HNHacker News
TopNewBestAskShowJobs

jimmytucson

1,470 karma · joined February 16, 2012

Software engineer specializing in data engineering, language technology, ML, and search. Reach me at my HN username at fastmail.com.
submissionscomments
jimmytucson··on Jack Ma Again Endorses Extreme Overtime as Furor Rages On
This is the essence of “fake news”.

Most people think fake news is reporting false information. But nowadays it’s relatively easy to fact-check something. Much more onerous is selecting facts to report on so as to tell a story that fits an agenda. It’s harder to recognize, and hey—it’s facts! What could go wrong?

jimmytucson··on The Difference Between SQL’s JOIN .. ON Clause and the Where Clause
As soon as you filter on the table you’re joining with in the where clause it becomes an inner join.

    FROM foo
    LEFT JOIN bar ON ...
    /* Makes it an inner join: */
    WHERE bar.baz ...
jimmytucson··on New study: Google manipulates users into constant tracking
> Eh, this has existed for decades at a macro level in over-the-air television and radio.

Advertising predates the internet, of course. What does not is the ability to personalize advertising to the individual using computer intelligence trained on volumes of data about people’s lives, recorded in microscopic detail. I don’t think the level of manipulation and personalization pre-internet is comparable to what advertising firms such as Google and Facebook are capable of now.

jimmytucson··on New study: Google manipulates users into constant tracking
> 3 or 4 rounds of A/B testing and you get the most efficient (read: manipulative) method of getting users "engaged"

Now replace “engaged” with “to buy stuff” and that’s essentially online advertising. You’re being tricked by gurus of the mind to send your money to someone else. How this is legal represents a gap between traditional laws and modern times.

The sad truth for us software engineers is, a lot of the most exciting work is at least tangentially related to refining this trickery: big data, machine learning, etc. For example, implementing GDPR where I work has been a bizarre experience from a human standpoint. On the one hand, we all surely feel privately it’s a good thing for people, but we talk about it to each other as if it’s a burden and a setback for our business (which it is).

jimmytucson··on Ray Kurzweil: AI is still on course to outpace human intelligence
What if human beings were physically connected to a growing network of perfectly organized facts and information? What if that network could even assist with basic logical reasoning tasks? Would that be enough to say we’ve reached the singularity?

It doesn’t seem like it now because our phones aren’t quite physically connected to our brains, and the internet is far from organized, but I’d venture to say we’re already pretty close. A more interesting question to me is whether this is really anything we want.

If human machines were infinitely intelligent then they would all agree. So you wouldn’t need more than one. Sure, the ultimate human might need 7 billion pairs of eyes but those eyes wouldn’t need any more brains than to fuel up and locomote. Furthermore any human-like tendency to doubt oneself or fall in love or “rock out” would be deemed useless and therefore overridden.

Understanding that the purpose of a human is to pass on genes, and that there’s little human or genes left in her, the ultimate human might therefore conclude that her sustenance only disrupts the purposes of organic life forms. Her next and final act would be to destroy herself.

This is how I see the end of humanity. Not in a violent clash between artificial and organic intellects, not in a symphony of mushroom clouds, but with the final flip of a switch, in cold, calculated resignation.

jimmytucson··on What People Say Before They Die
What you’re referring to as “The Christian bible” is roughly 77% the Hebrew Bible (referred to as “the Old Testament” by Christians). There are a lot of stories in the Hebrew Bible about The Lord “choosing” one people and bringing about the destruction of another. It sounds brutal and unjust (which, to me, are intriguing qualities to attribute to God—as opposed to a simplistic, purely benevolent God that a child might believe in) but the stories themselves are wonderfully interesting (e.g. Exodus).

As someone who read the Bible as an atheist, I strongly believe that faith is not required to appreciate, and even be guided by, the Bible. Much in the same way one does not need to believe in orcs to appreciate and learn from Lord of the Rings.

jimmytucson··on Ask HN: Going from Developer to Manager. What should I know or learn?
Leaders are declarative and managers are procedural?
jimmytucson··on What did Jesus really look like? (2015)
https://en.m.wikipedia.org/wiki/Historicity_of_Jesus

Pretty interesting:

> Virtually all New Testament scholars and Near East historians, applying the standard criteria of historical investigation, find that the historicity of Jesus is effectively certain ... with very few exceptions such critics generally do support the historicity of Jesus and reject the Christ myth theory that Jesus never existed.

jimmytucson··on Write good git commit messages
My take: these aren’t good commit messages. The verb isn’t supposed to be what you did, otherwise it would always be “add” or “change” or “fix”. The verb is supposed to be what the program does thanks to this change. E.g. “Check server fingerprint”, not “Add server fingerprint check”.

