Date: August 13, 2025
Status: Successfully delivered, collecting insights
Core Strategy: Position workshops at the intersection of rapid AI industry evolution and deep technical expertise. Each workshop serves as a vehicle for delivering distilled insights on how AI is transforming industries in real-time.
Evolution Framework: Through iterative feedback analysis across workshops, we're discovering the heuristics that differentiate transformative workshops from mediocre content delivery:
- Bad meeting β Could have been an email
- Bad workshop β Could have been a YouTube video
- Good workshop β Interactive intelligence transfer that changes how people think
Goal: Systematically refine these heuristics through cumulative feedback analysis to consistently deliver workshops that capture market signal, share battle-tested expertise, and create genuine cognitive shifts in attendees.
- Python steering scripts demonstration was a hit - audience loved seeing actual code running
- Q&A format with 2-3 minute audience discussions before steering back created energy
- Imperfections made it human and authentic (vs polished video content)
- Multiple attendees praised articulation and concept breakdown
- Enjoyable speaking style that resonated with audience
- Consistent feedback from last workshop - should lean into this strength
- Clear explanations that made complex topics accessible
- Runs enterprise B2B AI agent solutions
- Perfect role model for HasanLabs direction
- Only attendee familiar with OODA loop concept
- Shared deeper military history context (John Carver, Robert Corman)
- Potential collaboration opportunity
- Live code execution created "magic moments"
- Real-world implementation examples resonated
- Core insight: "A bad workshop could have been a YouTube video" (like bad meeting = email)
- Audience valued:
- Real-time interaction with presenter AND other attendees
- Authentic mistakes and how they were solved
- Raw, unfiltered experience sharing
- The presenter as "guide" through complex territory
- Key quote: "Anyone can read papers, but not everyone can get the raw, authentic human experience"
- Had prepared web demo but only showed Python scripts
- Missing visual impact of full system in action
- Lost opportunity for "click and see magic happen" moments
Instead of just "any questions?", implement structured checkpoints:
Checkpoint Structure:
- Display 2-3 thought-provoking questions on slide
- Open-ended questions designed to inspire new thinking
- Questions audience hasn't considered before
- Directly related to subject matter just covered
- Silent thinking time (1 minute)
- Let audience process and formulate thoughts
- No pressure to immediately respond
- Facilitated discussion (5 minutes max)
- Audience raises hands to share answers
- Foster discussion between attendees
- Cap at 5 minutes to maintain momentum
Checkpoint Timing:
- First checkpoint: After first third of workshop
- Second checkpoint: After second third of workshop
- Creates natural rhythm and expectation
- Let numbers sink in: "When showing numbers and demos on screen, let it sink in!"
- Show before/after clearly: "Didn't show where exactly the layer injection was happening"
- Pre-reading materials: "Send stuff out to people ahead of time for optional reading"
- Visual accessibility: "Slides should be light mode and dark mode compatible"
- Human element valued: "Personal experience matters because it's imperfect"
- Content saturation context: "We're already inundated with content online"
- OODA loops in AI: "Think about OODA loops in the AI loop"
- Service business opportunity: "Multi agent systems for service businesses managing multiple tools for payroll, different business things, etc. Complex enough to be annoying but simple enough for a multi agent system to handle"
- Create optional pre-reading package (send 3 days before)
- Test all demos (web + code) in checklist
- Create light/dark mode compatible slides
- Design interaction checkpoints every 10-15 minutes
- Craft thought-provoking checkpoint questions for each section
- Start with web demo for visual impact
- Show clear before/after for technical changes
- Pause after showing impressive numbers/results (5-10 seconds)
- Include 3-4 "click and magic happens" moments
- Implement Dr. Pingali's checkpoint framework (2 structured discussions)
- Lean into articulate speaking style (validated strength)
- Share authentic mistakes and solutions (guide role)
- Web interface first (visual hook)
- Show the problem it solves
- Reveal the code behind it
- Let audience process the impact
- Interactive checkpoint with thought questions
- Immediate feedback collection form
- Follow up with key contacts within 24 hours
- Document all insights in Linear
- Research his enterprise B2B AI agent work
- Document his approach to multi-agent systems
- Schedule follow-up meeting
- Thank him specifically for checkpoint framework suggestion
- Research John Carver and Robert Corman contributions
- Document military history context shared
- Create content piece on OODA loops in AI systems
The magic formula: Live demonstration + human imperfection + strategic interaction points = engaged audience
Workshop vs YouTube Video Test:
- Can audience interact with each other? β
- Can they ask real-time questions? β
- Do they get authentic mistakes/solutions? β
- Is presenter a guide, not just broadcaster? β
If yes to all β Good workshop If no to any β Should be a video
Unlike video content, in-person workshops thrive on:
- Real-time problem-solving
- Authentic struggles and recoveries
- Audience participation in the journey
- Tangible "click and see" moments
- Presenter as experienced guide sharing battle scars
- Research Dr. Pingali's work for collaboration potential
- Create workshop template with enhanced checkpoint system
- Build service business multi-agent demo
- Document OODA loop applications in AI
- Develop bank of thought-provoking questions for checkpoints
Note: Workshop attendee requested AI Research Hub feature on HasanLabs website - implementation completed during workshop debrief
- URL: https://linear.app/hasanlabs/issue/HL-185/workshop-post-mortem
- Identifier: HL-185
- Status: In Progress
- Priority: High
- Assignee: Unassigned
- Created: 2025-08-13T21:50:08.607Z
- Updated: 2025-08-13T22:21:27.177Z