Chain of Thought - Not the Model’s - Mine Envisioning a Diversified Frontier Firm (DFF): an integrated, AI-native business where humans and AI agents share capabilities, infrastructure, and objectives across multiple business lines. My first “org chart” had me as CEO, with an ever-evolving choreography of agents across media, education, consulting, and physical products. Honestly, “ever evolving” initially meant “ever darkening”: humans out of the loop. Then I realized I have more to offer than direction and permissions. Sorry, hopeful agentic business CEOs, that’s pretty much the job description. Coining a term: HaaIC. Human as an Individual Contributor. What I bring to the team, and love engaging in: ▪️ Principal, Ideas and Expression: Unless you’ve been asleep since 2023, philosophy and its kin are rising rapidly in significance and value. ▪️ Design Thinker: In my opinion, design thinking and data governance belong in the top five skills, regardless of the project. ▪️ In-Person Events and Networking: Until embodied AI becomes the norm, this is genuine space where humans can do what machines cannot. ▪️ iPhotographer: Since handing my Instagram and Facebook posting to Meta Muse, many of my accounts have become nothing but infographics. My apologies. 🔄 Let’s be honest: slop is AI-generated content that hasn’t had enough training and guidance. Human employees need guidance, context, and examples too. That choke point made me rethink my value. If an agent hesitates to use stock photos or repurpose photos full of people, I can give it a “social photos” folder with material I’ve created for that purpose. Stay with me: photography and video are interests I’ve never made time to explore. Why not capture interesting stills with my iPhone and give my social agents something personal to work with? 🔜 Those infographics will be replaced by amateur photography. Mine. I also need to coach the agents to use the hundreds of video clips I already have in Riverside. What are your HaaIC skills? Comment please.

Before you hand work to an agent, write down who receives the result and what they can inspect before accepting it. We're still designing and testing our firm, so this is a working practice, not a proven result. schoolsout.agipodcast.ai/new…

At review compare measured, delivered and accepted work against original criteria. Keep unfinished visible. A miss teaches without renaming it success. Focus on honest review and next step. schoolsout.agipodcast.ai/new…

Hudson Conkling is 21 and working with AI every day. We talked about starting an AI career, learning, and why real connection matters when people use these tools. AGI Season 2, Episode 1 is live. Stream on your favorite platform through the link in my profile.

Read the brief like you missed the meeting. The recipient should understand the delegated outcome without the original planning conversation. The owner decides priorities and tradeoffs. schoolsout.agipodcast.ai/new…

Set the review point before delegating. Define the success criteria and the review date when the recipient will compare the delivered evidence to them. Keep unfinished work visible. schoolsout.agipodcast.ai/new…

For delegated AI work, the blocker report must name the missing input and who can resolve it so responsibility stays clear. Add that requirement to your next brief. schoolsout.agipodcast.ai/new…

China, China, China is live now on The Weekly Blitz. Julie Koehrer and I talk China, the U.S., the AI race, regulation, and who benefits. Stream it on your favorite platform. Link in my profile. #AGIPodcast #WeeklyBlitz

Agree the outcome and evidence that counts before delegating. You own priorities and tradeoffs. Set the agent's authority boundary, recipient, and blocker route. Which decision must come back before work continues? schoolsout.agipodcast.ai/new…

An export notification says a podcast is ready to share. It does not establish the episode is live. Check the destination before recording completion. For the fictional show Daily Brief, verify the host page shows the new episode. What evidence supports a delivery claim?

In the fictional internal newsletter two agents both thought they owned the review step. Name the owner before the next run then record what happened after one test. What step still needs a single named recipient?

Start with one approved source when you brief AI. Name the goal, audience, time available, and output you want. Ask it to flag what the source cannot answer, then review and improve one instruction. Which source will you try this with first?

Ryan Powell Joined Me to Talk Multi-Player Agentic Workflows This episode of AGI Human in the Loop - anthropomorphism, harnesses, data engineering, multi-agent/sub-agent strains on traceability and multi-agent collaboration vs multi-player human collaboration with an agent. Join the full conversation where ever you stream your podcast linktr.ee/AGIPodcast (link in our profile)

Valentine's Day 2026 Ronda Swartz and I Unpacked Claude's Constitution Recent concerns raised by Mustafa Suleyman are bringing renewed scrutiny to Anthropic’s anthropomorphization of Claude. @kberlind of Castaway Podcasts created this cool video after he heard Jason talking about the newly rekindled conversation. castawaypodcasts.com/ Check out our full episode on the topic youtu.be/KnZN_tQorNw

An AI draft lands with clean sentences and one claim that feels a little thin. The useful skill is marking that claim before the draft moves on. What do you check first when the wording looks finished?

Linda Calvin calls communication, collaboration, and critical thinking “power skills.” How do we keep practicing them while using AI? AGI School’s Out Saturdays: youtu.be/qt1jHuJKxug #AGIPodcast #AIinEducation