Four Predictions For How AI Will Change Software in 2026
Timestamps are as accurate as they can be but may be slightly off. We encourage you to listen to the full context.
Podcast Summary
In this insightful conversation between Dan Shipper and Every COO Brandon Gell, they reflect on their company's remarkable growth trajectory in 2025 and share predictions for how AI will reshape software development in 2026. (01:00) The discussion covers Every's evolution from a flat revenue curve to doubling their business in six months, growing from the 60-70K MRR range to about 150K MRR. They explore how agent-native architectures will fundamentally change software engineering, the emergence of designers as AI power users, and the splitting of software engineering into distinct camps. (11:55) The conversation also touches on concerns about AI-generated content in elections and the timeline challenges facing true AI autonomy.
- Main themes: AI-driven business transformation, the future of software development, agent-native architectures, and the evolving role of different technical professionals in an AI-first world
Speakers
Dan Shipper
Dan Shipper is the founder and CEO of Every, a platform that has evolved from a newsletter bundle to an AI-first business offering software tools and consulting services. He has been building Every for six years and is known for his insights on AI development and business strategy in the AI era.
Brandon Gell
Brandon Gell is the COO of Every who joined the company in early 2024 and has been instrumental in their growth trajectory. He previously scaled a business to 100 employees and brings operational expertise to help Every navigate its transition from a 5-person to 20-person team while maintaining their culture of genuine innovation.
Key Takeaways
Agent-Native Architecture Will Revolutionize Software Design
Dan Shipper introduces the concept of "agent-native architectures" with three fundamental levels: anything a user can do, an agent can do; anything the code can do, an agent can do; and anything a developer can do, an agent can do. (11:55) This represents a paradigm shift from traditional software where agents are limited in their capabilities within applications. The context behind this insight comes from observing how tools like Claude Code have reached a point where coding agents consistently produce working code, unlike previous years when they would inevitably break. This architectural approach enables users to have natural conversations with their software to modify, customize, or extend functionality in real-time, making every app potentially agent-controllable and infinitely customizable.
Designers Are Emerging as AI Superheroes
Brandon Gell predicts that designers will become the next major beneficiaries of AI advancement, as they possess taste and vision but have historically been constrained by having to convince developers to implement their ideas. (17:02) He cites Lucas, their creative director, as an example of someone who has transformed into a "machine" capable of both incredible design work and "vibe coding" small applications that enhance his workflow. This shift occurs because code has become cheap through AI, allowing visually-oriented creative professionals to build complete experiences without traditional programming barriers. However, the challenge lies in creating interfaces that don't intimidate designers who may be afraid of touching code directly.
Software Engineering Is Splitting Into Three Distinct Camps
The software engineering profession is evolving into three primary categories beyond traditional non-AI engineers: traditional engineers using AI as an accelerant, vibe coders who create without understanding code, and a emerging third group called "compound" or "agentic" engineers. (23:22) This third group represents engineers who have completely reinvented their skill set for an agent world, moving up the stack to manage AI agents rather than writing code themselves. They've made the deliberate trade-off of abandoning hands-on coding in exchange for mastering agent management and direction. This represents a fundamental shift in the profession, where growth will primarily occur in this new engineering paradigm while traditional approaches slowly decline over time.
AI-Generated Content Poses Electoral Risks
Brandon Gell expresses concern about the proliferation of AI-generated content on social media platforms without proper labeling, particularly as it relates to upcoming elections. (19:42) He shares an example of his mother refusing to believe that a realistic video of gorillas caring for their children was AI-generated, highlighting how sophisticated these creations have become. The concern centers on the potential for deepfakes and manipulated content to influence political processes, especially since social media companies heavily invested in AI aren't requiring "made by AI" labels. This could lead to regulatory requirements for content labeling, particularly if problematic content emerges during midterm elections.
True AI Autonomy Remains Further Away Than Expected
While AI has made remarkable progress, both speakers agree that Artificial General Intelligence (AGI) is more complex and distant than initially predicted by industry leaders. (31:56) They note that major AI CEOs like Sam Altman and Dario Amodei have shifted their timelines, moving away from aggressive predictions about AGI arriving within 1000 days. The challenge lies in creating AI systems that can operate autonomously for extended periods without human intervention, requiring continuous learning, goal modification, and the ability to handle the accumulated quirks that emerge over time. True autonomy demands that AI systems develop their own sense of agency and individual decision-making capabilities, which current alignment training actually works against by prioritizing predictability and compliance.
Statistics & Facts
- Every's monthly recurring revenue (MRR) doubled from the 60-70K range to about 150K MRR starting in June 2025, representing their biggest growth breakthrough after five years of relatively flat performance. (02:43)
- The company grew from 5 people at the beginning of 2025 to almost 20 employees by the end of the year, representing a 4x increase in team size. (06:22)
- After appearing on Lenny's podcast in June, Every experienced 20% growth that month, breaking their historical pattern of 1-5% monthly growth rates. (04:17)
