Smaller teams equipped with AI tools can now match or exceed the productivity of larger developer groups. AI assistants reduce the need for outsourcing, lower development costs and help maintain in-house ownership of code. But it’s important for SMBs to understand the limits of today’s AI coding assistants or risk wasting their AI investment. That’s the experience of Mike Stone, co-founder of customer web and mobile development firm The Gnar Company, whose clients include the state of Massachusetts, Grubhub and AARP. With AI coding tools, Stone said the playbook has changed to this: Hire two exceptional senior developers; Equip them with AI tools; Watch them outperform entire teams; Ship product code from ‘Day One.’ Particularly salient is the latest trend called “vibe coding,” which is a new, more intuitive way for people to write computer code using natural language — like how one would talk to a friend. Instead of writing complex code in a particular syntax, users just need to describe what they want the software program to do, and an AI model helps turn that into working code. This speeds up the creative process and lets users focus more on ideas and less on technical details. Santiago Nestares, co-founder of DualEntry, told that his company was able to build an enterprise resource planning (ERP) system like NetSuite despite being told that they would need at least $100 million, which was out of reach. The company used ChatGPT daily to pressure-test its system designs, validate architectural decisions and explore edge (or one-off) cases. It used Cursor to write and refactor code with LLMs embedded into its development workflow. It also used AI to help run automated code reviews and maintain quality without adding management layers or quality assurance processes. “With a team of just 11 people and nine months of focused work, we’ve built a fully capable ERP that has feature parity with NetSuite,” Nestares said. “AI made that possible. If we had to hire for all the knowledge we now get from AI, we’d need a team two to three times the size,” Nestares added. “AI is the great equalizer.”