Alex Ray

  • Student

Submitted

Would you stay at a robot’s house? How Airbnb uses machine learning to drive product development
Alex Ray
Posted on November 13, 2018 at 2:08 pm
Over the past few years, Airbnb has developed machine learning algorithms that change how hosts and guests interact with listings. It's a major departure from their early human-centered design days. How should the young company balance data-driven design with human-driven [...]

Activity Feed

On November 15, 2018, Alex Ray commented on Boeing: Making Metal 3D Printing take-off :

This is super promising technology in the aerospace industry. I think an interesting use case would be eliminating interior, non-functional material in parts. That would reduce the weight, thus reducing fuel cost per flight.

3D Printed components have high variable costs, but low fixed costs. Do you think airplanes are made in small enough batches that the low fixed cost option makes more sense than a low variable cost (more traditional) option?

As Hanyin mentioned, this seems like an area that they need to invest heavily in — buy or invest in companies, get an internal group working on AM… they must do something to avoid the threat of disruption.

On November 15, 2018, Alex Ray commented on ADIDAS – Additive Manufacturing for the Masses :

Do you think Adidas will find a way to reduce production costs? AM seems to be low in fixed costs but very high in variable costs, especially the cost of the raw materials. Raw material prices seem to be outside of the company’s control — what other ways could they reduce costs to the point that shoes made with this technology rival the prices of shoes of similar quality?

Also, Adidas seems to be missing a major opportunity here. As you mention, they’re focused on standard shoe sizes. Why not go with people with big feet? The selection for shoes over size 13-14 (US) is extraordinarily limited.

On November 15, 2018, Alex Ray commented on The Beauty of Crowdsourcing :

This company surprises me. It seems to be full of contradictions. How can a MLM transition to a legitimate company? It seems to be doing that successfully. I’m also pleasantly surprised that open innovation works here. It feels like this would be a research lab-driven world, but open innovation, and especially the college collaborations, seem to be working well. Maybe beauty is a good fit for crowdsourcing: everyone is unique, and therefore need somewhat unique beauty products.

On November 15, 2018, Alex Ray commented on Amazon Studios’ experiment in crowdsourcing scripts: were they right to abandon open innovation in developing new content? :

This is such a popular post! Nice one. It’s cool that aspiring screenwriters had a chance for their biggest dreams to come true. While it’s a bit sad that this failed, I think it was inevitable. In fact, I think Roy Price was exactly wrong in your selected quote,“If you’re betting on one show, it’s got to be right down the middle of the fairway, but if you’ve got 14, you can allow people to try new things” [10].

With crowdsourcing, you tend to get the middle-of-the-fairway results. The things that people vote for the most are the ones that don’t extract extreme responses. Did you ever see the internet meme Unpopular Opinion Puffin (see link at bottom)? This is one of the few memes that have been banned by internet content aggregator Reddit, because of its lack of substance. People voted for opinions that everyone held. I posit that the Amazon Studios scripts that received the best ratings were for unsurprising, uninteresting scripts, and risk-tolerant taste-makers at Amazon recognized that.

https://knowyourmeme.com/memes/unpopular-opinion-puffin

On November 15, 2018, Alex Ray commented on Say “Helio” to the future of private investing :

Coming from a background with both VC (including some work with a CPG entrepreneur!) and machine learning experience, I am fascinated by this topic. I follow CircleUp’s CEO, Ryan Caldbeck, on Twitter, and have seen him fleshing out this idea over the past few years.

Previous commenter Emma is right — CircleUp will have the ability to undercut existing players through lower costs. Caldbeck has gone so far as to suggest “VC as a Service”, where the company could generate term sheets based on data. That could “disrupt” the VC’s. However, as you point out, this data-driven technique probably only works for CPG investors. Their investments have such similarity (especially in business models) that data is probably relevant from one firm to another.

I’d love to see what they have identified as the factors that correlate most significantly with success.

On November 15, 2018, Alex Ray commented on Machine learning in the Chemicals industry: Lyondellbasell :

Interesting stuff here. Your question – “what are the implications for proprietary data as companies partner with third parties for machine learning solutions?” – is an important one, too.

Recent technological advancements have commoditized machine learning. Now, basically any software developer with a basic understanding of statistics can build machine learning algorithms. Your article and the question paired with it imply that LYB should find expertise externally. Why not in-source the data science? This would protect their data, and potentially build even more of a moat between them and the competition.