Introduction

AI is changing so fast that old growth rules stop working. This story shows why you must learn new habits to keep up.

What's the problem it solves?

The article explains why old growth methods fail in fast-moving AI companies and what new skills and habits are needed to keep up.

Quick Summary

The writer joined an AI company called Lovable and quickly saw that almost everything she learned in past growth roles no longer works. AI tech and customer needs now change so fast that steady plans, old patterns, and slow playbooks fall apart. Instead of tuning funnels or running small tests, she now spends her time inventing new ideas, new product loops, and even new features.

Most classic marketing channels do not drive real growth for AI products right now. Word of mouth, creators, and community matter far more. Growth also happens deeper in the product itself because many AI tools look the same on the surface. The real difference comes from how smart and helpful the agent is. This pushes growth work into model choice, reasoning quality, and agent behavior.

The job roles inside AI companies are also blending. Designers, marketers, product people, and engineers all cross lanes. Plans shift every few weeks, not every quarter. You have to move fast, learn fast, and let go of your past ideas even faster. PMF is not stable in AI, so growth teams have to balance scaling today with constant change tomorrow.

Key Takeaways

• Old growth frameworks break in fast-moving AI companies.
• Real growth comes from product quality, word of mouth, and community, not old channels.
• PMF changes often, so growth is never stable.
• Roles blend and everyone must work across boundaries.
• Short plans and fast learning beat long plans and heavy process.
• The winning skill is letting go of old patterns and building new ones quickly.

What to do

• Plan in short 2 to 3 week cycles.
• Focus on learning faster, not being comfortable.
• Ship ideas quickly instead of polishing them forever.
• Watch for new distribution channels and test them early.
• Treat your current mental models as temporary.
• Work across roles instead of staying in your lane.
• Put more effort into agent quality and product depth, not just surface tweaks.

‍