# Four Predictions For How AI Will Change Software in 2026 Source: https://www.youtube.com/watch?v=srnTvlmTq3U Recap page: https://rapidrecap.app/video/srnTvlmTq3U Generated: 2025-12-31T16:03:31.542+00:00 --- ## Quick Overview The speakers predict that by the end of 2026, AI progress will still be fast, leading to a significant shift where AI agents become commonplace, drastically changing software development, but they also foresee that AGI will likely not be achieved by then, emphasizing that current AI systems still require human oversight and correct prompting. **Key Points:** - The speakers predict that AI progress will remain very fast through the end of 2026, causing rapid changes in software development. - They predict 2025 will be the 'year of agents,' where AI agents become a major development, fulfilling predictions made by leaders like Dario and Sam Altman in 2024. - The major challenge for software engineers in 2026 will be adapting to AI-driven development, specifically moving away from writing boilerplate code toward higher-level tasks like agent orchestration and prompt engineering. - The emergence of sophisticated AI agents, capable of performing complex tasks like writing and submitting PRs, will fundamentally alter the role of software engineers. - The speakers believe that while AI can generate code, it currently lacks the 'common sense' or comprehensive understanding needed for complex, novel software architecture, meaning human engineers remain essential for high-level design. - They predict that the industry will see a split between traditional engineers who resist AI integration and a new class of developers who embrace agentic workflows. - They definitively predict that AGI (Artificial General Intelligence) will not be achieved by the end of 2026. ![Screenshot at 00:07: The speaker emphasizes the mind-blowing pace of AI progress by comparing where we are now versus where the technology was just one year ago, setting the stage for their 2026 predictions.](https://ss.rapidrecap.app/screens/srnTvlmTq3U/00-00-07.jpg) **Context:** Dan Shipper (CEO of Every) and Brandon Gell (COO of Every) discuss their predictions for the future of software development and Artificial Intelligence, specifically focusing on the timeline for achieving AGI and the impact of AI agents on engineering roles by 2026. They reference past predictions from AI leaders and contrast the current state of AI tools with the capabilities required for true general intelligence. ## Detailed Analysis The discussion centers on concrete predictions for AI's impact on software engineering by 2026. The speakers firmly believe that AI progress will continue at a very fast pace, validating earlier predictions (specifically referencing Dario and Sam Altman in 2024) that 2025 would be the 'year of agents.' They anticipate that by 2026, AI agents will be sophisticated enough to handle tasks like debugging and submitting code changes, fundamentally shifting the role of software engineers. The key challenge identified is that while AI excels at generating routine code, it still lacks the holistic understanding or 'common sense' necessary for complex, foundational architecture, meaning engineers must adapt to orchestrating and prompting these agents rather than writing boilerplate code. They project a split in the industry: traditional engineers who resist AI adoption and a new class of developers who embrace these agentic workflows. Despite the rapid advancements, they explicitly predict that true AGI will not be achieved by the end of 2026. The speakers note that their company, Every, has already seen massive growth (doubling business from June to a certain point, reaching $150k MRR) by adopting these new AI-assisted methodologies, which they find more exciting than prior development successes. ### 2026 AI Predictions - AI progress remains fast; 2025 is the year of agents - AGI will not be achieved by the end of 2026 - The core challenge is adapting engineering workflows to be AI-native. ### Impact on Software Engineering - Engineers must shift from writing code to architecting and prompting AI agents - Traditional software engineering skills risk atrophy if engineers refuse to adapt. ### Every's Experience - The company doubled its business in a short period and experienced rapid revenue growth (up to $150k MRR) by adopting AI-first development practices, which they found more exciting than previous successes. ### AI Capability Limitations - Current AI agents, while powerful for routine tasks, lack the necessary common sense for complex architectural design or handling novel problems without human oversight. ### Future Workflow - The future involves a paradigm shift where AI agents handle the bulk of coding, leaving humans to focus on higher-level design and instruction. ![Screenshot at 00:02: The title card displaying the episode's focus: "2026 Predictions."](https://ss.rapidrecap.app/screens/srnTvlmTq3U/00-00-02.jpg) ![Screenshot at 00:05: Speaker emphasizing the rapid, mind-blowing pace of current AI progress.](https://ss.rapidrecap.app/screens/srnTvlmTq3U/00-00-05.jpg) ![Screenshot at 00:15: Graphic illustrating the complexity of AI development, showing a hand holding a puzzle-piece brain amidst classical architecture and circuit diagrams.](https://ss.rapidrecap.app/screens/srnTvlmTq3U/00-00-15.jpg) ![Screenshot at 00:37: Visual comparison of AI leaders Dario \(from Andreessen Horowitz\) and Sam Altman \(from OpenAI\) being referenced in the discussion about 2025 being the year of agents.](https://ss.rapidrecap.app/screens/srnTvlmTq3U/00-00-37.jpg) ![Screenshot at 07:00: Speaker holding up a flame-shaped object, humorously referencing a 'fire' to describe the rapid acceleration of AI adoption.](https://ss.rapidrecap.app/screens/srnTvlmTq3U/00-07-00.jpg)