Jon and Adam discuss whether the AI industry is heading for a dramatic crash or a slower deflation. Jon argues that as AI makes software cheap enough for anyone to build, SaaS companies will face relentless pricing pressure and the bubble will quietly shrink rather than burst. Adam pushes back, suggesting it's not one bubble but many smaller bubbles making up one large one, and that large platforms like HubSpot will be forced to pivot as smaller businesses become self-sufficient.

They explore the implications for software jobs, the dangers of moving too fast without review, and whether AI's rapid progress is outpacing humanity's ability to structure it responsibly.

• Why the AI bubble won't burst like the dot-com crash, but may slowly deflate as software becomes cheaper to build
• The HubSpot problem: when a small business owner can build a custom CRM over a weekend for £20, what happens to platforms charging £1,000+ a month?
• Adam's "mosaic of bubbles" argument: it's not one big balloon, it's many smaller ones being pricked from below
• How smaller businesses could become self-sufficient, cutting out the SaaS middleman entirely
• The snake eating its own tail: AI companies funding the tools that undermine their own customers' willingness to pay
• Why "pivoting" might not save legacy platforms built on monolithic codebases
• The Klarna cautionary tale: cutting staff for AI efficiency, then rehiring when customer satisfaction collapsed
• Where the real danger sits: people moving too fast, trusting AI output without review, and creating security vulnerabilities
• Why AI is a productivity tool, not a replacement for human judgement