sidrajaram

Their model was not logistic regression (the networks may have had a few logistic units in it, but it's hard to call that a logistic model). The logistic models they compared against were published models from the academic literature. I'm not enough of an expert in this specific subdomain to comment on whether these were the benchmark papers or not.

I don't know if you're trying to imply that the authors of this paper didn't know/know of survival analysis, or if it was a general rant. Looking at the names I know on the paper and the affiliations/backgrounds of the others, it's safe to say they are aware of proportional hazards models.

Survival analysis is not called for when predicting the outcome variables of interest in this study, and that seems to be your primary beef - that they chose the wrong outcomes to model in order to "make hospitals money". I would think that being able to predict outcomes help hospitals plan and manage their resources effectively. From your high horse this may appear to be a wasteful endeavor, but controlling costs will do much more to save lives by making healthcare accessible, rather than building survival analysis models for rare diseases that affect some trivially small portion of the population.

The truth is outside of tech, statisticials (or data scientists) are way underpaid relative to the training and specialization demanded of them. This is true for non-profits and academia. Note that administrators in both these fields are not underpaid to the same degree. Instead of money, they are expected to pay their bills with warm fuzzy feelings of doing good for the world, because of attitudes like the ones expressed in your comment.

Also, fun fact: survival analysis was developed for actuarial use to make ugh money, not bio/medical statistics.

Depends on what you mean by "trust with money" - trust them to spend it on something that will give back to society? In that case, yes...I trust almost anyone more than Jeff Bezos.

I think the point was that the chair viewed as a per-employee "tax" has benefit to the employer, but this head tax doesn't.

Yes, the cost of things is more expensive in dense cities, but by 2x-3x for some services and not at all for buying a pair of jeans from a retailer - those prices are pretty much uniform everywhere, modulo state/local taxes. I'll argue that one's quality of life is better in SV on $250k vs $70k in the rural midwest or $100k in a midwestern city - obviously this will depend on spending patterns, which depends on personal preferences. But for me, it's a conscious decision to make more and spend more on a west coast city.

Walmart competes with Mom and Pop grocery stores. Google is competing with god damn Amazon. And possibly future startups that might otherwise try to enter the smart speaker space, but decide not to because they can't compete at that price point. In which case, great...that is not a case of a large corporation leveraging their size/position to harm customers.

> unpaid data gathering tool

I get free unlimited google searching, gmail, google maps in return. I would pay a hefty amount if they starting charging for those and am happy to trade the use of my fairly benign data usage patterns in return.

> starts selling your information to the highest bidder.

I am confident they will never ever sell my personal information (not population-level aggregates, but actual raw data with PII) because they will lose their only source of revenue and will soon go bankrupt if they did so.

Funnily enough, your post is also missing some important details. The Canadian system is very transparent with its scoring. Unless you're from a natively English/French speaking country, you need test scores to validate your proficiency in English/French. Certified doesn't mean much. Also, you need degrees from accredited/verifiable institutions to really rack up points - if your friend is self-taught, it may be no good. Which does speak to the original point of Canadian conservatism.

> well, at least the ones that don’t abuse the system, like FAANGs

It's unclear whether you mean that FAANGs do or don't abuse the system. I've heard both sides of that.

That must've amounted to several tens of cents (per book) on the $200 price tag.

A CSR returns 4.5% on travel/dining, if you only redeem on travel through the Chase portal, which is not a ripoff based on when I've checked and what I've read (you can get even better effective rates if you transfer miles to airlines strategically).

That means if you spend > $6000/yr on travel/dining (which includes uber/lyft/taxis and hotels/airbnbs besides the usual), you're going to come out ahead of the $150 fee ($450 fee - $300 travel credit) as compared to a no-fee 2% cash back card. The $300 travel credit won't count, so change that $6k to $6.3k. This is well within reason for a lot of white collar folks.

That seems to be a precursor to the work mentioned in the article. This is the one that was presented at NIPS this year: https://papers.nips.cc/paper/6182-fast-mixing-markov-chains-...

The first year or two of math (and physics, and to some degree, CS) are roughly standardized. There are variations, not only across colleges, but within the same college (honors vs. non-honors, calculus for business, calculus for pre-meds who won't take any more math), and among professors teaching different sections of the same class even. But to a first order of magnitude...roughly the same stuff.

Many American students have often taken a year or so of calculus (sometimes more) before starting. But these are so-called "Advanced Placement" (or AP) classes which are considered "college-level" (even though at top colleges, this is basically a requirement to get in). At less competitive/less technically focused colleges/programs however, most students may not have taken it.

You can think of AP as an equivalent of A-levels in the UK. I'm not sure if calculus concepts are A-level or not? But either way, from what I know of UK education, you can get by not taking mathematics A-level...it's the same for AP mathematics here.

NIH too.

The current deep dreaming fad is going to last about one more week before everyone gets sick of looking at those stupid pictures. They really serve no purpose other than "look how crazy this stuff is" - deep learning is cray-cray.

I'm no expert, but afaik probabilistic programming isn't a new method or technique. It is just wrappers around existing statistical techniques, as an attempt to divorce the details of inference algorithms with model specifications.

I'm not buying in just yet, because although it's nice to talk about model specification as completely independent processes, the availability of fast inference algorithms sometimes dictates what models you should choose. Sometimes less exact models with a larger parameter space that allows you to crunch orders of magnitude larger datasets (with approximate inference algorithms) yield more useful results than better specified models...and sometimes not. The thing if one still needs to know the whens and whys of picking certain models over others, and can't just gloss over the inference details.