I thought this was an interesting read, but I’m conflicted on whether this use of ML makes sense for Sephora. On one hand, ML reduces the seemingly infinite amount of choices for makeup, and it seems like a genuinely good user experience for someone trying to buy makeup. On the other hand, it seems really creepy to have an ML program (which is basically an excellent pattern matcher) decide how women should look. We already have a huge amount of social norms for women’s appearances, and an ML program won’t be able to discriminate between what is healthy or unhealthy. After this week’s Dove case, it’s hard not to feel skeptical about Sephora’s approach.
This was a cool article. The processes in challenge.gov reminded me a bit of code.gov, which is a federal repository of computer code that agencies can share between each other. The idea is that the federal government is large enough to attract software solutions to big problems, but then these solutions can be spread out to smaller parties that may not be able to attract the same talent. Without the federal government’s help, I’m not sure smaller governments will have the resources to create programs like challenge.gov. Even if smaller governments can create these programs, it might only work for larger cities; I really like what Mexico City was able to do, but I’m curious if this strategy could work as well in a smaller city like Wichita, KS.
Thanks for writing this! I usually think of 3D printing in terms of smaller projects — i.e. prototypes of consumer goods — but you make a good case for larger scale projects. I was surprised that the construction time can be cut down so quickly. I had assumed that most elements of construction are quite basic — beams, arches, etc — so its development time would have been minimally low, but it looks like there may be ways to create buildings faster. With housing prices getting higher and higher, this might be an interesting way to dramatically increase supply of housing in a short time frame.
I thought your point about shipping was particularly good. We tend to look at 3D printing as this groundbreaking new technology for its advances in materials, prototyping, and design, but being able to move a factory closer to the point of sale may have a bigger effect on the bottom line that the more commonly discussed advanced. This point would be an interesting addition for classes in the future when people discuss the effects of automation on our economy. It’s usually focused on questions of labor, but we rarely consider questions of shorter supply lines.
Thanks for posting. I think OK Cupid is an excellent application of machine learning. Your point about unintentional misinformation is especially good; if we don’t know ourselves well, we might enter inaccurate preferences, but with enough data, a machine learning algorithm should be able to figure out which stated preferences tend to be reliable. It’s also worth noting that the downside here is relatively limited, which makes dating a good application for a technology still in its early phase. I care less about a bad date than I do about a misdiagnosed disease.
Interesting article. I think open innovation makes a lot of sense for a large-scale consumer goods company like Pepsi. I’m curious how they can keep control of the narrative (i.e. avoid a “Boaty McBoatface” scenario) when the people voting on ideas don’t have any skin in the game for the actual product. The incentives seem more aligned in R&D, though as one other comment mentioned, I think Pepsi will need to retain their own competencies in R&D. In either case, crowdsourcing offers a low-cost method of product development with a potentially high reward.