rm999

It's fun to dig up old predictions of where we'd be today. Here's an article from 2002: https://www.eetimes.com/document.asp?doc_id=1176510

>Moore's Law scales to at least 9nm: technologist

>It'll be at least a human generation before Moore's Law begins to run out of gas at around the 9nm and even then it may thrive, TSMC's chief technologist said Tuesday (March 12). Calvin Chenming Hu told an audience at the annual Semico Summit conference here that the 9nm node "can be ready more or less on time, in 2028 according to long-term forecasts or 2024 according to the 2002 (industry roadmap)."

Intel is off by 4-5 years according to their initial estimates (in 2011) of hitting 10nm, but they are 5-10 years ahead of where TSMC thought we would be 15+ years ago. TSMC was manufacturing at 10nm last year, incidentally (Qualcomm's Snapdragon 835).

>I'd take ISO/IEC 9075 as an official specification

Fair enough, but I'd argue it's not a de facto standard because the vast majority of implementers don't follow it. Case in point: SQL is rarely portable across databases, even if you're e.g. moving from postgres to a postgres-like system like bigquery or redshift.

>Do you have examples?

For me (as a machine learning + data engineer) it often comes up around aggregating with window functions. Things that would take a couple lines in R or Python or Scala can take dozen of extra lines with superfluous CTEs.

SQL is by no means perfect. For one, t̶h̶e̶r̶e̶'̶s̶ ̶n̶o̶ ̶o̶f̶f̶i̶c̶i̶a̶l̶ ̶s̶p̶e̶c̶i̶f̶i̶c̶a̶t̶i̶o̶n̶: some dialects are meaningfully different than others, and even similar ones are often full of implementation details about the underlying database. It has a ton of quirks, and isn't as powerful as I'd always want it to be. And sometimes it can be really hard to read.

But I still haven't found a better universal "language" to talk about extracting and manipulating data. The core set of functionality of SQL (selects, joins, aggregations) is usually adequate to express powerful ideas. Most importantly, a lot of people who work with data know it, and I've found semi-technical people (including non-programmers) can learn it quickly. Anyone who has been paying attention to trends in the data world in the last 5 years knows SQL is here to stay for the long-term.

I'd say we're both right: the article is based around a flawed study that hilariously contradicts the title and is full of circumstantial evidence. You are right that they cite plenty of other studies and quote qualified people. But I'd claim that the faults I cite instantly disqualify it from being a quality objective piece, and that there are likely better places to get a well argued viewpoint.

As a subscriber I love the New York Times, but like many articles in the Wellness section this is low quality.

The article starts off with:

>A recent clinical trial by the F.D.A. suggests that the complications experienced by Mr. Ramirez are not uncommon. Nearly half of all people who had healthy eyes before Lasik developed visual aberrations for the first time after the procedure, the trial found.

And then five minutes of reading later leads to this valuable insight:

>But many of the trial’s 574 participants reported having visual aberrations and dry eyes before surgery, and the study concluded that Lasik slightly reduced the prevalence of these problems.

Why... didn't they put that in the same paragraph? The title of the article leads me to believe this obfuscation was on purpose.

Then, a few minutes later into the article, they state:

>The study’s lead author ... said the researchers had concluded that the multimillion dollar trial was too small to produce meaningful results, and that the purpose of the study had shifted from determining how many patients have problems functioning to developing a questionnaire that might be used in future research.

Ugh. I just spent 10 minutes reading an article formed around a flawed study that contradicts the title. Filled with circumstantial evidence and discussion about clearly biased people (like the guy with seemingly no credentials who runs lasikcomplications.com).

I've had a tremendous amount of respect for Mark Zuckerberg as a leader, and have gone out of my way to defend him in the past (https://news.ycombinator.com/item?id=17128369). I always envisioned whatsapp as a defensive purchase that would stay true to its initial vision (perhaps being used to prop up their other products in an unintrusive manner). But this article makes me realize Mark is s a one-trick advertising pony like half the other tech leaders out there.

His speil about how he's "connecting the world" is undermined by his clear attempts to build a monopoly of social+communication, with an inefficient layer of distracting ads. Imagine a future where every time you want to electronically communicate with someone you have to spend mental energy filtering out ads. It's the equivalent of every road having a toll booth, directly paying into a rich person's bank account. This is the "connecting the world" Mark Zuckerberg has been trying to hide from those of us who aren't paying attention.

