This “default alive” advice is repeatedly shared. One thing it obviously does not address is the human element of who you cut and whether they will be “default alive” unemployed in a recession. A huge amount of YC advice in general positions founders as protagonists and employees as NPCs then are shocked people pick Google over their startup offer. Funny thing is I’ve seen this exact advice destroy a company. In Marc…
That's just bad software engineering. In general, I would expect software to scale well as far as traffic is concerned esp in today's environment where scaling infra is not really a big deal. Coping with feature requests on a shoestring staff is a different story.
For example, one outage related to Google Ads API changing their parameters. This led to ads not being run which directly cost revenue since those ads were profitable. The outage went on for much longer than it needed to since people with expertise on the marketing pipelines were gone.
Id say the “bad engineering” here is mostly Google ads who should version their API changes. But if Google can’t do good engineering, I’m not counting on too many other companies to do so. This idea that infra is a “set and forget” operation because autoscaling exists is a fantasy for conference talks, not reality.