shas3

Ed Witten, arguably one of the greatest mathematicians/physicists/mathematical physicists alive majored in liberal arts, worked for Nation and New Republic, enrolled in the economics PhD program at UW-Madison, before he switched to math!

Also, hyperparameter optimization is a bit painful. Zoubin Ghahramani of Cambridge is a champion for Gaussian process models which are as flexible as SVM/SVC, but have more disciplined approach: via well-chosen priors (Bayesian structure) and single-step hyperparameter optimization.

This is one of the dumber articles I've read on Uber. I don't think the author has used cabs much and is just theorizing. In a mid-size American city, I've had cabs cancel on me around half the times I've called them, leaving me stranded. I've been through an experience where someone with a fake taxi license fleeced me charging me 3X with physical intimidation. I know someone who was robbed at gunpoint by his fake cab driver. The author misrepresents the taxi industry as some sort of a philanthropic order of saints. The only times cabs have been reliable for me are when going to and from airports.

Uber is fundamentally different from regulated taxis. Also, there's no moral framework for the taxi-related regulations (medallion, commercial insurance, etc.). They're both a relic of an age that has passed and a prime example of regulatory capture.

It is worth noting that the byline for the article states this:

> Benjamin Edelman is an associate professor at Harvard Business School and an adviser to various companies that compete against major platforms

I wonder what 'major platforms'.

Dunno if this is his real account, but it is supposedly not working out https://twitter.com/clattner_llvm/status/877341760812232704

This has happened to me several times (not always in interviews). Especially when someone asks about machine learning and I start the explanation with regression, they get pissed. Nothing shows that you are trying to jump on a bandwagon than expecting needless complexity in explanations.

These lectures are different from his famous Lectures on Physics. The lectures you linked to were compiled in "Character of Physical Law" and are not related to his Caltech undergraduate courses where he presented his "Lectures on Physics" content.

You link to: https://en.wikipedia.org/wiki/The_Character_of_Physical_Law

"Feynman Lectures on Physics" are: https://en.wikipedia.org/wiki/The_Feynman_Lectures_on_Physic...

#3 is very hard if you are living more than 8 hrs drive from the totality zone. Population-wise way more than 10% of the US lives more than 8 hrs from the totality zone.

Eh? He and others were looking at a rape victim's medical records, apparently with the conviction that she was making it up and was paid by their competitor to smear them. If that doesn't raise concerns about the company's culture, I don't know what will. There are few precedents for such shitty conduct in the 21st century corporate world.

I see that no one points this out: Dan Luu and others in this post are using 'Public Company' to refer to only a small high-growth subset of all public companies, namely Google, FB, Netflix, etc. (and perhaps Apple and Microsoft). Most other public companies: track S&P 500 or you joined at a time when your company's growth slowed down to S&P 500 or if you work at a place like IBM that underperforms S&P 500. In these cases, you aren't necessarily assured millions.

I subscribe to The Economist and WSJ and have pretty much stopped visiting Google News, except when there is important breaking news. Google News really isn't for someone who subscribes to news journals. Apple News is much better for this purpose.

I'm super excited about Vive and general VR support for Macs. My interest in VR is not airplane mechanics, gaming, or anything of that sort. I'm just interested in the VR experience. Conceivably, there'll be many VR documentaries, VR puzzles, etc. as there already are. To see that I can buy a Mac and not be locked out of the VR revolution is a relief.

Given the R&D firepower of Oculus and Facebook, they are in a better position to execute. I'd bet on them winning the VR/AR/MR race. I think with continuing hardware developments, PC VR will plateau and become a commodity and the real battle will be on mobile VR where Oculus will win. If I had to bet, Valve may never venture into that space.

I am deeply skeptical that anything other than R&D and patents will come out of this move by Apple. I am bearish on the prospects of using technology in healthcare. Primarily for these reasons:

1. Regulation: it is very hard to innovate at the pace of the consumer-electronics industry while satisfying FDA regulations.