> I believe that the success of the company is incidental to who ran it

This is usually not true for successful companies, but it's especially not true in facebook's case. Mark Zuckerberg has kept incredibly strong control of facebook, holding more than 50% of voting control even as it's grown to a half trillion dollar company - so he definitely deserves credit for what his company has become.

Also, he has made some strong, unintuitive decisions along the way (no advertising in the first several years, real name policy, dropping out of college, moving to silicon valley, hiring experienced tech executives, buying a dead-simple photo filter app for 1 billion dollars) that can directly be attributed to facebook's success today. It's hard to remember today, but in the first ~5 years of Facebook it was very far from obvious that facebook would dominate the social space.

I know a few earlyish employees at Facebook, and they are easily the happiest people I've met at any company. In some respects this can be credited to joining the right company at the right time: they got in early at a company that makes insane amounts of money and has a huge influence on the world. But this seems to be the icing on the cake to them. Everyone I know at Facebook who got in before 2010 credits Mark Zuckerberg ("Zuck", ugh). They call him a friend and say he never wavers in his passion or mission. They seriously look up to the guy and won't work anywhere else, even though they have built up generational wealth.

Honestly I don't care much for Facebook's methods and I think they are a net negative on the world, but I still really look up to Mark Zuckerberg as a leader.

Google dominated non-SMS chat in the late 2000s/early 2010s. I don't think they realized how insanely valuable that was - they were basically sitting on the future of communication. All they had to do was sit back and let network effects build their monopoly (with some more tech build-out and straightforward marketing in international markets).

Instead, they myopically saw Facebook as a threat and started one of the most user-hostile rampages I've been victim to. They hobbled existing products and over-complicated new products to force people onto a platform no one asked for. Facebook (and others) used this opportunity to grab communication away from Google.

It's all an interesting business case study of missing the forest for the trees.

>Is it because the board was full of connected people?

The board was scammed by her and lost billions of potential dollars, so that's probably not why. If anything, I wonder the opposite, why the connected people didn't do more; I'm guessing it's because destroying her further isn't worth their time?

I'm not one to normally defend Facebook, but I disagree. Advertising was not prioritized for a huge chunk of Facebook's history, so I'd argue it's not part of their core ethos (it's a necessary evil, if you will). While a lot of their more recent product decisions have been made around revenue generation, I don't think that's what keeps their leadership up at night. Keeping a monopoly on user attention (euphemism is "connecting the world") is their biggest challenge, especially in the ever-evolving social industry.

To put it in perspective: today you are in the global top 1% if you have at least $744,400 in household assets.

You are in the top 10% if you have at least $71,560; top 50% (median) if you have at least $2,220. And the mean wealth is $52,819 per adult.

source: https://www.economist.com/news/business-and-finance/21710771...

This wouldn't make much of an impact because the statistical distribution of music listening time has an extremely long tail, i.e. the most popular songs are far more popular than even other very popular songs. In a given market, something like 90% of listening time will come from the top few hundred tracks, and 99% will come from the first couple thousand. Those top couple thousands tracks are already well-served by alternative distribution channels (youtube, "indie" labels, etc), so there isn't much room for disruption there. Starting a service to fight for that 1% of listening time will not move the needle.

source: I used to work in the data side of the music industry.

I'm surprised that AOL isn't really discussed in this article. I always assumed Yahoo's ethos, as the one-stop shop on the internet ("We didn’t want to call it a portal, because a portal is a door to somewhere else, and we wanted people to stay there") came straight out of AOL's playbook. What I know for sure is that I, and a lot of other people, immediately felt comfortable with Yahoo when we first started using the internet because we were so used to AOL's vision of how you connect "online".

I think the connection run strong - I'd go as far as saying that Yahoo and AOL failed for the same reason. To survive at their inflated valuations they had to be everything to everyone, which stretched their vision thin. Meanwhile, it was always impossible to do everything, so smaller, more agile players came in and etched away at their business models.

It's in the original complaint filed by federal prosecutors. See page 16:

https://assets.documentcloud.org/documents/4406486/Vincent-R...

Note that this is in Celsius - for us Americans, this is 45 degrees Fahrenheit above normal.

