John R

  • Alumni

Activity Feed

On November 23, 2015, John R commented on SmartMat – quantifying the unquantifiable :

This is a great post. I never thought someone would actually do this. The points you raise are good ones: they aren’t broadening access, the mat is ludicrously expensive, they are quantifying a quasi-spiritual experience (although this may be a stretch). If you are rich enough to buy this SmartMat (think of a better name for God’s sake!) you can probably afford to be trained live. Practicing on your own means you lose much of the motivation to improve that seeing other people bent into a pretzel provides. I thought yoga was supposed to be about finding inner peace and centering oneself. Since when does my downward facing dog need to be quantified? Why can’t it just be *my* downward facing dog?

Also, this mat may receive backlash from PC practitioners like those at an Ottowa College (http://ti.me/1NM4Gzj).

On November 23, 2015, John R commented on Flo Is Hungry for Your Driving Data :

But the metrics progressive is the most interested in are inherently the most incriminating ones; otherwise there’s no point in doing this at all. Progressive can counter the selection bias by weighting participants by their previous driving history (#claims, severity of accidents). Customers that drove well before signing up will be the gold standard for good drivers, and customers that drove poorly before adoption (assuming they do adopt) will be the gold standard for reckless drivers. The differential can be used to adjust the likely positively-biased results towards the true population mean.

As you say, the threat of big companies like Google and Facebook building a universal shopping cart into their platform is very real and looming. Especially concerning is that retailers would probably prefer to work with Facebook because they can always be a click away from perusing your friends’ photos and finding that dress you really want. If I have to go to another site to start shopping, my impulses are harder to capture and impulse buys are especially big for fashion.

On November 1, 2015, John R commented on Reddit: A bastion of free speech :

Reddit certainly is not a bastion of free speech, but that’s probably a good thing. 4chan and its famous /b/ forum are true bastions of free speech, but 4chan is also the internet’s armpit (I wanted to say something else), having done great things for us like hacking Trayvon Martin’s email, hoaxing Steve Jobs’ death, and facilitating numerous school shooting threats

One of the big issues with crowdsourcing is the reliance on random strangers creating your content. When given anonymity, people have a tendency to lose all decency and self-restraint. I think it’s good that Reddit is heavily moderated because otherwise it’d descend into a nihilistic pit like 4chan. A lot of people nowadays (boy I sound old) do their growing up on the Internet and on sites like Reddit. While I don’t use it myself, I’d much rather peruse a site that’s civil than one where the next comment I read could make me question whether there is any good left in the world. That being said, I just went to Reddit to see what I could find for this response and saw numerous “(USER WAS BANNED FOR THIS POST)” messages; although I was on the very meta “What’s the worst thing you’ve seen on /b/?” thread. ¯\_(ツ)_ /¯

If you want to learn more about why these forums should be moderated, check out Parmy Olson’s book:
We Are Anonymous: Inside the Hacker World of LulzSec, Anonymous, and the Global Cyber Insurgency
http://amzn.com/0316213527
which gives a great overview of what happens when these forums get out of control.

On November 1, 2015, John R commented on How Not to Crowdsource : The Demise of Quirky :

Interesting! I’d never heard of this company before and can’t believe they thought this was a good idea. One thing that’s important here is that crowdsourcing from inventors is inherently biased towards “forward thinkers” (there’s probably a better word…) who would adopt strange new products. Inventor-types are definitely not the norm, so one could deduce that their buy-in is not representative of broader marketability. Also, it’s incredible they didn’t figure out that by manufacturing in-house and splitting resources across a variety of unrelated products that they’d lose all economies of scale and ultimately have marginal costs that were way too high. Add on the margin cuts you mentioned and it’s clearly unworkable.

On November 1, 2015, John R commented on StackOverflow : Q&A site for programmers :

I take the opposite perspective, Allen. I see the critical nature of the respondents as a good thing. It forces respondents to think carefully about their responses before submitting them. Also, it’s not rare that the respondent’s responders give great optimization tips, not only helping not just the OP but also the initial respondent.

