HNHacker News
TopNewBestAskShowJobs

mlb_hn

726 karma · joined December 29, 2017

playing with little space aliens at julius.ai
submissionscomments
mlb_hn··on Facebook Gave Device Makers Deep Access to Data on Users and Friends
Agree with FB on this one. The privacy issue with 3rd party developers was that friend data was sent to the 3rd party developer databases. Here, the data just stays on the device.

The way the article is titled can make it look like the device makers actually got to make their own database of Facebook users.

Edit: I read it again, the article does say "Facebook acknowledged that some partners did store users’ data — including friends’ data — on their own servers," but they don't follow up on that at all, they go back to messing with a BlackBerry. That's the big question.

mlb_hn··on Elon Musk has a very bad idea for a website rating journalists
It makes sense that something like that is needed. Right now, the media is essentially throwing the bones - there isn't really a way to evaluate much less track what's been correct and incorrect over time, so there's no justification on why one system of interpreting the world is any better than any other.

Of course, as the article points out, evaluation is hard. I don't think there's enough of the guts of how this sort of program would work to really call it 'bad' though. (potentially assuming tracking user votes, running in conjunction with something like doc2vec would lead to some cool results for understanding splits in beliefs).

mlb_hn··on All images of the Facebook ads Russians purchased during the election
To clear up your question, I believe it was Facebook that identified the ads as being of Russian origin and specifically purchased by Internet Research Agency (and then provided to Congress)
mlb_hn··on Two Koreas Agree to End War This Year, Pursue Denuclearization
I fundamentally disagree with that position, although I suspect that question may play a factor into the current partisan divide and the upcoming elections.
mlb_hn··on Two Koreas Agree to End War This Year, Pursue Denuclearization
That could be accurate. That's not the point though - the issue is that there are all these articles being written about North Korea wanting a peace treaty as if it's something new, when North Korea's always wanted the peace treaty.
mlb_hn··on Two Koreas Agree to End War This Year, Pursue Denuclearization
North Korea has wanted a peace treaty for some time. E.g. 2016 Rand did an analysis on North Korea wanting a peace treaty to get the US to withdraw from South Korea (https://www.rand.org/blog/2016/05/behind-north-koreas-bid-fo...). They wanted a peace treaty under Kim Jong Un's father as well (https://www.reuters.com/article/us-korea-north/north-korea-c...).

Everyone seems to be forgetting that the reason that WE have refused to sign a peace treaty for decades is because of the human rights violations in North Korea. I have no idea how the narrative switched to North Korea wanting a peace treaty as a concession from them.

mlb_hn··on Zuckerberg won’t give a straight answer on data downloads
Has anyone been able to figure out where the ad interests from web traffic are?
mlb_hn··on Facebook Can’t Be Fixed, It Needs to Be Broken Up
That and one can't help notice the irony that on all these articles they're still trying to load the Facebook trackers in addition to the others (not to beat a dead horse).
mlb_hn··on Homeland Security to Compile Database of Journalists, Bloggers
Answering that question may be the purpose of this request - since no one knows the state of discourse prior to or after any identified propaganda/ads, it's rather hard to assign any impact. By monitoring the discourse, DHS can then measure changes in sentiment on salient features in order to measure the impact of various propaganda/ad campaigns.

That being said, I think DHS might be going a bit overboard as far as the monitoring goes.

mlb_hn··on Internal Facebook posts of employees discussing leaked memo
There's also selection bias at work here though - the only responses that'll get media attention are the ones that are novel or 'newsworthy' - we aren't seeing the normal boring comments.
mlb_hn··on Beware the smart toaster: 18 tips for surviving the surveillance age
This seems like folk-remedies for issues rather than getting into encryption, time-deleting messages and ad-blockers (well, I suppose the last one would probably be bad for their business model).

The issue being that despite all the coverage of data collection with Cambridge Analytica, the public still doesn't really understand this. Especially considering The Guardian was one of the publications that broke that story.

mlb_hn··on Palantir worked with Cambridge Analytica on the Facebook data it acquired
It doesn't look like he was jobless the whole time.