Hey, I just installed and played around with reactionRNN; I hate to be negative (especially to people open sourcing models, kudos!), but your model seems to perform quite poorly. It immediately failed my "easy" smell tests: https://i.imgur.com/FvfuZgy.png, and didn't really work on most of my other tests: "This book sucks" is 0% angry, "I'm going to go home and listen to emo music and cry" is 0% sad, "Check out this hilarious youtube video" is only 26% haha. Your example "He was only 41." is 100% sad, but "He was only 42." is 0% sad. These aren't hand-picked, these are literally things I just typed in. From what I can tell it usually gets anything negative wrong, and usually picks "haha" for the positive ones.

Subjective performance is worse than your included examples. I've been building models my whole career, and what I've learned is most people will take claimed performance at face value until it burns them. It's beneficial to no one if someone comes up with an idea based off your repo description, builds it out, then finds it doesn't work adequately. My advice is to update your examples and test cases, and keep finding ways to improve the model.

>For more than 30 years, Geoffrey Hinton hovered at the edges of artificial intelligence research, an outsider clinging to a simple proposition: that computers could think like humans do—using intuition rather than rules.

This is so disrespectful to the 1000s of researchers who have been studying machine learning since well before 2012. It was well established that the future of teaching computers came from statistics and not rules in the 80s/90s by researchers like Michael Jordan (https://en.wikipedia.org/wiki/Michael_I._Jordan) and his students.

It was even engrained in popular culture: neural networks are how the AI brain worked in Terminator 2, in 1991! https://www.youtube.com/watch?v=xcgVztdMrX4

edit: I don't want to downplay Hinton's accomplishments, I've been lucky to have been surrounded by and motivated by his work since I started learning machine learning. I did my masters research on neural networks that were partly inspired by his work, and it was a deep networks paper he presented at a NIPS 2006 workshop that got me really excited to stay in machine learning while I was starting my career.

I can't speak for OP, but a big one is how to deal with the check at a restaurant. For people who are lower/lower middle class (or at the least not financially secure), you pass around a receipt and take turns deciding who owes what. In the middle/upper middle class, there is an unspoken rule the check gets split, and anything else is awkward. In the upper class, oftentimes one person will pay for the check and it's awkward to object.

>That whole ethos of just having informal personal conversations with college friends on Facebook is gone.

The wall was never a good way to do this. It's neither as powerful nor as flexible as pretty much any competing way for a small group of people to communicate, including chat programs, email, txt, snapchat, etc.

Facebook's deprioritization of the wall and buy-in on messenger/whatsapp, instagram, and the newsfeed was very smart. I suspect they are ok with people moving off posting on facebook to posting on instagram and chatting privately.

In one project at a previous job we were able to pinpoint the genre, mood, instrumentation, etc of any new song with (usually) better-than-human accuracy in milliseconds using deep neural networks. This is better in pretty much every way than the 1000s of music experts that are employed by competitors and vendors.

https://tech.iheart.com/mapping-the-world-of-music-using-mac...

Not to mention personalized recommendations, which basically aren't possible at scale without some level of ML:

https://tech.iheart.com/mapping-the-world-of-music-using-mac...

https://news.ycombinator.com/item?id=12269568

The thing to keep in mind is every machine learning practitioner who is worth their salary is doing stuff like this. A lot of our every day work isn't as sexy as teaching a computer Go, but it's game-changing to most industries.

This is a negative value article. It spends a lot of time dwelling on a low-value semantic argument that "when people in business say 'AI' they really mean machine learning". Sure, but these words have been used synonymously for a long time - my grad degree ~15 years ago concentrating in machine learning was called "AI"; who cares? Then the article goes on to claim:

>I anticipate that after this passes, we can start to do the right thing — focusing on using machine learning to build things that are meaningful and realistic.

UGHHH. Why did you hide this in a long article making the 180 degree contradictory point that AI (which really means machine learning, remember?) is dumb?

This article is a disservice to its reader because it downplays a huge shift in the world that any ML practitioner should understand very well: machine learning/AI will ultimately replace a lot of what humans do, and it's moving at a faster pace than ever. I've led several small-ish projects (1-2 people, 3-6 months) that could replace dozens or even 1000s of experts in their respective fields. There are thousands of people like me, and there are more every day. The buzz may wear off, in the same way Time Magazine stopped talking about how the internet was going to take over the world after the dot-com crash, but the effects and the efforts will continue unabated.