People answer questions because they earn bragging rights, they like helping less experienced programmers (“pay it forward”), and (in my case) it’s a great way to learn new things about languages you already “know”. Often the best way to learn new techniques is to attack tricky problems in an area you are less familiar with. It motivates me if I’m solving a problem for someone else.

I practically live on StackOverflow because of just what you said: I’m always learning new languages. Here’s my latest one. Props to anyone who knows what it is. 🙂

+++++ +++[- >++++ +++++++ +++++ +.— ——- -.+ +++++ ++ +++++ +++++ ++.++ +.+++ ++++. — —
-.- —– — —– .+++ ++++++. + +++++++ +++++ .- ——. ++.++ +++++ +.—- — –.++ +++++++ .- —-.
+++++++ .– — –.++ ++++. +++.+ +++++ ++.—- —– — —.<

Hint: you'll probably need SO to figure it out.

On October 5, 2015, John R commented on Get In Ma Belly: How A Little Blue Card Helped Out Mom & Pop :

As Belly grows, they are trying to bring on board the big players as well as the small. If you read my comment above, I think it may undercut some of the “I shop locally” cred users would accumulate with their friends; but hey, those big chains have a lot of cash and scale, making them hard for a profit-oriented Belly to snub.

On October 5, 2015, John R commented on Get In Ma Belly: How A Little Blue Card Helped Out Mom & Pop :

As an retailer that uses Belly, I don’t really have any incentive to recruit new SMBs to the Belly platform. The network effects are “pulling” SMBs into the fold rather than “pushing” out to them, if that makes any sense. For example, a coffee shop in downtown Seattle sees the number of Belly Card users increasing and wants to attract their business as well. The coffee shop then signs up for Belly to encourage customers to gain rewards from their shop rather than going to another Belly-enabled shop down the street.

I agree with you on the “spammy” point though. I think because the focus is on SMBs, Belly users get to brag about supporting local retailers rather than nationwide chains like Starbucks or Walmart.

On October 4, 2015, John R commented on Daily Fantasy Sports Compete Head-to-Head :

You raise a very interesting point that often gets overlooked by online betting proponents. These “sharks” are hard to identify and can shapeshift freely to avoid being banned. In traditional gambling houses, you knew who you should never bet against; but with millions of players online you have no idea. This reminds me a lot of the SEC’s definition of “sophisticated investors” who must have enough AUM or annual income to be deemed “knowledgeable” enough to purchase certain securities. I wonder if similar rules will be put in place for online gambling, with users having to select which stakes they want their account to be able to invest in. High stakes accounts can’t bet in low stakes pools and vice-versa. Obviously this isn’t foolproof but it’s a start.

On October 4, 2015, John R commented on Duolingo: Sheer Genius :

For the most part I agree with you; however “positive feedback” attracting new users is not a direct network effect. If that were the case, essentially all products would be utilizing network effects. My buying a Fiat because my friend told me it has served him well (unlikely) is not a direct network effect unless my Fiat experience somehow improves because he owns one too. You are spot on when you say that more user translation data improves duolingo content for everyone though. 🙂

Their model to make money based on micro-transactions is brilliant. While I don’t play “Clash of Clans”, I have played other games where its very easy to justify spending $1 to catch up to a friend or getting that extra sweet building that everyone else is jealous of and just has to buy for themselves. I bet if people looked carefully at their credit card statements they’d be horrified how much money they’ve cumulatively put into a silly mobile game just to gain instant gratification.

The key to the game is the “Clans” part though. Human beings are psychologically inclined to codify themselves into groups for a sense of belonging and even defense (street gangs). The fact that the game has both component is crucial to attracting and maintaining a large user base. You’re right that other games are just too late, well, at least direct competitors are too late. CoC is just another incarnation of Mafia Wars or Farmville. It will fade and be replaced by another GDP-sapping digital distraction within a year or two. I shudder to think what the next one will be. Altercation of Agriculture? Harmville? Angry Kurds?

On September 14, 2015, John R commented on Strivr: Revolutionizing Sports from Behind-the-Scenes :

Agreed.

