Who Directs the Agents? Vibe Coding, AI Orchestration, and Human Judgment with Brent Rager
AGI - Advance, Grow, Innovate with AI
Who’s talking
2 speakers1Transcript91 turns · 1 hr 36 min
Welcome back for another edition of Humans in the Loop. I have a great interview for you today. We're going to get into Astra and Fable 5.1, but we're also going to take a walk down memory lane and just kind of go through the progression of artificial intelligence from the 1956 Dartmouth Conference to the launch of ChatGPT, November 30, 2022, to what Greg Brockman is now calling AGI. So my guest is going to be Brent Reger. He is a software engineer with Topstaff, and he is also a good friend and a consultant with Chakra AI Solutions, one of the companies I work for, who's now a proud sponsor of this podcast. Chakra AI Solutions has a proprietary way of mapping business processes with small to medium sized businesses, helping to gain clarity on where AI could help your humans do less of what their busy work seems to be and more of what. they want to lean into and are good at. So your salespeople can spend less time in their CRM and more time in front of customers. Your admin people and your CFOs and your CEOs can spend more time in strategy and planning meetings and less time drafting PowerPoints. Chakra AI Solutions, amazing organization. We'll put their links in the show notes. And they will be sponsoring the AGI podcast as well as a LinkedIn Live coming up October 27th. That's going to be a conversation kind of Halloween themed called Data Centers Behind the Cape. Are they really scary? We're going to host a LinkedIn Live podcast. with representatives from Indiana University, Purdue University, Meta, Google, maybe some local city officials around Indiana. We want to have a pragmatic conversation about what the true impact, both positive and negative, of data centers are so that we as Hoosiers can have an informed conversation and make sure that if we are embracing data centers, we're getting the most out of that. So all that to be said, Chakra AI Solutions, definitely worth checking out if you're a small to medium-sized business in the Midwest. And Brent Rager, a new friend, he's the kind of guy who really leans into that developer vibe coder workflow. And so here he is.
Brent Rager, good morning. Morning, Jason. How are you? I'm good. How are you?
I'm doing great. It's a lovely morning today. software engineer for Top Step and founder of SMU AI. Before we get into all that we're going to get into, which we're definitely going to talk about Astra and Fable, we're going to talk a little bit about the trajectory of artificial intelligence since the 1956 Dartmouth conference. Tell everybody a little bit about yourself because what I want to say is there's a lot of talk out there about who the ideal employee is now, and a lot of people are saying it's like part dev, part project manager, part designer. I would argue it is a vibe coder who was a creative turned vibe coder combined with a software engineer who work well as a team. And I think you and I have hit it off pretty good, pretty big in the last month or so.
Yeah, I agree completely. I've been in the industry for 25 years. I started as anything from a team lead of 15 reports at a company called Genesis here in Indy to CTO at some pretty successful startups and revenues in the hundreds of
2Episode notes
AI can now write code, call tools, coordinate other agents, and keep working long after a single chat ends. Brent Rager and Jason Padgett trace what that shift means for builders, businesses, junior engineers, educators, and the people responsible for directing the work.
Brent brings more than 20 years of engineering and technology leadership to a conversation that moves from transformers and GPUs into agent harnesses, model routing, autonomous workflows, local AI, cost, safety, and the future of technical work.
The practical thread running through the episode is simple: the human role is moving upward. More people will define outcomes, select systems, judge results, and decide when an agent needs to stop and ask for help.
WHAT YOU'LL TAKE AWAY
Why customer understanding may matter more than coding speed
How agent harnesses turn language models into systems that can act
Where human review belongs in autonomous workflows
Why junior engineers still matter when AI handles more of the code
How local models, smaller models, and routing affect business costs
What educators need to teach when AI becomes part of every profession
Try this after the episode: choose one recurring workflow and write down the goal, the tools an agent may use, what success looks like, and the moment it must stop and ask a person.
ABOUT BRENT RAGER
Brent Rager is a Staff Software Engineer at