organizations, emphasizing the need for C-suite buy-in and focusing on low-hanging fruit to build trust. Oliver also highlighted the importance of addressing trust issues within organizations by identifying areas where people feel comfortable with the technology and its impact on existing workflows. AI Implementation Strategy Discussion Oliver discussed strategies for implementing AI in organizations, emphasizing the importance of starting with non-disruptive, high-value tasks that people find unpleasant, such as filling out expense reports. He noted that focusing on big data problems and legacy data management helps build trust in AI technology without causing resistance. Oliver also mentioned that their company specializes in private AI models, particularly serving clients in aerospace, defense, health, and education sectors, with a strong focus on advanced manufacturing. Unconventional Working Setup Discussion Oliver discussed his unconventional working setup using a reclining position with a computer close to him due to his bullet-resistant glasses requiring large font sizes. He explained that this lying-down typing method is relaxing and effective, though he acknowledged the challenges with speech-to-text technology due to human speech patterns containing fillers like “uh” and “um.” AI Problem-Solving Walking Strategies Oliver discussed how taking walks helps him overcome mental blocks when solving complex AI problems by breaking out of tunnel vision and allowing him to see the problem from a wider perspective. He explained that he has standard walks of different lengths near the Royal Botanic Gardens in Edinburgh, which he uses for mental resets. The conversation ended with the coach beginning to ask about working in bursts and its relationship to maintaining quality in AI models, but this part of the discussion was not captured in the transcript segment. Bursty AI Development Approach Oliver explained that the bursty approach to AI development has advantages over traditional software development because AI systems can provide value after just 5-6 hours of work, unlike conventional software which requires a minimum viable product before being useful. He noted that AI development follows an opposite process to traditional software, where getting initial low-hanging fruit is prioritized over building a polished final product. Coach asked about the mindset required for incorporating this different way of working, but the discussion ended without a complete answer. AI Problem-Solving Challenges Oliver discussed the unique challenges of working with AI, explaining that unlike traditional software development where problems are clear-cut, AI work involves more exploration and uncertainty about whether issues stem from the technology, implementation, or understanding. He shared an example of how his team used creative problem-solving techniques to get LLMs to solve the Einstein riddle by having them think about the problem differently rather than trying to solve it directly, which predates current agentic AI approaches. AI Orchestration for Information Systems Oliver discussed how AI systems can be more effective when they serve as orchestrators of information rather than needing to know answers directly. He noted that modern models contain vast knowledge databases but tend to regress toward common approaches rather than suggesting optimal solutions. The conversation ended with Coach asking about balancing a joy-led approach with the structured requirements needed for large manufacturing grants, but this part of the discussion was not captured in the transcript segment. Business Strategy and Market Focus Oliver discussed the challenges of running a business compared to an engineering program, noting that business requires more flexibility due to external factors like the sudden AI trend following ChatGPT’s release. He observed that focusing on specific markets, particularly aerospace and defense, has been more effective for their company. Oliver shared an example of how market focus led to opportunities with a major auto manufacturer through engineering organization connections. Smarter AI Business Development Strategy Oliver discussed how Smarter AI’s business development has been organic rather than planned, with opportunities emerging through connections with aerospace and defense companies. He explained that while AI language models have been progressing too quickly for fine-tuning to remain effective, vision systems continue to benefit from customization for specific clients. Oliver shared his approach to managing morning thinking time with coffee and planning, acknowledging the challenge of business planning when everything changes so quickly. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit rvhoss.substack.com/subscribe [https://rvhoss.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2]
46 Episoder
Kommentarer
0Vær den første til å kommentere
Registrer deg nå og bli medlem av No Doubt, Let's Go! sitt community!