World's Fair 2026
Everything we knew about software has changed — Theo Browne, @t3dotgg
Overview
The talk argues that advancements in AI models have fundamentally changed software development, shifting the paradigm from incremental improvements to a need for developers to think bigger and wider. The speaker posits that current developer workflows and ingrained opinions about tools and languages are becoming obsolete as AI capabilities rapidly surpass human development speed. This necessitates a reevaluation of what is possible and how we approach building software.
Who should watch
- AI Engineers
- Product Managers
- Builders and developers
- Those feeling overwhelmed or experiencing "AI psychosis"
- Individuals questioning traditional development workflows and tool choices
Key takeaways
- AI models have progressed through distinct eras: tool calling (Sonnet 3.5), long-running task completion (Opus 4.5), and orchestration with self-awareness (Mythos).
- Developers often cling to outdated tools and workflows due to ingrained habits and a "skeuomorphic" design approach, mirroring early iOS interfaces that mimicked physical objects.
- The rapid improvement of AI models means they are advancing faster than human developers, requiring a shift from personal skill improvement to building larger, more ambitious projects.
- Many complex software products or services that previously required significant engineering effort can now be achieved with simpler tools, even just a markdown file executed by an AI.
- The concept of "too big" to build is shrinking, with capabilities like building a database platform into a product becoming feasible in days.
- Developers should focus on building wider, covering a broader spectrum of functionality, rather than solely competing on depth within a narrow niche, as users can extend functionality themselves.
- The current development landscape encourages building platforms that allow users to integrate AI agents and features, similar to how Slack has become a de facto platform.
- Ideas that do not feel a bit "stupid" or overly ambitious may not be large enough to leverage current AI capabilities.
Notable quotes
*The models are getting better faster than we are.*
*We're currently in our skeuomorphic phase as software developers.*
*If your idea doesn't feel stupid, it's cuz your idea's not big enough.*
Unofficial community note. Prefer the recording for nuance.