"Training an AI model's final run for $6M does not mean the entire model costs $6M"

In today's episode of Bricks and Bytes, we had Alan Mosca, CTO and co-founder of nPlan, who shared fascinating insights about AI developments, construction tech, and the future of project planning.

Tune in to find out about:

✅ The truth behind DeepSeek's training costs and why media reporting missed the mark

✅ How nPlan's Barry agent helps construction teams make better decisions through meeting notes, WhatsApp integration, and email connectivity

✅ Why the future of software isn't SaaS - it's buying outcomes through agents that perform specific tasks

✅ nPlan's game-changing "Schedule Studio" that can generate detailed project schedules in 45 minutes

Listen to the full episode to learn how AI is reshaping construction planning and why transparency, not blame, is key to better project outcomes.

----

Chapters

00:00 Intro

05:26 Introduction and Overview of Enplan

07:12 DeepSeek and the Reality of AI Training Costs

15:36 Open vs Closed Source Models in AI

26:23 The Rise of Agents in AI and Construction

31:23 Future of Startups: Agents vs Traditional Software

36:01 The Evolution of AI and the Path to AGI

44:01 Defining AGI and Its Implications

45:26 Introduction to Nplan and Its Mission

47:24 The Functionality of Barry: AI in Project Management

49:27 Continuous Learning and Model Optimization

53:23 Demonstrating the Value of Planning in Construction

57:28 Customer Perspectives on AI in Project Management

1:00:31 Product Development and Future Directions for Nplan

1:11:39 Cultural Change in Construction Through AI