I think this would be better phrased as "don't start a [product] business". A services business can't be easily reframed as a side project and doesn't carry the same risks of product-market-fit etc.
That's fair. The point I was trying to make was that it's better to start any new thing in small stages, because you don't know what you don't know. For service businesses, the default mode of operation is starting as a side project - for example, starting a marketing consulting service.
Of course none of this applies to things with default huge capital expenses like starting a restaurant so it's not an exhaustive observation. Mostly stemming from my own experiences with software product dev.
