Conversations on Applied AI

The episode contains a handful of actionable tactical tips (voice dictation for prompts, a NotebookLM→PDF→Canva workflow, the Plaud recording device, the 7-9 word average prompt statistic) but the bulk of the runtime is motivational pep talk, origin story, and generic adoption advice a smart operator has heard a dozen times. The ratio of novel-to-filler is low.

“the average prompt is seven to nine words. People do it like, 'cause that's what a good Google search is... where an average good prompt is at least 21”

“you export it as a PDF? You can import that into Canva as a PowerPoint, and then you can hit the Magic Editor in Canva and lift the text”

Most of the thinking recycles widely circulated ideas - habit stacking, dopamine loops for behaviour change, AI as thought partner, 'you don't need to be technical.' The 'context engineering' reframe and the voice-dictation-as-prompting angle are mildly interesting, but the CRAFT framework itself is a thinly differentiated acronym over standard communication advice.

“Context Engineering is the word I like. Really, it's about good communicating”

“I use it by like voice dictating my thoughts and brain dumping, and then keeping my voice and my ideas, but having them tightened, rearranged, and streamlined”

Tyler Moberg is a credible AI-adoption trainer with real client volume (10,000+ professionals claimed), but his background is instructional design and middle-school science teaching, not operating or scaling a B2B business using AI. He is a competent practitioner-educator, not a senior operator who has driven measurable business outcomes with AI at scale.

“I started as a middle school science teacher, I spent about 13 years just in the classroom”

“I'm a solopreneur, but I think I operate like I have. Five employees”

A few concrete client vignettes add grounding - the Public Works engineering team saving '80 hours' using NotebookLM, the relocation company cutting 1.5-hour post-call write-ups - but these are single anecdotes with no corroborating data, no revenue figures, and no rigorous before/after metrics. The named tools and the prompt-length statistic are the strongest specific evidence.

“The engineer was like, you know what? We just saved like. 80 hours right now in terms of like literally how this is what he had budgeted his time”

“they were literally spending an hour and a half after every call doing this and now. Boom”

The host asks almost exclusively open-ended, promotional questions and frequently pivots to share his own anecdotes rather than drilling into the guest's claims. There is no pushback, no challenge of the '80 hours saved' figure, no follow-up on the CRAFT framework's evidence base, and the guest literally reverses the interview at one point. Mutual admiration dominates the final minutes.

“By the way, before I, I don't mean to interview you, but I'm always like, what's like a, a mistake or a misconception that you're finding out there?”

“I wanna say this publicly, like I highly respect you the way that you lead. I just feel like you lead by example”