DOP 368: The AI Productivity Paradox

Episode 368

Show Notes

#368: You’re faster with AI. Your teammates are faster with AI. More tickets closed, more deployments going out than a year ago. So why can’t anybody in the building answer a simple question: are we shipping more value?

Jeff’s take is that the bottleneck didn’t go away, it moved. Code got cheap. Deciding what to build, why, and whether it actually worked did not. That’s not an engineering problem anymore. It’s a product problem.

In this episode, we speak with Jeff Keyes, Field CTO at Allstacks, about where the bottleneck went once AI made coding cheap, and why product people have to pick up the tools next.

Frequently Asked Questions

Why do developers feel more productive with AI while their teams ship no faster?

Jeff Keyes of Allstacks tells DevOps Paradox episode 368 that more deployments are going through and more tickets are being closed, but teams still cannot answer whether they are shipping more value. Individuals report better personal productivity while throughput and quality do not visibly improve. Viktor Farcic’s reading is that if each person gets more productive and the combination does not, the company’s process is what does not work.

Does AI make software delivery less stable?

Jeff Keyes describes AI on DevOps Paradox episode 368 as an amplifier that magnifies the good and the bad, with the bad currently offsetting much of the good. He points to research showing individual effectiveness rising while software delivery instability rose too: pipelines that were only mostly right produce more failures, higher incident rates and more deploy failures once more work runs through them.

Where has the software delivery bottleneck moved now that AI writes the code?

Jeff Keyes argues on DevOps Paradox episode 368 that what needs to change next is not an engineering problem but a product problem, a definition problem, and an organization and team topology problem. He says Allstacks saw customer token spend start out on code and backlogged features, then shift toward deciding what to build and whether people were actually using it. Answering that reaches upstream to business goals, target customers and identity.

What is flow distribution in value stream management?

Flow distribution, as Jeff Keyes explains it on DevOps Paradox episode 368, is how the work in any given cycle breaks down across innovation, technical debt, risk and defects. Generate only features with AI and there will still be bugs; address only defects and the debt is never reduced. He says product should own that mix. Jeff helped start the Value Stream Management Consortium, since reformed as Flowtopia.

If AI makes fixes cheap, why do companies still carry so much technical debt?

Jeff Keyes answers on DevOps Paradox episode 368 that teams have not discovered all their defects, do not know the impact of fixing each one, and cannot trust any agent to patch a bug and deploy it to production unattended, because most pipelines have too high a failure rate. Viktor Farcic, who raised the question, replies that he trusts humans no more than agents, and suspects companies keep spending new capacity on new features instead.

How does AI change the product manager's role?

Jeff Keyes says on DevOps Paradox episode 368 that product people now need to vibe code mock-ups to put in front of customers, run evals on AI components to judge the quality of their output, and orchestrate and interpret data themselves, all under a taste for what the product should be. He expects older roles, such as synthesizing feedback into a backlog or building grandiose roadmaps, to change or disappear.

What is the DevOps Paradox podcast?

DevOps Paradox is a weekly podcast co-hosted by Darin Pope and Viktor Farcic, covering DevOps, platform engineering, and modern software delivery. Episode 368 features Jeff Keyes of Allstacks on the AI productivity paradox: why individual developers feel faster while teams cannot show more value shipped, where the delivery bottleneck moved, and how product roles are changing. Every episode page carries audio, video, and a full transcript.

Episode Transcript

Read the Full Transcript

Share and Download

Guests

Jeff Keyes

Jeff Keyes

Jeff Keyes is Field CTO and Product Leader at Allstacks, focused on modernizing product management for the agentic era. AI made code cheap, and PMs are becoming builders: proving ideas with customers, handing engineering fully fleshed plans, and managing to outcomes. He drives Allstacks Product Studio, draws on 30+ years at Microsoft, Atlassian, Kaseya, and Captura (acquired by Concur), and hosts the Stacked Sessions podcast.

Hosts

Darin Pope

Darin Pope

Viktor Farcic

Viktor Farcic

Viktor Farcic is a member of the Google Developer Experts and Docker Captains groups, and published author.

His big passions are DevOps, Containers, Kubernetes, Microservices, Continuous Integration, Delivery and Deployment (CI/CD) and Test-Driven Development (TDD).

He often speaks at community gatherings and conferences.

He has published DevOps Paradox and Test-Driven Java Development.

His random thoughts and tutorials can be found in his blog The DevOps Toolkit.