This is a really good point, Chris! Sometimes I get into a bit of a music rut where I’ll listen to a lot of top 40-ish music and want spotify to redirect me to more interesting songs, but instead it gives me more Bieber, and I further disengage from the service 🙂 Definitely a real flaw.
KT – what a cool concept! I’m curious how they’ve designed this product to work in emergency situations. For example, does it run without electricity, which I would imagine is a constraint that exists in most emergency response scenarios? How accessible is the real-time transfer of design files in the (presumed) absence of internet immediately after a catastrophe? Or is this product meant to serve communities slightly after the “first responder” scenario?
In any case, it is incredibly interesting and seems to hold a lot of promise for both disaster-struck communities, as well as rural / isolated areas that lack easy access to important goods on a more sustained basis.
Nice article! An additional application of ML I’ve heard of within a security context is algorithmic processing of would-be passengers’ facial expressions to detect nervousness, aggression or other signs that they be in the process of committing a crime. I wonder if this sort of approach may be more prone to adoption, as it could (at least initially) be run in parallel with current bag / document screening procedures, to provide an extra layer of security rather than a replacement for current practices. The AI bag screening could also likely run in parallel to a human-backed check, which to me also seems like a more likely path-to-market than an immediate replacement. Unfortunately, both of these approaches, while perhaps decreasing false negatives and improving safety, do nothing to shorten our wait times in the TSA line. Guess I’ll renew my Pre-Check for now 🙂
Love this! It’s fascinating to hear that Duolingo is working on applying their technology to other subjects. I wonder how much value machine learning will add in other areas of study, or whether the complexity of language learning is uniquely fitted to ML. For example, in elementary math, it seems intuitive that counting must come before addition which must come before multiplication (whereas, as you mentioned, this linear relationship doesn’t exist as clearly in languages between adverbs, past tense words, etc). Regardless, any level of increased individualization in this sort of highly-accesible learning platform will be great to see!
Super interesting. Full disclosure – I know very little about baseball – but I wonder how this approach to pricing would affect revenue for teams when they perform poorly, or when other factors affecting the algo like weather underperform. Assuming that fans have heterogenous preferences on what makes a game “worth it” (good weather, strong team performance, etc), and that they can’t predict how these factors will play out in the future, they may wait for prices to go down before they make purchases, reducing overall revenue for the team – for example, if I care about team performance but not weather, but I assume the “price” of a good weather ticket is baked into the pre-season price, under dynamic pricing I may wait until I see good team performance and bad weather so get my ticket at a discount. Even putting preferences aside, any user that expects a team to do worse than the pricing reflects would be better off waiting to buy their ticket. Have we seen examples of how dynamic pricing plays out when a team has especially bad scores, poor weather, etc?
Great article! It is interesting to see ML-driven recommendations applied in a less ‘triable’ product type than what we may be used to in media/digital-driven plays (Netflix, Amazon, Spotify, etc.) – whereas I could easily listen to 15+ songs in an hour of Spotify, and therefore generate 15+ data points to feed their model, StichFix only gets a handful of data inputs per month. Anecdotally, I’ve observed that StichFix does tend to stick to one, fairly narrow aesthetic in their recommendations – I wonder if this may be due to the fact that their model doesn’t have enough data to develop truly nuanced user ‘types’ that deviate meaningfully from their core users’ preferences (e.g., the model just converges towards the mean user, causing other users to drop out because they don’t like their clothing selections, which further reinforces homogenous user preferences in their model). This would be another argument supporting your recommendation for greater human curation!
Very interesting article, Russel!
While suicide prevention seems like a valid use case for Facebook AI, counter-terrorism comes off to me as a bit of a post-facto rationalization for extreme levels of data collection on the part of Zuck. Terrorism prevention is a role that has typically sat in the hands of governments – for good reason – and it’s not obvious to me why Facebook would be better at collecting, interpreting and (especially) acting on related data than the many public agencies that are currently dedicated to this purpose. Given that Facebook has adamantly protested requirements to share data with the NSA and similar organizations, I wonder how the company intends to act on potential threats they may identify? They can cut off information flow, but this introduces the additional ethical dilemma of whether they have a further responsibility to report possible attacks to those actually equipped to enact a more robust prevention plan (i.e., the military). It will be interesting to see how this plays out going forward!