Software is DEAD. SaaS is DEAD. AI has solved EVERYTHING!

Summary

The speaker argues against claims that “software is dead” or that AI has made software development obsolete. AI has lowered the cost and effort of writing code, enabling individuals and small teams to build personal tools and products more quickly. In the speaker’s view, that means more software is being created—not that software is disappearing.

Key arguments

• Personal tools and micro-SaaS are more exposed. If someone needs only one feature from an expensive product, they may now build a small alternative themselves. This puts pressure on narrow, easy-to-copy products, though larger, dependable products can still offer value through breadth and reliability.
• Writing code is only one part of development. Understanding business needs, choosing architecture, planning, testing, reviewing, maintaining, and supporting software remain substantial tasks. Enterprise products add further complexity, including multiple stakeholders, integrations, security requirements, compliance, and operational risk.
• Cloning software does not reproduce the whole product. A clone may lack ongoing maintenance, security updates, a roadmap, integrations, support, and compliance. Organizations may prefer a paid vendor rather than take responsibility for maintaining a clone themselves.
• AI can speed up teams, but introduces new bottlenecks and costs. The speaker says their AI-focused agency delivers faster and may need fewer developers, but code review, testing, quality control, and security still require human judgment. They cite a Harvard study reporting a 35% increase in pull requests and a 56% increase in average time from submission to approval after AI-agent adoption. AI tools also add costs, which the speaker says may rise as usage expands.
• High-stakes software requires particular care. Drawing on work on a pharmaceutical management system for dispensing and manufacturing medication, the speaker describes a multi-year effort involving developers, product owners, project managers, and subject-matter experts. In that setting, bugs can have serious consequences, so generated code cannot simply be accepted without review and testing.
• The likely result is a software boom—with more low-quality output, too. Easier development may produce more products, clones, and unfinished or poorly maintained systems. The speaker expects some organizations to need experienced developers to repair, support, or replace these efforts. Their core distinction is: code may be cheap, but quality software is not.

This is an opinion and industry analysis, not a tutorial. The speaker also discusses Adobe Experience Manager from personal experience, criticizing the product while arguing that large companies are unlikely to abandon deeply integrated enterprise systems for clones alone.

Main speaker and sources

• Main speaker: Dorian Develops, drawing on experience in agency work, enterprise software, and a pharmaceutical software project.
• People and sources discussed: Peter Levels, Simp for Satoshi, Nikita Bira, a cited Harvard study, and comments attributed to the Replit CEO and Elon Musk. The speaker also mentions GitHub cloning and reverse-engineering projects, CodeRabbit, Copilot, Adobe Experience Manager, and Beehiiv.