We already know you changed or added something, it’s a git commit.

jimmytucson··on Serial Port SDR
I don't think transmitters are made artificially expensive to keep out the newbs. As you said, it's illegal.

And if you're a newb (like me) interested enough to try this at home, you might have read to the end of the article where it states:

> DISCLAIMER: It should go without saying that you should never transmit on any frequencies or at any power levels for which you are not authorized by the local governing body, but I guess I just said it anyway.

jimmytucson··on Measuring the “Filter Bubble”: How Google is influencing what you click
If the entity is uncovering a fact then what does it matter whether it's a foreign or domestic entity who uncovered it? The fact itself is what's interesting.

I get that an entity could be more interested in uncovering facts about one party than another but in this particular case, the fact was that the Democratic Party was acting "unethically" in some people's eyes. If they hadn't been then it wouldn't have "influenced" people's opinions.

Another way to look at it: that entity didn't annotate the emails, the content of the emails was enough to anger people.

I'm not try to fork this into a flame war, I'm genuinely interested in what other users of this site think, particularly those who think differently than I do.

jimmytucson··on Measuring the “Filter Bubble”: How Google is influencing what you click
When they say "bubble", I think of groups of users with fewer differences within the group than outside the group, sort of like the Wall Street Journal study showed. If they're finding variation among users, but not predictably more or less variation between any two of them, then that isn't a "bubble" to me, it's just customization.

Customization is troubling, but less so than bubbling. (Hey now...)

jimmytucson··on The Devil's Hair Dryer: Hell is other people, with leaf blowers (2016)
> David Dudley is the executive editor of CityLab. He is the former editor in chief of Urbanite magazine

Sounds like he spends a ton of time around lawns.

jimmytucson··on People are “consistently inconsistent” in reasoning about controversial topics
This comment is not satire, I'm being brutally honest: I don't believe everything for the right reasons.

I believe in global warming, but not because of research papers or books I've read on climate change. I've clicked on research papers but I often find the scientific terminology and information in them overwhelming. I think, "It looks like they know what they're talking about." Sometimes I don't even click. I look around at people commenting on it, and if an articulate person makes a case one way or another, it influences my beliefs.

I know this is wrong but frankly it would be impossible to be educated enough on all these matters to justify having an opinion. Imagine if everyone who felt strongly about global warming had read and understood at least one research paper on it. That would mean half my friends, family, and coworkers have read and understood a research paper. I'll take the under...

What makes matters worse is being around people socially who have strong opinions. Not having one can make you look boring or dimwitted. If the conversation wanders into an area in which you do have an opinion--based on facts and research--citing those facts can feel like reading off box score statistics at a Yankees-Red Sox game. No one gets excited!

That makes me cynical. I think, people just want to feel strongly about something, kind of like they want to have a baseball team. It binds them together. I guess that's fine, I just wish we treated it more like sports and less like religion.

jimmytucson··on S3 trickery: using it as a scheduler
This isn’t using S3 as a scheduler. Cloudwatch already supports cron and rate expressions, as the post alluded to. This is a hack to schedule something once.

All Cloudwatch would have to do is implement a recurrence = 1 feature but I’m guessing it's not a common enough use case.

jimmytucson··on Machine Learning in Google Bigquery
Plenty to chew on here. Thank you
jimmytucson··on Machine Learning in Google Bigquery
> I could see people saving money in some specialized use cases, but for most people BigQuery is almost certainly going to be cheaper and faster in almost every way

That’s exactly what the TPC benchmarks are designed to show: how different appliances perform under the same diverse set of generic workloads. As of 2 years ago, they show that Redshift is faster.

> With BigQuery, you don't really worry about that crap

The benefit this depends on your organization’s level of expertise. If you grok sorting and distribution then you can leverage those to increase performance—but it’s not a prerequisite.

For example, in Redshift, when you bulk load data into a table, if the user didn’t specify a compression scheme in the table definition, Redshift will analyze the data and find a scheme that works best, and automatically apply it to the table for you. BigQuery almost certainly does something similar. The difference is, with BigQuery, you’re not invited to take part in that discussion. And you’re charged as if the data is uncompressed. Psychologically, this is a huge relief if you don’t (want to) know how compression works but rest assured you’re paying for it somehow.

To draw a tired analogy, vehicles with automatic transmission still have to shift gears. If you’re driving to the grocery store, not having to worry about that is a win. But if you’re racing stock cars, you’re definitely going to want a stick.

jimmytucson··on Machine Learning in Google Bigquery
This blog post from 2 years ago from AWS has Redshift beating BigQuery pretty handily on TPC-H and TPS-DS: https://aws.amazon.com/blogs/big-data/fact-or-fiction-google...

