Hacker News new | past | comments | ask | show | jobs | submit login
How journalists use youtube-dl (freedom.press)
1651 points by thisisparker 36 days ago | hide | past | favorite | 328 comments

I have friends in law enforcement who use youtube-dl to save copies of videos involving crimes or criminal allegations, from quite a few different social media and video sites.

Speaking with them, it seems the associated metadata (json/xml?) and output log contain valuable information to go along with the video from an evidentiary perspective. Also the fact that youtube-dl has many test cases in the code goes towards demonstrating its reliability as an instrument for collecting evidence.

Certainly seems a better approach than installing a 'DownloadVideosEzy' extension for chrome or similar.

I am not in law enforcement myself, but do some online investigations, and like many others in my field we use YouTube-dl to save a copy of video evidence relevant to a case we are working on. It can be instrumental for archiving the evidence if it is ever removed or taken down, can also be used to grab extra info (CC text log is one example) and then manually searched for specific strings. This tool has made ma y investigations pay off in ways they never could have otherwise.

How is this significantly different than placing a video camera in front of a screen?

Yt-dL is surely nicer and more convenient, but from an evidence standpoint, a one time $500 (to go wild) setup should have you covered, no?

(Mourning the loss of ytdl myself but trying to be realistic)

Eek, the thought of that makes me cringe. I guess it would be passable in some circumstances, but I can imagine a bunch of reasons why it would be miserable or useless:

* the quality drop; recompressing, non-matching frame rates, non-matching resolution--all the same goes for audio. You're very likely clipping (losing) data. This is assuming you're doing screen-capture. If you're literally video taping a monitor you will get moiré in the video, room tone in the audio, losing any stereo separation, and other audio/video artifacts.

* performance; must be done in real-time, cannot queue up multiple sources. This is likely the biggest efficiency killer and makes things 100x more labor intensive.

* reliable Internet; if you get a blip or have a slow connection you have to hopefully catch it and start over. With youtube-dl you can pause, resume, confirm even on the slowest, spottiest connections.

* metadata, organizing, indexing; likely hand-typed separately, prone to error, prone to not knowing if you've done that video already.

* Chain of custody; grabbing the original video allows you to prove two identical copies match (using file hashes or other comparisons) screen recording makes that difficult to impossible to confirm--maybe with fancy AI you'd have to run by the courts?

Eek, you are responding to my comment as if it was a freestanding response about archival copy and law enforcement work, when it was specifically in response to someone saying he was using it for neither. It's not surprising it makes you cringe, but please consider in context.

> the quality drop; recompressing, non-matching frame rates, non-matching resolution--all the same goes for audio.

Are you trying to preserve quality or prove something? My response was in context for "gathering evidence" but not police work, and not archival quality. Would such a copy cause your problem to prove libel, copyright infringement, illegitimate disclosure, etc?

> performance; must be done in real-time, cannot queue up multiple sources

Can most definitely queue up multiple sources. Just make a youtube playlist and record it. Yes, it takes "real time latency", you'll take 10 hours to download 10 hours of video in general -- that's not an issue for evidence or gathering in a non law-enforcement context.

> metadata, organizing, indexing; likely hand-typed separately, prone to error, prone to not knowing if you've done that video already.

Again - consider the context of my answer, NOT archival quality anything. The "cc" stream GP mentioned, which can be searchable etc - has also seen many revisions for many files when the Google STT algorithms are revised, and with corrections.

> Chain of custody; grabbing the original video allows you to prove two identical copies match (using file hashes or other comparisons) screen recording makes that difficult to impossible to confirm--maybe with fancy AI you'd have to run by the courts?

You have no chain of custody. You can prove two downloads are the same, but YouTube does not guarantee they keep the file the same (indeed, they've modified files several times, changing formats and even remastering old '80s videos). If a file is later pulled (which is what GP was talking about), what are you going to compare it to?

Chain of custody is law enforcement business. They'll get the files from YouTube directly, with affidavits and statements about it and any modifications, if they need it in court. You are going to civil court, and youtube-dl is not making your evidence more valid than a screen recording.

Is it possible to replace youtube-dl with a video camera in front of the screen? Sure.

Did I expect anyone, ever, to propose that? Let's just say, 2020 is full of surprises. I want off this wild ride.

This, along with screen capture, is known as a form of rebroadcast and it's used to obscure and obfuscate digital alterations, watermarks, deepfake artifacts and the like. When doing media forensics, it's optimal to get as close to the raw source as possible.

Isn't this just due to more compression? What's stopping someone from turning down the bitrate and re-encoding a video into different formats a few times to kill the quality (which will still look fine on a 6 inch phone in portrait orientation)?

