Legal ML

  • Student

Activity Feed

On November 15, 2018, Legal ML commented on Additive Manufacturing in Construction :

This article is a great overview of the development and possible implications of 3D printing in construction. To your questions, I think it’s fair to always anticipate some regulatory pushback at the start. That sad, unlike automated vehicles where there is a clear likelihood — and recently, realization of the risk of death to consumers, that seems less likely here (though of course still possible during construction or if construction is shoddy, later on). Nonetheless, I’m very optimistic the regulatory environment would prove very favorable, especially given this technology has the potential to bring tons of manufacturing jobs back to the developed world. While some jobs, as you write, might be lost to machines, many will be won back from the developing world. I think this will mitigate some of the concern other commenters above expressed about pushback from unions (and, quite frankly, for better or worse, unions have the least influence they’ve ever had in this country).

On November 15, 2018, Legal ML commented on Beyond Bureaucracy: Open Innovation in the U.S. Government :

I think you identified the rub when you wrote that so far the open innovation program has avoided the particularly polarizing issues. The challenge with questions like this — organizational innovation in government — is it is merely a method of achieving a goal. But it doesn’t tell you what the goal is. Yes, if the goal is to tackle global warming or save social security, then this may be a far more efficient way of sourcing the best ways to do those things. But, excuse my cynicism, that’s not really the challenge in our government today. The challenge is agreeing whether we should fight global warming, or whether social security should die and be replaced. These are the principled questions that ultimately matter and no organizational or process change to operations can really change that. It’s important, no doubt, but won’t actually decrease polarization by itself. It reminds me of another oft-discussed process/operational change that’s been touted as the silver bullet of polarization: changing the process by which we draw districts (gerrymandering). But most political science literature says gerrymandering isn’t really the problem, though that would be great if it were. The problem is we tend to be a more polarized electorate that live, eat, school near people that think like us — and increasingly think different than others (think urban vs rural political divide that has grown, whereas both constituencies broke more evenly for both parties in years past). I think ultimately then that open innovation will help us come up with new solutions to problems we agree on tackling; but it won’t do much to help reduce our disagreement about which problems to tackle.

This is a fun example of one industry (perhaps unintentionally) borrowing the model of another. We often think of big pharma when we think of companies that outsource their R&D (big names buying the patents of little biotech startups), but it seems strange to see it in an industry we don’t associate much with product innovation (how much can potato chips, for example, innovate?) But as consumers demand more, this seems like one possible avenue to keep up. That said, I wonder long-term how this will play out. There’s certainly a cost to sacrificing control over your own creative process, or not investing as much in your own internal R&D, especially if in the medium-term you abandon this idea and find yourself with an R&D team depleted and atrophied. I also think — much like we saw with the marketing case about whether or not to have a creative director (top down product changes) — sometimes the customer does not even know well what he/she wants. That’s another risk of overindexing on external customer recommendations (as Ford famously said, his customer would’ve said he wanted more horses on the buggy!).

On November 15, 2018, Legal ML commented on Anticipatory shipping—retail’s crystal ball? :

This is very cool and almost has a minority report feel to it — predicting our behaviors ahead of time. Exciting, scary. It also relates a bit to some of the supply chain challenges we’ve seen throughout the course, including this week with the beer game. It’s counterintuitive on one hand to think of Amazon guessing when and what you will order, potentially wasting time and money shipping something you ultimately don’t want (and may need to return or will just keep and demand no charge). But in reality, after having studied some TOM, we see that holding inventory has its own (often less immediately apparent) cost. Further, there is a qualitative aspect to being able to service customers more quickly (and Amazon already does this to some degree with subscription services and one-click purchases). Finally, in a real way, this is merely an extension of existing inventory management logic; as the beer game showed, forecasting is inherently an educated guessing game so this new Amazon move isn’t as risky (or, if you like, even as revolutionary/novel) as it seems at first read.

On November 15, 2018, Legal ML commented on MACHINE LEARNING IN RETAIL: AN EDITED APPROACH :

Thanks for sharing this trend; I think it’s an important development in an industry that needs to evolve. I recall speaking to the founder of an image classifier about the possibilities of ML in retail as well and lots of ML engineers seem to agree this space is ripe for automation. My first question is more a business strategy one: how do companies in an industry being squeezed find the courage to actually invest even more in a new development like retail machine learning? I wonder if companies will be able to measure the value enough at a time when they are looking to cut costs and just survive. Moreover, I wonder how companies like EDITED will manage different clients whose goals are competitive, as well as interests of — and possible improvement in welfare of consumers. For example, will I scrape one client’s data to help another? Will I distort recommendations (“right product at right time”) for commercial reasons because my client happens to, say, pay a premium over another company. The latter question has some interesting ethical, strategic and even legal (antitrust) implications that have been explored already in the case of SEO and companies like Google providing hits to their own products over others.

On November 15, 2018, Legal ML commented on Flying High: GE’s Billion Dollar Bet on Additive Manufacturing :

This article is relevant for two seemingly opposite reasons, brought together in this example: additive manufacturing is one of the most innovative developments in industry today — and GE is one of the old guard that seems unable to keep up. I’d be curious, like the author, to see where additive manufacturing makes sense and where it doesn’t. On one hand, I’m inclined to agree it makes most sense in production of complex items that would otherwise require too many disparate parts. That said, one thing we learned early on in TOM was that parallel assembly lines can be created to save time when a product is modular — say, a blender that can be built in different parts simultaneously and then brought together at the end. The interesting thing about 3D printing is that in some ways it seems organizationally less advanced (even if technologically more advanced) — like a job shop building a custom baseball bat where the same piece of wood needs cut, shaped, stained, imprinted, etc. all sequentially without being able to do any of those steps at the same time.