Weird - I visited Chicago a couple years ago from NYC and was shocked at how pedestrian-unfriendly central parts of Chicago seemed (e.g. River North). There were parking lots everywhere, with aggressive drivers pulling in and out. I walked through a parking lot to go to a liquor store and was almost run over. Sometimes crossing a large street felt like a race against the green light. Jaywalking signaled to drivers that they should speed up, which is the mark of a driver-run city IMO.

The whole experience contrasted heavily with the east coast cities like DC, Baltimore, Philly, NYC, Boston. Maybe it's changed in the last few years, though.

Likely true, but it also depends on how you define it. The OPHI (Oxford Poverty & Human Development Initiative) has been tracking what they call "multidimensional poverty", which is a holistic view of poverty over health, education, and living standards, and is far more insightful than the X dollars a day metrics. They have found the poverty rate (and destitute rate) in India is very high and is comparable to that of Sub Saharan Africa: https://www.theguardian.com/world/2010/jul/14/poverty-india-...

This doesn't prove the Economist's claim per se, but it indicates to me that it's not absurd.

A lot of the extreme poverty in India is concentrated in the northeast and in rural populations, so it's not always super visible to visitors outside the segregated slums of cities.

> How is Google even remotely comparable to Amazon???

If you're already scared of Google and Facebook this is an easy one: Amazon's growth will increasingly depend on advertising, and they will continue to grow. Pretty much everything they have done in the last couple years moves in this direction: a search page riddled with sponsored results, Echo (https://gizmodo.com/yes-your-amazon-echo-is-an-ad-machine-18...), my new 30 dollar Fire tablet that always shows me ads, retargeting, etc. Amazon has Google-level information on its users, and will use it in ways that are simialrly invasive.

I see a lot of people writing off Amazon because it's not as ubiquitous as Google in their lives. This is a mistake: their dominance will continue to grow. They have the best product strategy I've ever seen, and everything they are doing will move them in the direction of being a central part of everyone's lives (even non-Americans - when I was in India recently and I saw an Amazon billboard or bus ad every 5 minutes, even in smaller towns).

That's an interesting one. I'd vote:

Apple > Microsoft >> Google == Amazon >> Facebook

Facebook is dead last by a mile.

Beyond the lack of smartphones/earbuds, the two big things that stick out to me:

1. The cars. Sometime in the 90s and 00s cars started becoming a lot less boxy (compare the Camry from 1990 to 2000 to 2010 to see this).

2. Way more formal dress wear. In today's NYC financial district you plenty of jeans, and even a lot of financial professionals have given up the suits and ties. The financial district today also has more tech workers, and I believe more tourists.

For sure. In another office they use a geiger counter (a much more standard way to get randomness, and probably easier to setup).

That said, this doubles as a public art installation in their lobby - something many companies spend thousands (or millions!) of dollars on. So it's not just PR, it's actually surprisingly practical.

> What's the balance? I've no idea,

An obvious solution is rural development, looking to India as an example. The accelerating urban migration in India in the 90s/2000s was not organic: young people were forced to leave their homes because they didn't have job opportunities or enough land to farm, and then older people had to follow for cultural and economic reasons. The ultimate goal is sustainable urban growth, but cities can't immediately handle what this would require. Instead, India has instituted fairly successful rural development programs that are aimed at increasing jobs, improving social safety nets, and empowering women. I think China should seriously consider ramping up similar efforts.

http://blogs.worldbank.org/endpovertyinsouthasia/what-india-...

>I came to the conclusion that area under ROC is pretty much garbage.

I cannot disagree more. IMO AUC is a rare great metric: it's principled, useful, universally applicable (e.g. invariant to class imbalances), and easy to explain adequately to non-statistician ("probability of choosing a positive sample over a negative one").

>What you really want is two separate steps: estimate the probability that instance A belongs in classification X, and then a decision step where you decide how to classify A based on a loss function (this varies depending on how harmful a false positive is).

Yes, but those are two inherently separate steps and should be measured separately. AUC is a metric of the first step (the "model"). The second step is a business decision and will often be made separately from the modeling process, by different people, with a different cadence, and with a different goal in mind.

For example if I am designing a model to find a disease, I just want to make the best prediction I can, which is cleanly measured by AUC. Then, when it comes to actual diagnosis, someone else will choose cutoffs based on various factors like false positive costs (treatment cost, human toll), false negative cost (disease damage, death toll), supply of treatment, etc. I can picture scenarios where the same model is used for decades, but the cutoffs change seasonally.