Has anything changed on BigQuery since then that would warrant rerunning those benchmarks? If Amazon’s largest cluster outperforms BigQuery then is your decision about which service to use just a cost calculation of dedicated per month for Redshift vs. projected units scanned per month on BigQuery?

jimmytucson··on Canada hits back at 'punitive' U.S. tariffs in 'turning point' in relations
Is Trump’s point that the prices of these goods abroad are artificially low, so he’s going to implement tariffs to correct for that? If so, do counter-tariffs strengthen this argument? This would seem to create an infinite protectionist feedback loop.
jimmytucson··on Python’s For - Else
Here’s why (Hettinger):

> In 1991, “else” was the obvious choice because of the traditional way compilers implemented while-loops: pastebin.com/tY35CTJ4

That is, “if I haven’t finished the loop, GOTO the body.”

He says if he had a time machine he could tell Guido in the future we’re all using structured programming so no one will find “else” intuitive; call it “nobreak” instead: https://m.youtube.com/watch?v=OSGv2VnC0go#

Guido says if he had a time machine he would not have included the feature at all: https://mail.python.org/pipermail/python-ideas/2009-October/...

This whole thing was Knuth’s idea, not Guido’s: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.103...

jimmytucson··on 'I’m not having children because I want to save the planet'
Millions of species flourish and procreate without destroying the planet. Their populations rise and fall in harmony with other species and the resources nature provides. Only one species is destroying the planet, and it’s not because we’re acting like animals (procreating and such).

It’s because we’re acting like humans.

We grow food where it’s not supposed to. We survive disease, natural disasters, shortages. We live too damn long—not only as a species but as individuals. Think of one natural check or balance we haven’t either turned off or are actively working to. Even minor inconveniences like boredom or unwanted facial hair we’re putting coal to flame to destroy.

If you want to save the planet, have lots of babies, live a simple life, and die young. And if you’re wondering where I got this from, watch Disney’s The Jungle Book. It’s all you need to know.

jimmytucson··on Testimony of Mark Zuckerberg – Hearing Before US House of Representatives [pdf]
There are so many other ways to "influence" people's opinions by distorting what they see on their news feed, front page Google results, etc.

For instance (from Wikipedia):

> In April 2016, Correct the Record announced that it would be spending $1 million to find and confront social media users who post unflattering messages about Clinton.[1][4] The organization's president, Brad Woodhouse, said they had "about a dozen people engaged in [producing] nothing but positive content on Hillary Clinton" and had a team distributing information "particularly of interest to women".

jimmytucson··on We Dare You to Explain Luke’s Plan to Rescue Han in ‘Return of the Jedi’
This sounds like a question Sci-fi Stack Exchange would make mince meat out of. For example: https://scifi.stackexchange.com/questions/182014/why-does-th...

But seriously, little inconsistencies like these make me think Star Wars was just meant to be a kids’ movie and not a sacred document to be handed down and analyzed by generations like the Tanakh or something.

jimmytucson··on A surprisingly potent technique can boost short and long-term recall
This. A lot of people get so caught up in the benefits of doing nothing they forget it’s dangerous to do too little. One technique that’s worked wonders for me is breaking up doing nothing with 15 minutes of doing something. That way I’m not doing nothing or something all of the time.
jimmytucson··on Dow plunges 1000 points
Right.

Better yet, wait for the end of this correction, then buy more stock.

Better still, sell any stock you own, and at the end of this correction, buy it all back and more at a cheaper price. Yeah, that’s the ticket!

Guys, all we need to do is find out when this correction is going to end. I feel like we can narrow it down to between one day and one year from now.

jimmytucson··on Minigo: An open-source implementation of the AlphaGo Zero algorithm
Sounds like this meets the requirements: https://github.com/niklasf/python-chess

> very carefully respects all possible game-ending pathways

Not sure exactly what this one means but it implements a is-game-over method, so you could do MCTS with it.

> a translation of a chess board into an array that a NN could understand

Chess positions are represented by sets of 64-bit integers so I don’t think this would be a blocker.

> and a schema by which to flatten the array of all possible moves (both legal and illegal) into a single vector

The list of legal moves is a set of bitboards as well.

jimmytucson··on You’re Descended from Royalty and So Is Everybody Else
I read something while researching my DNA results that made me realize how little my ancestors have to do with me. Maybe somebody here can tell me if it's actually true.

