
Sign up to save your podcasts
Or


Key topics:
Timestamps:
00:00 - Introduction and motivation for AI harnesses in startup and enterprise contexts
02:10 - Reimagining the SDLC in the era of AI and real-time feedback loops
04:00 - Achieving over a thousand story points per engineer and what it means
05:20 - The evolution of story points and complexity estimation with AI's aid
07:00 - Transitioning from waterfall to agile to continuous AI workflows
09:30 - Defining and refining product requirements documents in rapid cycles
11:15 - The concept and importance of outcome attribution and real-time feedback
13:40 - Comparing model importance versus harness design in frontier AI systems
16:15 - Anatomy of a AI harness: context, control flow, action, persistence, observation
18:00 - Building multi-agent, long-running harnesses for complex projects
20:05 - Platform tools: Deep Wind harness setup, rebranding, and collaborative features
22:00 - Practical roles and skills for managing AI harnesses (front-end, back-end, DevOps)
23:45 - Strategic positioning frameworks like Blue Ocean and market differentiation
26:00 - Observing how roles like Product Owner and Scrum Master evolve into more integrated positions
28:00 - Rapid setup of harnesses: from project briefing to execution in hours
32:00 - Managing work, parallel execution, and continuous iteration with AI
36:00 - The importance of detailed planning, hypotheses testing, and validation
40:00 - Challenges and opportunities in AI testing, governance, and external research
44:00 - The critical role of gauntlet reviews and critical assessment in harness design
50:00 - Orchestration of AI tasks via harness coordinator and agent management
55:00 - Estimating time, costs, and token consumption in practical harness deployment
60:00 - The future of autonomous AI agents and reducing human oversight
61:30 - Resources available: platform tools, CLI, and free resources for rapid implementation\
Resources & Links:
Connect with George Filipovich:
Connect with Steve:
By Steve SaperKey topics:
Timestamps:
00:00 - Introduction and motivation for AI harnesses in startup and enterprise contexts
02:10 - Reimagining the SDLC in the era of AI and real-time feedback loops
04:00 - Achieving over a thousand story points per engineer and what it means
05:20 - The evolution of story points and complexity estimation with AI's aid
07:00 - Transitioning from waterfall to agile to continuous AI workflows
09:30 - Defining and refining product requirements documents in rapid cycles
11:15 - The concept and importance of outcome attribution and real-time feedback
13:40 - Comparing model importance versus harness design in frontier AI systems
16:15 - Anatomy of a AI harness: context, control flow, action, persistence, observation
18:00 - Building multi-agent, long-running harnesses for complex projects
20:05 - Platform tools: Deep Wind harness setup, rebranding, and collaborative features
22:00 - Practical roles and skills for managing AI harnesses (front-end, back-end, DevOps)
23:45 - Strategic positioning frameworks like Blue Ocean and market differentiation
26:00 - Observing how roles like Product Owner and Scrum Master evolve into more integrated positions
28:00 - Rapid setup of harnesses: from project briefing to execution in hours
32:00 - Managing work, parallel execution, and continuous iteration with AI
36:00 - The importance of detailed planning, hypotheses testing, and validation
40:00 - Challenges and opportunities in AI testing, governance, and external research
44:00 - The critical role of gauntlet reviews and critical assessment in harness design
50:00 - Orchestration of AI tasks via harness coordinator and agent management
55:00 - Estimating time, costs, and token consumption in practical harness deployment
60:00 - The future of autonomous AI agents and reducing human oversight
61:30 - Resources available: platform tools, CLI, and free resources for rapid implementation\
Resources & Links:
Connect with George Filipovich:
Connect with Steve: