• Alumni

Activity Feed

On November 20, 2016, stong commented on TED – A Company Worth Spreading :

While I completely appreciate TED’s efforts to “democratize” knowledge and information, I am also a bit wary of the downsides associated with rapid over-proliferation of TED content. TEDx is a wonderful way to tap into communities and individuals around the world, but it lacks the “curation” element of the original TED talks, which vetted, screened and evaluated the quality of the content presented. Because of the greater variability in the quality of TEDx talks, I have found that content to be less valuable and brand-eroding. In that vein of thought, I wonder if Carlos’s suggestion of “dynamic iteration” between speakers and audience can help maintain a certain level of quality, especially in community generated content. Perhaps TED can explore other elements of “crowd-sourcing” to improve access and quality of their content moving forwards. (LTKI for short henceforth) has clearly managed to solve a pain-point for Instagram influencers (and followers) but I, like the other commenters, worry about its long-term sustainability given its over-dependence on Instagram. To start, LTKI was created as a workaround for Instagram’s restriction on posting external links in captions. From a technical perspective, enabling links in posts is an extremely simple feature that Instagram developers can implement in days, if not hours. In the past year or so, Facebook has been ramping up on developing an ad platform for Instagram, and I imagine that it’s only a matter of time before they offer an elegant way for influencers to share affiliate links on their posts.

On November 20, 2016, stong commented on Evaluating New Drugs with Wearable Technology :

As an early-adopter of wearables, I have always been bullish about “quantified health” and its potential to radically improve scientific research and advancement. I’m excited about the promise behind Litmus Health’s product, but I think a big initial challenge will be around data collection and accuracy (as mentioned in other comments). Based on Figure 2, it appears that Litmus Health will be relying on both “active” and “passive” data collection (i.e., user submitted inputs and pre-programmed, tracked data from Fitbit). While active data may offer more granular and specific insights, accuracy and reliability are often less robust due to users’ own biases (self-perception may not be objective) and behavior patterns (patients may not remember to log data correctly or consistently). Passive data is less susceptible to those pitfalls, but the pre-programmed nature restricts flexibility and limits customization, especially if it turns out that the tracked metrics are not relevant in evaluating a given drug’s efficacy. I’m curious to see how the company evolves to address these challenges.

On November 20, 2016, stong commented on Netflix Can’t Chill Anymore :

First off, awesome title (and post). One thing that wasn’t covered in your post is Netflix’s changes to their pricing plans. An initial disadvantage of digitization was the greater ability for “abuse” through sharing accounts (e.g., where multiple individuals share the same login credentials to access Netflix – essentially splitting up the cost of one subscription amongst all of them). I thought it was really smart of Netflix to offer differences in plans by the “number of screens” that can simultaneously access Netflix content (Basic – 1 screen, Standard – 2 screens, Premium – 4 screens). I’m wondering if Netflix can do more around this to capture more value from its customers – for example, enable “add-on screens” similar to how you can add lines to an existing cell phone plan, or implement more sophisticated account validation to enforce unique user access to accounts/ profiles (e.g., 2 step verification for new devices). Curious to hear your thoughts on this aspect.

On November 20, 2016, stong commented on Is Facebook becoming a news site? :

This is a really interesting, timely, and important discussion on what role social media sites like Facebook should have in an increasingly digitally-connected world. Other commenters have already raised some of the big challenges in fixing Facebook’s “fake news” dissemination (e.g., how to decide what’s fake and what isn’t, who makes the decision, how to break “echo chambers” effectively), but I find myself more concerned with the underlying implications on a more macro level.

I am most anxious about Facebook’s growing role in its users’ news consumption and the dangerous potential of someone abusing that influence for their own interests. By aggregating content, Facebook has become a powerful intermediary that controls the bulk of the user experience and commoditizes actual content suppliers like newspapers. As people started spending more and more time consuming content through Facebook, basically anyone who wanted to get the attention of users (newspaper outlets, businesses, etc.) had to be on Facebook as well. By focusing on user engagement, Facebook created a never-ending virtuous cycle for their favor: the more time users spend on Facebook, the more placement influence Facebook has over its business customers. With news specifically, as users spend more time getting news via Facebook, the less time is left to read anything else, thereby further increasing media outlets’ dependence on Facebook.