At 3:01 in the hearing, they get into a discussion regarding Wylie pitching the company he was at in January, 2016, to the VoteLeave people (but not being accepted). Wylie was doing something in Canada at the time though.

mlb_hn··on It's not just Facebook. Thousands of companies are spying on you
It's continuing to like the most recent Facebook exposure is going to get a lot more public interest than Equifax or other previous exposures over the last decade.

The author suggests that the key difference now is GDPR compliance will ultimately force the public to routinely pay attention to the personal data industry.

mlb_hn··on Propaganda on Tumblr
Although you're going to get down-voted, that's an actual problem. We don't know the full state of discourse before interference, during interference, or currently so it's a bit hard for anyone to say 'this is specifically what they changed'.

It's a bit analogous to having a server that lots of people access. Then you notice that a bunch of people have been accessing it with write permissions that weren't supposed to write to it, but the data's totally disorganized, you don't have a backup to compare it to, and reconstructing the changes they made is a pain.

That being said, there's been quite a bit of individual changes identified at this time by going back and looking which are evidence that changes were made, but as far as how any of the changes add up to significant functional differences, we don't know.

mlb_hn··on How can Facebook change when it exists to exploit personal data?
The correct title is "How can Facebook change when it exists to exploit personal data?" (Looks like a 'N' got typo'd in there instead of a 'H')
mlb_hn··on [dead]
There was some really cool research on changing food behavior after World War II by Kurt Lewin.

They found that while price mattered, they also had to figure out the dynamics of who bought food and why all the way from the store to the pantry to the dinner table. "Frontiers in Group Dynamics: II. Channels of Group Life; Social Planning and Action Research" is worth a read if you're interested in that sort of thing, (http://journals.sagepub.com/doi/abs/10.1177/0018726747001002...)

mlb_hn··on Mark Zuckerberg apologizes in full-page newspaper ads
Also, he can't really use Facebook for it if his target audience is the people who stopped using Facebook (they wouldn't see it).
mlb_hn··on [dead]
Gotta say, tried running for a while and I hated it. Tried lifting and I hated it.

Then some friends showed me how to actually lift with a real routine, gained a ton of muscle and kept the belly in check, plus it was fun. Might be a similar thing with running (still hate it) but... if you're having a hard time with weight, it can help to find a buddy who knows how to lift and does so daily or get a trainer.

Doesn't help the 2-5 age group though.

mlb_hn··on It’s time for a digital protection agency
The author suggests fines as the third suggested purpose. The author's take on cleaning is mainly restricted to credit monitoring type deals, which overlaps with the second purpose. (which is part of why I'm saying I don't think it's really thought out all the way)
mlb_hn··on It’s time for a digital protection agency
The author's suggested fix is "Let’s make a digital Environmental Protection Agency. Call it the Digital Protection Agency. Its job would be to clean up toxic data spills, educate the public, and calibrate and levy fines."

A couple upfront issues with this: 1) "cleaning up toxic data spills" - this doesn't seem well worked out, unless the author suggests going and deleting the stolen data off others' computers

2) "educate the public" - the author suggests explaining how to deal with identify theft. That's great, but doesn't address the secondary issues of advertising/propaganda/other clever unintended uses of data.

mlb_hn··on Problems coming for online publishing
> Google News made it quite obvious how redundant a lot of content is. Often articles are just exact copies of Reuters, AP, ... texts.

Reuters and AP are newswire services which is meant to be used by multiple pubs because local pubs don't have reporters all around the world. The local papers' reporters are busy telling you what your local schoolboard is up to and whatnot.

But I agree more pubs will likely disappear or be merged into larger media conglomerates, and that leads into issues like Sinclair forcing all their pubs to push certain agendas.

mlb_hn··on FTC Probing Facebook for Use of Personal Data, Source Says
To your second point, I argued this sort of data was weaponsized and not just in an election (http://www.armyupress.army.mil/Journals/NCO-Journal/Archives...).

To the third point, focusing on Facebook seems like that scene from Casa Blanca though: "There's gambling going on, I'm shocked, shocked" "Your winnings, sir" Not confident FTC fines would actually change any trends.

mlb_hn··on Four Years After Declaring War on Pollution, China Is Winning
Just putting this into context, NYT has an article "How China Is Challenging American Dominance in Asia" on their front page. Not saying it's intentional on their part, but it may be useful to understand how the title "China is Winning" impacts the publication's larger narrative and how that can shape peoples' perspectives.
mlb_hn··on The New U.S.-China Rivalry: A Technology Race
I see that as well. There's been an interesting trend over at least the last several months where NYTimes has been promoting fears about China (No judgment as to the veracity of those fears).

