3 ms·

It gets even more interesting when you think that by definition, our economy is supposed to be a long term win by having completely separated companies building

• cinquemb 5y ago
  Coordination and consensus can have non linear costs, and will vary with the amount complexity of the decisions that need to be made.
• simonh 5y ago
  Competition in an economy works because the market picks the winners and allocates resources to them. Often that means several winners because different competitors have complementary strengths and weaknesses. Nobody knows their requirements and price tolerance better than the customer. Incentives of the producer and customer are aligned. Inside an organisation there’s no real mechanism equivalent to a market within which these alternatives can compete. It almost entirely comes down to politics, not relative cost, features and technical suitability. If a team’s solution is adopted by another department does their revenue go up? Do they get extra resources? How do they advertise their product? What’s their incentive to provide customer service and support? If they get a new project or strategic direction and this ‘product’ doesn’t fit their vision anymore what’s to stop them just dropping it? After all none of their budget actually comes from this ‘product’ no matter how many other teams use it. It’s not that these problems can’t be solved, but there’s no real market or real revenue stream there so market forces have to be consciously emulated by management fiat. If enough layers of management lose focus on that, or management changes, it all breaks down. It’s basically the same problem as with planned economies.
  • dv_dt 5y ago
    With internal completion aren’t there internal customers, or even external customers too? I think you bring up some interesting points, but think there’s some deeper considerations than market vs planned. For instance at a macro scale, China has a hybrid approach and for many planned industrial growth areas have been successful from furniture to solar to being on their way with chips. Planned investment is arguably how many companies work very successfully. I’m not arguing for all planned control, but suspect some set of features that make some competitive approaches work well and some planned approaches work well.
    • lmm 5y ago
      > With internal completion aren’t there internal customers, or even external customers too? Internal "customers" are generally spending someone else's "money" and their own success or failure is often more dependent on politics than their merits. Competition works if your success metrics are measuring the right thing. But internal popularity often isn't.
• mirker 5y ago
  It’s risk vs. efficiency. Duplicating work lowers risk at the cost of efficiency. Without redundancy, the market is unstable. Look at the pipeline shortages of the past year. A chip shortage screws up many industries because all eggs are in one basket. Farmers have government incentives to overproduce for to prevent food instability. A company always has competition due to promotions, by the way. Competing at a team level makes sense for R&D, since products can end up different. But I’m not sure how it makes sense for a big, resource heavy project.
• puchatek 5y ago
  Yep. If they had all coordinated their efforts then they might have settled on some inferior design and stuck with it for longer than necessary. That third group might not have been around to look into that CMOS stuff