You share about 0.5^n percent of your DNA with each member of generation n, where n = 1 are your parents, n = 2 your grandparents, and so on. The explanation that was given to me is that each of your parents contributes 50% to your genes. Depending on which of their parents' genes they gave you, you share anywhere from 0 to 50% of your DNA with each of your grandparents, but for simplicity's sake let's call it 25%.

The point is, most of the people we're trying to find out more about through ancestry websites and DNA tests are barely related to you. Searching for your great, great, great grandfather from the old country? Fine, you only share about 3% of your DNA with him.

jimmytucson··on How Do You Vote? 50M Google Images Give a Clue
The title reads a little disingenuous. Aren’t they using deep learning and Google Street View to predict the make, model, and year of a vehicle? Then, separately, they’re using predicted vehicle information to guess the demographic traits of the person who lives there.

Still, applying image recognition on Google Street View is a pretty cool idea. The location information gives you a whole other set of data to join with. For example, you could recognize flags hanging on people’s houses and join that with immigration data (there are a lot of Italian flags where I live), or recognize species of trees and join that with climate data.

jimmytucson··on A Message to Our Customers about iPhone Batteries and Performance
Child, please. Apple has taught a generation about planned obsolescence. Observe:

2017 lawsuit claiming Apple broke FaceTime on iOS 6 to force users to upgrade to iOS 7. [1]

2015 lawsuit alleging iOS 9 degraded performance of iPhone 4s to force users to upgrade their device. [2]

2014 article from the New York Times showing how Google searches for "iPhone slow" coincide perfectly with the release of new devices. [3]

2013 Brazilian lawsuit claiming Apple could have implemented 4th gen iPad features on 3rd generation model. [4]

2013 article from the New York Times about performance degradation coinciding with Apple's release of iPhone 5. [5]

2011 article on Apple's "innovation" of a proprietary 5-point screw to thwart hardware upgrades on iPhone 4, Macbook Pro, and Macbook Air. [6]

2009 TC article on UI becoming sluggish, random restarts, device heating up (is this starting to sound familiar?). [7]

2008 Movie rentals not being available on 5th generation iPods which were released just months prior. [8]

I even found an article from 2005 complaining about the high cost of replacing the iPod battery ($250!). [9]

[1] https://www.reuters.com/article/us-apple-lawsuit-facetime-id... [2] https://www.usatoday.com/story/tech/news/2015/12/31/apple-la... [3] https://www.nytimes.com/2014/07/27/upshot/hold-the-phone-a-b... [4] http://appleinsider.com/articles/13/02/21/brazilian-lawsuit-... [5] http://www.nytimes.com/2013/11/03/magazine/why-apple-wants-t... [6] https://www.ifixit.com/blog/2011/01/20/apples-diabolical-pla... [7] https://techcrunch.com/2009/06/04/the-short-lifespan-of-the-... [8] https://www.wired.com/2008/01/ipod-rentals/ [9] https://www.technewsworld.com/story/42328.html

jimmytucson··on Please Don't Show Us Another “Typical Family Earning $270,000 a Year”
The author makes a valid complaint about the examples these senators used to promote the tax bill but he doesn't further my understanding of the bill. In fact, what little he says about it seems wrong:

> All changes to deductions, for example, only affect people who itemize their deductions (almost always high earners)—and within that group, only the subset whose deductions topped the new, doubled standard deduction. And this means that if your income is in the bottom half, your tax return won’t showcase many of the ways the law is changing.

Let's break it down:

> Most people who itemize are high earners.

From what I found[1], 1 in 3 tax filers itemize. Most deductions come from mortgage interest and state and local taxes. I pay a mortgage and a hefty NJ state income tax. Am I a high roller?

> The amount you pay in taxes will not change unless your itemized deductions top the new standard deduction.

That's exactly wrong, right? If your itemized deductions top the new standard deduction then you will still itemize, so your deduction won't change. If you do not itemize (looking at you, non-high rollers) then your deduction will double. If your itemized deductions are lower than the new standard deduction (peasants like me), your deduction will increase to the new standard deduction.

Also, I can't tell if he's saying you won't pay less in taxes, or if he's saying "your tax return won’t showcase many of the ways the law is changing". The latter is vague enough to be non-falsifiable. The former is just wrong because the tax rate is changing for every bracket but the lowest one.

This is a contrived example, but a single filer making $50k who took the standard deduction last year would have remained in the 3rd tax bracket (25%) without the bill. With the bill, he'll deduct an additional $6k, which drops him into the 2nd tax bracket, which is now 10% (down from 12% a year ago). That's a nice chunk of change for a deadbeat!

1. http://www.taxpolicycenter.org/publications/who-itemizes-ded...

← PreviousPage 7 of 12Next →