That's one of the effects we model in fact, "social media laundering" is the term of art.

Various detectors are more or less thwarted by it. It actually surprised me how strong the artifacts from some GANs are - they can survive several passes of re-encoding, but accuracy does suffer.

But I still need the raw 720/1080 stream for training.

Nothing. But his point is that raw files are better than screencapture/record

I'm afraid to ask how you take screenshots.

With the appropriate tools, such as gnome-screenshot.

But when I find myself on a locked down computer - e.g. watching a movie on an AppleTV, or when I was shown surveillance video but was refused a copy for some bureaucratic reason (or it required a different license to export, reasons were unconvincing) - I use a mobile device.

Yt-dl lets you download whole accounts or playlists.

Using a screen recorder like shadowplay or obs also doesn't seem like it would be out of the realm of possibility.

I think this is absolutely reasonable question.

Filming the screen, means that in order to fake it, you have to setup something that routes youtube.com to your own fake version of youtube, before filming. To me, that sounds much harder than say "this file was downloaded from here on that date"

"oh nice, a youtube-link from one of my sources, let me get my camera set up to archive it ..."

I kind of expect a serious investigator to archive these materials just for the sake of it. I don't expect them to make it harder on themselves for no good reason.

> Mourning the loss of ytdl myself but trying to be realistic)

It's just a GitHub repository which is lost, not ytdl itself.

So - if the issue is really the marketing around youtube-dl, does this mean someone can create a fork named something else, use different marketing, and carry on?

I can see no reason why not. Just don't put it on GitHub, put it on a private hosting in a country unfriendly to copyright trolls.

Obviously CC stream is not readily searchable if you record it with a video camera.

Well, sure, but google/YouTube search does find in it. The GP was talking about their work collecting evidence - they can find it just as well, and record a copy for posterity just as well.

I am not saying it’s as convenient (more options >> less options except for analysis-paralysis). But I don’t understand how it tips the scale to making any archiving or evidence gathering unusable or uneconomical. (I am not saying GP is wrong - just want more explanation so I can understand)

Law enforcement already has an anti-circ exemption under 17 USC 1201(e), though a finding against youtube-dl would limit it's (or similar tools') availability and/or functionality.


Honest question: How is Section 1201 anything other than incoherent?

Suppose there is some legitimate non-infringing use of some material. Interoperability, law enforcement use, fair use, accessing public domain material which is distributed using the same technological measures, whatever.

You create and distribute a piece of software to interface with the material for those purposes.

It's the same piece of software as one designed for any other purposes, because the software has no knowledge of your intent and your intent has no technical effect on its operation.

If that is allowed, doesn't that make 1201 a dead letter? If you make the tool available for non-infringing uses, the tool is available.

But if that isn't allowed, doesn't that make 1201 a contemptible offense against all of those important interests? And possibly unconstitutional as a result?

Section 1201 requires that the piece of software is meant primarily for infringing purposes. Which means that it being able to infringe alone isn't enough for it to violate 1201, but also that because it can be used for non-infringing purposes doesn't mean it doesn't violate 1201. As with many many laws intent matters and deciding intent is why we have judges and juries.

That sounds a lot like saying that it is, indeed, incoherent, but don't worry because we can get a judge to decide whether or not you're going to jail ex post facto, at a cost to you of more than you paid for your house (win or lose).

Computer people probably see the nondeterministic behavior in words like “primarily” as evidence of a bug. To the law school graduates who write the laws, it’s a feature and not a bug that these decisions are made by a judge in a courtroom.

It helps to realize that "human reality" is nondeterministic in the sense that we lack sufficient computing power and modelling capability to work with it. It's not random, it's AI-complete. We literally can't handle it formally, and that's why we need to defer to the only "general AI" we have - or as its ordinarily called, human judgement. And human judgement is not random - it's taking into account vast amounts of data, like ethics/morality and social context, which for now escape our attempts at formalizing them.

And the concept of case law / precedent is an attempt to synthesize a point of view from multiple different judges. “Training data,” if you will.

I don't see where you're getting that. GP rightly pointed out that 1201 prohibits tools "primarily" designed for prohibited circumvention. That's a perfectly coherent thing to prohibit, and that has the perfectly clear, logical implications that GP described.

Now, if what you really mean is that is is vague, then I agree. But I don't see any case for it being incoherent.

> GP rightly pointed out that 1201 prohibits tools "primarily" designed for prohibited circumvention. That's a perfectly coherent thing to prohibit, and that has the perfectly clear, logical implications that GP described.