On September 14, 2015, John R commented on WTG, Valve! Steam FTW!! :

To answer your question, I’ll pose my own: where else are small indie game companies going to sell/promote their products if not on Steam? For all intents and purposes, Steam has locked up the market for digital computer (though not yet console) game distribution. Smaller studios use Steam’s enormous network effects to promote their games, even if they get little or no money for them, in hopes of gaining recognition, loyal followers, and social media publicity within the gaming community, ultimately leading to lucrative contracts. If I’m running a small shop, I’d choose to distribute via Steam over some no-name platform or (God forbid) retail chain. I (and Richard Stallman) agree that Steam uses its monopoly position to squeeze developers; but what are their alternatives? It’s be squeezed by Steam and make something or be free and make nothing.

Your point about “wait until the sale” is valid though. This chart (http://ubm.io/1J8YnTz) shows just how much sales events affect game downloads. Some games may state that they’ll never go on sale to discourage gamers from biding their time, but I don’t know that for sure.

Hugs oo

On September 14, 2015, John R commented on How is Slack killing email? :

This is very interesting. I had heard vague whisperings about a company like this a few years ago and thought it was a great idea then but hadn’t heard about it since. The product itself seems phenomenally well-designed and clearly valuable to businesses no matter the industry. It reminds me a lot of the value prop of Dropbox, “no more file transfers between machines using USB’s”. Coming from a company that used Outlook (“a curse upon its home”) to send everything via emails, I definitely see the value in real-time chat and drag-and-drop file sharing. Hitting refresh repeatedly while waiting for the 11th version of a giant PPT file to arrive in my inbox is something I never want to experience again.

Your point on stickiness is also key. Once a company commits to this software and their employees grow attached to it, the company can forget about canceling its Slack subscription. While I doubt they’ll be the coup de grâce to email, I don’t doubt that inboxes will be significantly smaller as a result.

On September 14, 2015, John R commented on BigML – Early to the Party :

Really liked your post, especially the playful prose! BigML is certainly winning by expanding powerful data science (ML) tools to users without millions of dollars to invest in their own scientists and infrastructure. As you say, a lot of important research is facilitated by BigML, which is great. My two reservations revolve around unstructured data and inexperienced users. I’ve only given their site a cursory glance; but it would appear that their tools, while sophisticated, don’t help users where they most often need it – when “wrangling” disparate, often dirty, unstructured data. As someone mentioned in class, this is the newest “wave” in business analytics and poses one of the greatest challenges to even the most seasoned data scientists. Though it may be the case that their target market is only interested in highly structured data such as transaction history.

Secondly, one of the most common pitfalls we see is the old “correlation does not imply causation” issue. Data science is an art as much as it is a science, and it’s often the case (spoken from personal experience) that results that come out of ML algorithms are almost too good to be true because they are. Perhaps an incredible correlation appears between bank patrons experiencing bad service and time of day on weekdays, leading the bank to invest heavily in customer service reps only to discover that people going to the bank at 2:30pm on a Wednesday are probably out of work and likely to have more financial stresses than employed individuals. Patrons are unhappy no matter the quality of service because their external situation taints their banking experience and so the bank has wasted a lot of money attempting to address an issue that isn’t as easily remedied as they thought. An experienced data scientist would account and test these hypotheses, but a green ML user would not. BigML certainly offers its users a lot of rope for analyzing data, I just worry it may offer more than enough for customers to hang themselves.

On September 13, 2015, John R commented on Transforming your behavior into digital information :

I’m always very skeptical of machine learning startups that claim to have solved hard problems in human-machine interaction such as speech recognition over imperfect connections and language-based emotion detection without supplemental facial expression data. Having worked in this field, I’ve experienced how infuriatingly difficult it is to accurately elucidate complex signals from noisy, often incomplete data. Not to say this startup hasn’t made significant progress in that area; but I’d be very surprised if they have improved the state-of-the-art NLP algorithms enough to actually add value to businesses. Oftentimes its more about style over substance, especially when the underlying tech is so complicated that nontechnical customers can’t understand how it works. I’d be concerned about Cogito’s machine misclassifying a customer’s emotional state and adjusting accordingly, leading to confusion and poorer service overall.