Definitely something to keep in mind when reading pieces like this.

mlb_hn··on Secret Documents from Russia’s Election Trolls Leak
TLDR: A bunch of purported IRA docs ended up on an online auction; Daily Beast went and talked to some of the Americans listed and confirmed the documented interactions with the Russians.

The docs indicated that the IRA was targeting Reddit and Tumblr, previously undisclosed.

mlb_hn··on Introduction to Data Driven Propaganda
I was a human intelligence collector, although I'd also been running a small cell out of South Korea doing analysis in the region in 2016.

The paper was in response to the July prompt of a quarterly contest, "Is the Army too antiquated in how it fights the Information War?" (http://www.armyupress.army.mil/Journals/NCO-Journal/NCO-Jour...). This piece went through several extra months of trimming and editing and the program was subsequently discontinued.

mlb_hn··on Introduction to Data Driven Propaganda
I wrote this piece as I was leaving the Army, I was planning on letting it sit in the archives for a while but I guess I’ll go ahead and explain what it’s talking about.

1. The Russian information operations are not about the American elections. It’s been a sustained information warfare campaign for the last several years, and the elections were part of it.

2. The Russians are playing on existing differences in society. The key difference between their modern campaign and traditional campaign is they have population data to target their propaganda, so they can identify arbitrary fracture lines in groups.

On a side note, this is the same technique that traditional human intelligence uses, which the Russians are good at. Most people are generally good people and don’t manipulate others intentionally, so if this is new to you the US Army’s FM 2-22.3 covers about 20 different methods (non-exhaustive) of manipulating individuals with this sort of information, intended for interrogation. It covers playing on peoples’ fears, their ego, love, etc. The Russians are just doing this on a wide scale, albeit it’s also the same thing other advertisers do, just more widespread and coordinated.

3. Undermining groups is not random. It’s using an evolutionary strategy of undermining groups threatening to Russia’s interest rather than creating groups out of thin air. Without having the data that they’re using for analytics for identifying sub-populations, it’s rather hard to figure out what group a given set of divisive propaganda is targeting. Also, it isn’t just the bots or the social advertising. It’s full-spectrum.

4. The Russians are not the only ones that are using these techniques. However, they are more aggressive about it because Putin freaked out that the US was going to try to overthrow him. They’re in a corner and acting out of weakness, not strength. That being said, we’re not all that great about responding to events in the world so there isn’t a guarantee we come out of this with a good outcome.

I didn’t choose the top picture, the wording just frames it for why it was relevant to the Army (keeping it apolitical).

mlb_hn··on Twitter Followers Vanish Amid Inquiries into Fake Accounts
I don't get the justification for labeling the initial block as organic growth, unless they're claiming that only the horizontal stratification is signs of bot activity. The vertical stratification should also be signs of bot activity, and as they point out, many of the supposed bots were in the vertical stratification groups.

For those not familiar with the graphs, the graph shows date followed on x and date created on y. Where there's horizontal stratification, NYT noticed it means the follower accounts that all began following the account at the same time were also created around the same time, indicating a strain of bots. However, when there's vertical stratification, there's a shift in the rate at which accounts are following the account. When the vertical stratification is followed by horizontal stratification, it's an indication that both sets are bots - for example, one scenario is that instead of providing a mix of bots created at different time, someone got lazy and just grabbed a list of bots all created at the same time. However, vertical stratification could also just indicate that the person did something good or bad to change the rate of acquiring new followers so it isn't clear cut. That being said, I'm not sure that justifies labeling the initial section which lacks horizontal stratification as organic.

mlb_hn··on The Follower Factory
On the graphics, it looks like the density changes for the rest of the dates whenever the identified bot families show up (creating the differently colored columns). When the identified bot families stop, the density changes back. Is that an artifact of how you're displaying it, or is it an indication that the followers before/after are also bots, but they've been mixed with bots created at other dates, where the bot families identified are just the low quality (non-mixed) bot families?
← PreviousPage 4 of 4