2. For general purpose applications, signals and measurements are highly noisy and sensors are still deeply in the academic realm, working only in very special use-cases, as opposed to 'in the wild'. It is a different matter for special applications such as Type-I diabetes or other diseases because devices are critically important and patients are willing to tolerate discomfort (of wearing probes, pumps, etc.)

3. In general, predictive power of algorithms tends to be lower for non-genomic health care applications because of variability across individuals, errors in installing devices, rarity of certain types of data for training and validation, etc.

4. Some of the key benefits are hard to grasp for users because of the slow time scales over which things like weight loss occur, and also due to the multi-factorial/multi-component nature of human well-being.

I've seen colleagues try < 100% work and what would end up happening is that they would work the usual hours, but just get part-pay. There are really no solid checks and balances other than a manager's goodwill when you start working part-time. The only way you can actually curtail hours is if the manager says "ok, s/he is working part-time, so I'll give them fewer responsibilities." Quite often, project deadlines and not-so-nice managers ensure that you end up working full or more hours even if you are officially part-time.

Honestly you can take the $10k+ difference and donate it to charity or save it up for your kid's college, if you have children. $10k can buy malaria nets for 4000 kids in Africa, if you donate to some place like Against Malaria Foundation. At the moment, I don't see any altruistic outcome from buying EVs. Personally, I am also against using any non-economic arguments in choosing cars, unless you are really rich and the difference is only a drop in the bucket.

Lovely piece!

The addictive element of chess reminds me of Satyajit Ray's movie "Shatranj Ke Khiladi" (The Chess Players). The movie is set in the 19th century at the onset of British colonization of the Indian heartland. It is about two noblemen of the Indian province of Awadh who are addicted to chess to such an extent that they have become oblivious to the dramatic changes happening in their households, the province, and the region in general. There's also a decadent dance-obsessed Nawab (provincial ruler). All three elite aristocrats are consumed in their hobbies and addictions while the British are maneuvering to subjugate their province.

Jagadish Bhagwati of Columbia has a ton to say about how India screwed up on trade. Basically there are many trade-unfriendly laws dating back to 60s and 70s which are hard and slow to undo because they have to be done via parliamentary procedures (kinda how Congressional gridlock makes lawmaking slow in the US). Here's a good interview of his with Russ Roberts on Econtalk: http://www.econtalk.org/archives/2013/08/bhagwati_on_ind.htm...

Edit: this is in contrast to China where laws are made by diktats.

Great read! Many Asian countries have the concept of live-in housekeepers. I think the 'best' (scraping the bottom, here) arrangement is in Singapore, where the government, laws, and law-enforcement ensure that immigrant live-in housekeepers are treated fairly- kinda like au pairs, etc. in the US. I knew someone in India who brought their live-in maid's sister to the US in some capacity (and AFAIK, paid her wages- she returned happier and was able to put her kids, etc. in better schools).

I am pretty sure some of these arrangements, especially in poorer Asian countries like India, Indonesia, Philippines, Cambodia, Vietnam, etc. are effectively, slavery.

In most cases in Asian countries, live-in housekeepers are paid wages and their families are provided with various amenities and help- kinda like a very small-scale version of the servants' lives in Downton Abbey. But, there are at least two very serious problems: 1. cultural norms allow the servants to be treated very poorly- as second-class citizens, 2. while obscure well-meaning laws exist to prevent trafficking in India, law enforcement doesn't care much if they are treated shittily (physical/mental abuse, nonpayment of wages, etc. often go unreported or are ignored by cops).

Thanks for the link to Kiparsky's paper on mAheshvarasutraNi. Starting from the system of grouping and ordering, sandhi rules became straightforward to understand. When I first learned the sutras in school, we were presented with a lot of hand-waving and circular reasoning (sandhi rules- sutra grouping - sandhi rules) about why they were ordered in that seemingly arbitrary manner. It is cool that Kiparsky has solved the inverse problem of figuring out Panini's generative system for the grouping and ordering of syllables!

