Pace of innovation in the world of bits and atoms

Are people’s living standards improving? Silicon Valley says yes, but their story is at odds with what people are experiencing on the ground.

There’s been a lot of innovation in the world of bits and software, but not in atoms — real, hard engineering problems.

That was from “Notes from Peter Thiel’s speech at the National Conservatism conference on July 14, 2019“. Certainly we are deriving a lot of benefits from innovation in bits and software. If hardware innovation is lagging, perhaps that is because of a difference in project management cultures.

Silicon Valley tends to rely more heavily on interpersonal trust and a tolerance for trial-and-error learning. Hard engineering projects, however, seem to be run according to the kind of rigid processes taught by the Project Management Institute. There is a perception that the physical materials and machining are prohibitively expensive, such that small changes to plan are made to seem astronomical. The expectation is to get it right the first time. Thorough planning, tight controls, and error suppression are all regular parts of the hardware engineering project.

When looking back to the iPhone, for example, one sees quite a bit of innovation in the world of atoms. It would seem like magic to people just a couple decades before. But the iPhone largely recombined existing technologies. Projects in the Department of Defense rarely recombine existing technologies in that way. Instead, because the acquisition cycle is so long, projects seek to design special components based on technologies they expect to be around.

Perhaps one of the problems in the world of atoms is that project management fails to recombine science and technologies emerging from the labs. Instead, the requirements from various stakeholders seeks to direct the S&T activities. For example, the DOD is all excited about hypersonic vehicles and is now consciously accelerating development. If it weren’t for China and Russia, perhaps the DOD wouldn’t have recognized that the technology is within grasp.

Consider, however, that it may not have taken an external shock for the DOD to pursue hypersonics had it allowed for independent development of engines and airframes — for their own sake. Feasibility of new systems could then emerge from new components and architectures.

Breaking the tasks down into smaller components and moving judiciously along with the growth of knowledge seems to be the way software engineering has moved forward. This, of course, requires a certain culture of interpersonal trust and a tolerance for trial-and-error. That culture and related ecosystem of components is perhaps what is missing from the world of atoms.

Here was another interesting part from the notes on Peter Thiel’s talk:

Google is building artificial general intelligence at DeepMind. We need to be asking: Is this a military weapon? Is this a dual-use weapon?…

We now need to be asking: How many foreign intelligence agents have infiltrated AI and DeepMind at Google? Is Google’s senior management infiltrated by Chinese intelligence? Is it because Google is so infiltrated that they have decided to work with the Chinese military and not the U.S. military? Are they worried that their technology will be stolen through the backdoor anyway?