As such, I question if “top down” solutions are realistic – Facebook simply has no incentive to shift its focus away from user engagement when its business model is entirely driven by that metric. I think the only real “solution” will come individuals resisting the laziness of passively consuming content that’s been “curated” for us and actively seeking out reliable sources of real information and different perspectives.

On November 6, 2016, stong commented on Gundersen Under the Sun :

This was an eye-opening post that effectively highlighted the glaring contradiction between healthcare and the environment. As a society, we have grown to expect healthcare services at any cost, but the resulting pollution and waste directly contribute to harming the overall health of the population. In a similar vein, I wonder if a cultural shift within the healthcare system is required to embrace sustainability initiatives. Medical providers often play a critical role in resource decisions and present a great opportunity for waste reduction. Seemingly insignificant changes in how services or supplies are utilized and delivered can aggregate into substantial benefits in resource conservation and public health. Going forwards, perhaps institutions and hospitals can better educate providers on the impact of wasteful, low-value-added practices and perpetuate a culture of conservation across all staff.

It will be very interesting to see Planet Labs’ progress over the next few years. They seem to have successfully established themselves in generating high-frequency, high-quality satellite images and making that content accessible. However, the true value of this innovation will be the ability to extrapolate agronomic decisions from these images, and emerging agricultural startups are starting to focus on this function as satellite images are becoming more accessible. For example, Descartes Labs aggregates imagery from companies like Landsat, Planet Labs, and the European Space Agency and then uses machine learning techniques to provide data-driven insights and analyses on farming and land use. Perhaps Planet Labs should seriously consider acquiring a target in this space to further bolster their product offering.

Using existing nuclear waste as an energy source is a ground-breaking concept, but I am cautiously optimistic about the idea given the enormous technical, regulatory and temporal risks involved. If successful, TWR will not only reduce the amount of nuclear waste but also risks associated with nuclear proliferation and transportation. However, not all risks will be mitigated as a TWR reactor will still need fissile material (i.e., enriched uranium) to initiate a reaction. Moreover, the eventual nuclear decommissioning of a TWR reactor will also present a lot of risks and challenges. Scientists and researchers must thoroughly test and plan out the full, end-to-end implementation and operation strategy before this technology can be rolled out en mass.

On November 6, 2016, stong commented on H&M: Environmental Steward or Greenwashing Marketer? :

Thanks for bringing to light the fast fashion industry’s impact on the environment. I’m curious to know which facets of the fashion industry operations contribute the most to climate change. My bet is in manufacturing, which makes the prospect of closed-loop production most promising. However, it appears that progress in this type of production is greatly hampered by tradeoffs between sustainability, cost, and quality. The one company that I think is a real leader in this space is Patagonia – they have been doing closed-loop recycling for synthetic textiles like elastane-nylon blends. To chemically process polyester into core components and spin it back into polyester thread is, as you’ve mentioned, very expensive. The process only works with high-quality polyester textile (Patagonia’s own fleeces) as input and isn’t viable with cheap polyester textiles typically used in fast-fashion. But the important distinction here is that Patagonia has been implementing this method out of principle, not for profit. I wonder how much H&M can move the needle without acknowledging that sometimes they just have to do what’s right and not just what’s right for profitability.

On November 6, 2016, stong commented on Tesla: A Two-Step Plan to Transportation Takeover :

While building manufacturing capacity is essential to growth, gaining government support is most critical and fundamental to making Tesla’s vehicles more mainstream, especially in China. Hong Kong is an excellent example of this – the city alone accounted for ~10% of Tesla’s overseas sales last year, compared to 12% from the entirety of mainland China. In just two years of launching in Hong Kong, Tesla sales have more than quadrupled and account for over 80% of the electric vehicles in the city. Musk attributed much of the success to the Hong Kong government’s role in public education for early adoption and setting up required charging infrastructure. Hong Kong also offers tax waivers (non-electric cars have a 100% environment tax, which makes means Tesla cars, despite higher retail prices, end up being cheaper than other comparable luxury vehicles. In contrast, purchasing a Tesla is much more expensive in mainland China because of the lack of subsidies, and 25% import duty on foreign cars. Local production will certainly help by circumventing the 25% import duty, but that alone won’t be enough. To stimulate demand for electric vehicles and truly break into the Chinese market, Tesla will need to convince the government to set up better-aligned incentives such as attractive tax incentives and increased development of charging facilities for consumers.