Except that it isn't, as already discussed, because the code doesn't change based on the design intent. The technical operation of the tool is identical whether it's used for interoperability or fair use or to rip songs for The Pirate Bay. The technical operation of the tool is identical whether it's produced by a university researcher or the developer of a media player app or a literal sea-faring pirate with a parrot and a peg leg who is the founder of a political movement to abolish copyright and promote civil disobedience.

But prohibiting a specific person (who may have had a particular intent) from distributing it is incoherent (or at least futile) if it means you can just get some other person with some other intent to distribute the exact same thing. And it's also incoherent to say that the other person can't distribute it even if they genuinely do have a different intent, if the intent is supposed to be what matters.

None of which depends on what you judge a give person's intent to be after the fact, because it's a defect in the legislation, not the judgment in any particular case.

He probably consider any human judgement incoherent.

When "guess, it will work somehow" is all you can say, it sounds like you agree that it's incoherent.

Maybe. But nobody said that.

>Section 1201 requires that the piece of software is meant primarily for infringing purposes

It requires that the piece of software is meant primarily for circumventing technical measures in place to prevent infringement. Just infringing copyright isn't a violation of the section.

The tool itself is not inherently illegal. Intent and marketing is the only important factor.

A tool designed to be used by police and marketed towards police is legal, as its intended purpose is not infringement. A tool advertised as "hey, kids, use this to infringe those copyrights" is totally different, even if it's literally the same tool.

But then most of the defenses of the RIAA's actions here don't make any sense. When the purpose of youtube-dl is to enable the likes of mpv interoperating with YouTube then having RIAA music in the unit tests would be apropos of nothing because it would be completely valid to want to test the interoperability with RIAA music from YouTube.

Moreover, that would save the law from being a dead letter in the sense that there would be circumstances when it could theoretically be enforced (i.e. when someone is overtly marketing it as a tool for infringement), but wouldn't anybody making the tool then just not do that? It's still exactly the same tool. And then what does the tool have to do with anything anyway, when what you really have is a law against speech promoting copyright infringement?

> most of the defenses of the RIAA's actions here don't make any sense.

HN is not a court of law, most people here are not lawyers, let alone well-paid and competent lawyers like the ones the RIAA can afford.

> wouldn't anybody making the tool then just not do that?

Sure - which is why you're not seeing many ads for tools to steal cars or pick locks. But it's not only about the way the material is promoted, it's also about how it's obtained, whether the seller could reasonably know what it would be used for, and a bunch of other stuff.

You can't look at this from a binary perspective; despite depictions in popular media, the law is typically not about "aha" moments, but rather about putting a number of coherent pieces together to paint a certain picture.

From youtube-dl readme:


As a matter of policy (as well as legality), youtube-dl does not include support for services that specialize in infringing copyright. As a rule of thumb, if you cannot easily find a video that the service is quite obviously allowed to distribute (i.e. that has been uploaded by the creator, the creator's distributor, or is published under a free license), the service is probably unfit for inclusion to youtube-dl.

A note on the service that they don't host the infringing content, but just link to those who do, is evidence that the service should not be included into youtube-dl. The same goes for any DMCA note when the whole front page of the service is filled with videos they are not allowed to distribute. A "fair use" note is equally unconvincing if the service shows copyright-protected videos in full without authorization.

Support requests for services that do purchase the rights to distribute their content are perfectly fine though. If in doubt, you can simply include a source that mentions the legitimate purchase of content.


Nit: I presume you mean 17 USC 1201 (e) " Law Enforcement, Intelligence, and Other Government Activities."

Good catch, fixed, thanks.

Doesn't law enforcement have a means to subpoena Google for such evidence?

A few points you might want to consider:

* youtube-dl is not just for Youtube (ie. google). It supports heaps of sites, and youtube is probably less likely to have the sort of content I'm referring to than twitter, pornhub, liveleak, who knows.

* I am in Australia, so for a subpoena (or similar) my friends would need to issue an MLAT request, which would indeed take a very long time. In fact, this often does happen, but in the meantime a formal preservation request (to the service provider) and a local copy (via something like youtube-dl) are important steps.

* Many investigations don't go to court, for a litany of reasons (that's a pun I guess). In such cases, a subpoena isn't in context: law enforcement have to investigate what has happened, to figure out if it needs to go to court.

* Some investigations are important but do not meet the 'serious crime' threshold for forcing companies or parties to provide information. If you can imagine being the victim of a minor crime, that someone filmed and put on twitter, I'm sure you can imagine the local police may want a copy of the video before the tweet gets deleted - but also know that there will never be an international cooperative justice process. It's just a local issue.

* Publically posted video is very common and often important, and it is not surprising that the police need a tool to download it. There are so many sites and web browser technologies, and the cops are not efficient spending their time coding youtube-dl equivalents.

