Executive Overview

However, this unprecedented velocity brings a host of systemic challenges. Organizations face a massive spike in cognitive load, the erosion of traditional team dynamics, escalating technical debt, and profound ethical dilemmas concerning environmental sustainability and human well-being.

In a recent InfoQ Culture and Methods panel discussion hosted by Shane Hastie, industry leaders and engineering experts gathered to dissect these paradigm shifts. The consensus is clear: while AI acts as a powerful force multiplier for experienced practitioners, it threatens to exacerbate systemic flaws in organizations that never fully embraced foundational agility. As code becomes an abundant commodity, the software industry must pivot its focus away from raw output metrics and toward human collaboration, psychological safety, critical thinking, and sustainable engineering practices.

Detailed Chronology & Panel Discussions

The panel discussion brought together an international group of engineering leaders, sociotechnical architects, and principal engineers to explore how human systems are adapting to the AI era.

Introductions and the All-or-Nothing AI Trap

The conversation opened with round-robin introductions showcasing deep industry experience spanning from the 1980s to modern FinTech, MarTech, and government delivery roles. Vanessa Formicola, an engineering leader and sociotechnical architect, immediately flagged the industry’s tendency toward an "all-or-nothing" approach to AI adoption.

"We are all excited about it… but it’s very important to start considering what are the risks and to what extent what is appropriate in what circumstance, for which software, for which company, and for which group of people." — Vanessa Formicola

Formicola advocated for maturity frameworks in AI acquisition, assessing software risk profiles in the same deliberate manner engineering teams have used for decades rather than blindly applying AI to every codebase.

Psychological Safety, Cognitive Load, and Hyper-Velocity Engineering

Ben Linders emphasized that as collaboration patterns shift between teams and automated agents, psychological safety and transparency remain paramount. Linders noted that recent global disruptions have challenged team safety, making it difficult for engineers to feel secure enough to raise issues or question automated outputs.

Building on this, Philip Mortimer, a machine learning engineer specializing in document intelligence, highlighted the crushing cognitive load imposed by hyper-velocity engineering:

  • The Pull Request Deluge: GitHub predictions indicate a staggering leap from 1 billion pull requests annually to an expected 14 billion by 2026—a 14x increase driven largely by AI-generated code.
  • Context Switching: Engineers are increasingly forced to manage multiple concurrent work streams and agent swarms, buckling under the sheer volume of code review and context switching.

Rafiq Gemmail echoed these concerns, drawing a direct line to Jim Highsmith’s warning: "If you failed at agile, you will fail catastrophically at AI." Gemmail stressed that organizations skipping foundational agile practices—such as fast feedback loops and robust observability—are struggling to manage the ballooning risk of massive, AI-generated pull requests.

The Evolution of Teams: From Two-Pizza Teams to "Builders"

The panel debated the disintegration of traditional team structures. Once characterized by "two-pizza teams," cross-functional roles, and T-shaped skills, modern engineering units are increasingly shrinking into "one-person teams"—single humans working alongside swarms of AI agents.

Philip Mortimer pointed out that the boundaries between product managers, designers, and engineers are blurring into a unified class of "builders." Product teams can now commit code directly, while traditional engineers find themselves acting as platform architects—providing the underlying APIs and data primitives for human colleagues and AI agents alike.

However, this hyper-individualized model carries a dark side. Shawna Martell warned that stripping away multi-person teams removes the organic collaboration and human connection that make software engineering enjoyable:

"Part of the reason that we like to go to work is that we like the people that we work with. And I can’t imagine feeling that way about an agent."

Craig Smith echoed these fears, noting that organizations risk replacing genuine collaboration with an endless loop of AI-generated documents and code slop, ultimately losing the art of building great software together.

Career Paths, Junior Engineers, and the Death of Categorization

The transition from junior to senior engineer is undergoing a radical stress test. Without the traditional journey of writing code, making mistakes, and building foundational intuition, how will the next generation of engineers develop the judgment required for system scalability and maintainability?

  • The Senior Force Multiplier: Currently, senior engineers with deep domain expertise can use AI as a massive force multiplier.
  • The Plasticity of Youth: Philip Mortimer predicted that within six to twelve months, younger engineers—unburdened by legacy mental models—will master agent orchestration and "run rings" around older developers.
  • The Loss of Diversity: Rafiq Gemmail raised concerns that relying entirely on AI agents trained on homogenous corpuses will destroy personal diversity of thought within teams, leading back to toxic "rockstar" dynamics.

Despite these disruptions, Shawna Martell argued that the core attributes of leadership remain unchanged: "The folks that are leaders aren’t the ones who write the best code all the time… We’re looking for folks who can have a vision of what our product is supposed to be."

Vanessa Formicola agreed that while leadership values remain constant, traditional job categorizations—frontend, backend, QA, engineering manager—are effectively dead, replaced by a requirement for broad, cross-functional competence.

Supporting Context & Metrics

The panel’s qualitative insights align closely with broader industry trends and data points shaping the software engineering landscape:

  • The 14x PR Explosion: GitHub’s projected scale from 1 billion to 14 billion pull requests by 2026 underscores an unsustainable trajectory for human code review without automated governance and guardrails.
  • Work Slop and Productivity Paradox: Harvard Business Review data highlights that generative AI often intensifies work rather than reducing it, leading to widespread employee anxiety regarding job security. Furthermore, studies on "work slop" reveal that uncritical AI generation frequently creates more downstream cleanup work than value delivered.
  • The Trust Paradox: While AI adoption rates among developers remain exceptionally high, developer trust in AI-generated outputs continues to lag, creating an urgent need for robust verification, testing, and observability frameworks.

Official Statements & Industry Perspectives

The InfoQ panel captured definitive viewpoints from globally recognized engineering and culture experts:

  • On Agility and AI: "If you failed at agile, you will fail catastrophically at AI." — Jim Highsmith (cited by Shane Hastie and Rafiq Gemmail)
  • On the Shift in Engineering Roles: "We really just have this one class of people which [Mark Anderson has called] just builders. So we’re all builders now." — Philip Mortimer
  • On Environmental and Resource Ethics: "We’re burning tokens. We have a metric right in front of us that’s a proxy for the impact of the work we’re doing… Are we being cognizant of the fact that we’re coupling ourselves with this huge investment?" — Rafiq Gemmail

Future Outlook & Industry Wishes

As the industry looks toward the remainder of 2026 and beyond, the panel offered closing wishes focused on restoring balance, mindfulness, and humanism to software engineering:

  1. Mindfulness and Intentionality: Vanessa Formicola wished for greater mindfulness in how organizations wield AI, balancing business growth with respect for people and the environment.
  2. Mental Health and Peer Support: Ben Linders urged engineers to check in on one another, ensuring colleagues do not feel overwhelmed or left behind in the hyper-fast tech landscape.
  3. Rigorous Quality and Testing: Philip Mortimer called for a renewed focus on building robust, well-tested, and truly valuable software rather than simply chasing raw output velocity.
  4. Humanism and Compassion: Rafiq Gemmail and Craig Smith emphasized that the ultimate mission has not changed: serving real human needs. True innovation stems from human collaboration, empathy, and keeping humans firmly in the loop.
  5. A Return to Collaboration: Yinka Omole and Shawna Martell expressed hope that the industry conversation will shift away from exhausting output metrics and back toward creativity, teamwork, and enjoying the collaborative craft of software engineering.

As encapsulated by the Māori proverb highlighted by Shane Hastie—"He tangata, He tangata, He tangata" (It’s the people. It’s the people. It’s the people)—technology is merely a tool. The true measure of engineering success remains rooted in how we serve, support, and collaborate with one another.