Elad's pretty clear in laying out how pre-deep learning denoising research is differently elegant compared to CNN, etc. You should also consider the context, SIAM- the applied math community, which has seen a lot of work in deep mathematical connections between PDEs, denoising, etc. The math is certainly more advanced than required for understanding deep learning and applying it to achieve better results.

I am one of those people who got horrible nausea eating a couple of batches of the Soylent bar. I really liked the taste and the nutrition profile. But I simply cannot tolerate any Soylent products after the episodes of nausea. There is only one other food (an obscure Indian snack) that ever created a similar aversion in me. I don't know when, if ever I'll get over the association of Soylent with nausea in my head. I wish it had never happened so that I could have continued using their products.

Jeez. High skilled jobs outside software engineering pay approximately 10-20% less than software. How do you account for that in an 'auction'-based system? It is just a hard truth that unlike a software company, individual talent is proportionally less valuable in a manufacturing company, leading to lower wages (compared to software developers) for: materials scientists, chemical engineers, mechanical engineers, some electrical engineers (power systems, power electronics, etc.).

The reason for this is that a larger portion of a manufacturing company's competitive strength comes from having great manufacturing equipment, ins into markets, etc. vs. just great engineers or scientists. H1B is necessary because many of these special-skills engineers and scientists are foreigners.

This is also the same reason that hedge fund managers or doctors are paid much more than software engineers- the individual's contribution is more important than in other professions. Does this mean that hedge fund managers are more valuable to society and should be preferred over software engineers in a H1B lottery?

United has a history of being a company with a crappy culture. Basically, Uber today is the United of the tech world and United has been the Uber of the airlines industry for the last few decades. Stacking metaphors, even when there were many fish in the aviation sea, United was the rotten one.

Holman Jenkins wrote an informative and funny piece in today's WSJ on this: https://www.wsj.com/articles/make-america-great-boycott-unit...

Greyhound is super sketchy. Service is terrible. About 10 yrs ago, I made the near-mistake of taking a Greyhound to a different city to catch a cheap international flight. I was going from a remote city to a big one. Luckily I had given myself a buffer of 8-12 hours. The bus was initially a couple of hours late for a 5 hour trip. Then, at a transit point we had to change buses and we were told that ticket doesn't guarantee seat and it's first come first serve! And there's only one bus per day from the less remote connecting city! I was lucky to get into the bus that day and not missing what turned out to be a memorable vacation. I only rode Greyhound one other time. Other observations: Greyhound stations are super sketchy in many cities. Buses are sometimes poorly maintained. Tickets are pricey. Etc.

Well, there are two ways of looking at this: 1. Decision to use Python created unnecessary extra work (your point); 2. Executing (not just prototype) with Python, then re-writing and re-optimizing your code in a 'leaner'/lower-level language is a valid approach given that's how Youtube and other highly successful Google products and services evolved.

I think both are valid approaches depending on your application, available resources, nature of incentives, and the speed at which your market moves.

Mirror in brain surgery reminds me of Ted Chiang's (he's been mentioned on HN a fair bit of late) story Exhalation http://www.lightspeedmagazine.com/fiction/exhalation/

While it's hard to believe, there are entire fields where top journals can take 2 years to publish (no early publications, no culture of preprints, etc.) For example, in one subfield of EE where I once worked in mid oughts, the choice ws between an IEEE journal that took 2.5-4 yrs to publish or lesser known IET and online journals that published faster.

Such fields are few in 2017. But it's not implausible that in 2014 the only two quick forums in statistics he was aware of were Arxiv and lesser known journals.

That was my point, only my wording was suboptimal.

In case you missed it, John Baez's collaborator in that post is Greg Egan, who is well known for his brilliant and visionary science fiction. http://www.gregegan.net/

[Spoiler] A rather technical and intricate story is generically spiced (masala) up with typical thrill-tropes to make it appealing to the masses: military attack on aliens (wise aliens only send video-conference units to the earth in Chiang's story), an international affairs angle, melodrama of General Shang, etc.

I was using it in a sense similar to this: https://en.wikipedia.org/wiki/Masala_film