The article looks at the benefits of distributed computing and makes the case that similar benefits could apply to distributed working (eg. scalability). Then i

The article looks at the benefits of distributed computing and makes the case that similar benefits could apply to distributed working (eg. scalability). Then it is only fair that we also look at other aspects of distributed systems and try to identify constraints on this model. I can think of: 1. Just because a system is distributed does not mean you have a lot of very small machines (eg. a system with 10 machines each with 1MB RAM will beat one with 1000 machines each with 10KB RAM). Similarly, having 5-10 distributed offices could net all the benefits mentioned in the article while avoiding costs of having 500 employees working from home/wework/... 2. Just because your system is distributed, you don't locate one machine on each continent. Machines are still arranged in proximity and having cross datacenter communication significantly erodes performance. Similarly, having distributed offices in far away timezones (eg. US west coast and India) imposes tremendous costs on collaboration. Having office in nearby timezones would be much better (eg. Seattle, Denver, SF). One particular case where far-away timezones help is ease of having 24-hour oncall support. 3. Different machine profile in your system, or different composition of your clusters, makes it harder to tune the system for performance or quality. Similarly, having employees in vastly different jurisdictions would mean you have additional costs on compliance / regulations / etc. 4. When you are prototyping or bootstrapping a new product, you do that quickly over a small setup (one machine). Similarly, it might make sense to bootstrap your startup in your garage (or a single city) until it is ready to scale.