I feel quite strongly about youtube-dl being taken down for copyright infringement by the RIAA, but have tried to answer in an informative way. Sorry if it comes across harsh - no ill will intended :)

The process for a subpoena takes forever and isn't guaranteed.

If I have a subject who has relevant video on a service like YouTube, and I decide to go the subpoena route, here's the process:

- Find the exact video, document the URL, Content creator's account name, and video upload date.

- Submit a preservation letter to Google to ensure the video will still be accessible when the subpoena is auctioned.

- go to my prosecutor/legal team to determine what jurisdiction the subpoena has to go through. It could be the county I'm in, the county the subject lives in, the county the video was uploaded in (unlikely), or the county where Google hosts the video.

- draft the subpoena and get it reviewed by whichever judge/magistrate applies.

- wait who knows how long for the subpoena response, hopefully it comes back with the real video and important content.

- face defense arguments in court that my subpoena was improperly obtained/submitted

"The process for a subpoena takes forever"

No, it doesn't. And that "forever" are human rights that are established and should require a subpoena to contest.

I mean, we’re talking about publicly viewable stuff, here. Both the citizenry and law enforcement ought to have ways of archiving stuff like this.

(I can imagine lots of cases how this could be used to fight police corruption.)

That's for private, non publicly accessible, content.

If you post a video of murdering someone publicly - no subpoena should be required.

They have a handy upload site for your requests!



Maybe save that argument for where it applies?

In this thread we're talking about publicly available data.

The fact that the path that works for non-public data is burdensome is good because it reduces abuse. Demanding that it be used for public data too just creates more pressure to undermine the (already insufficient, IMO) protections that exist for private data.

I mean, 'taking a copy of publicly-available evidence' instead of asking someone else to provide it isn't exactly 'not doing policing properly'.

I think that using youtube-dl to download a video that a suspect has uploaded to YouTube, and presenting it in court is an example of doing policing right.

If you don't want your videos to be used against you in court, I would recommend not publicly broadcasting them to the entire world.

Assuming the content is public, why would they need a subpoena?

you mean "because they could just use youtube-dl to download it?" I think that's why we are arguing that ytdl is a tool with legal and moral purposes and should remain accessible.

Is availability on a locked, walled garden "public"?

I think we can tell what the RIAA thinks about that.

RIAA doesn't dispute that it's public.

Being public doesn't mean publicly owned. It means publicly accessible... Like your local shopping mall.

Because the RIAA has shown that this specific tool can't be 'public'.

If we have to play by those rules, LEO should too. Boo hoo if it's harder.

> RIAA has shown

They haven't shown anything. They made a claim. We'll know what rules we have to play by if it comes to a court ruling.

I mean, GitHub took it down, no?

Why would that have any bearing on the validity and legality of RIAA's logic?

You can claim anything, too. GitHub only checks if the Takedown is valid in format with necessary information provided.

Has someone abused this to harass big-name Github projects with takedown notices?

They follow the DMCA process.

The RIAA isn't part of the government.

What if this tool remains public in a different part of the world? Are you going to wall off from the rest of the world so you can no longer see the tool and pretend it doesn't exist?

Which way do you think takes less time and effort?

It seems the referenced URLs in question in the source are part of unit testing.

Aside: Wonder how gitlab handles DMCA requests.

It looks pretty consistent with GitHub


For a self-hosted instance, obviously it would be different. In that case, I assume the notice would go to the webhost.

Why was my first though here "I bet Palantir have their own fork of youtube-dl which they rent out to cops in a SaaS thing for the low low price of $ASTRONOMICAL-SUM-COMPARED-TO-ALL-THE-DONATIONS-THE-PROJECT-HAS-EVER-RECIEVED ???:

"Please don't use uppercase for emphasis. If you want to emphasize a word or phrase, put asterisks around it and it will get italicized."


(I understand the environment variable convention but if you think about the reason for the above rule, it's clear what needs to take precedence.)

Well the RIAA will most likely create the opposite outcome by removing YTDL. Now boat loads of people who did not know about this tool will find it. Am I mistaken in thinking this source code only existed on one website .. Now owned by m$ ?

Imagine if source code was posted on some type of blockchain like ETH2.. Or a decentralized file sharing service, like IPFS. I wonder if websites or self hosted interpreters using some kind of services like those would make serving a take down request nearly impossible.

Well for now we can pay m$ to host code in ways people rely on for the long term. Until, they don't and the codes all taken down in something analogous to modern book burning.

Software Freedom Conservancy are asking Microsoft to resign from the RIAA over the youtube-dl takedown demand:


That Microsoft is a member of the RIAA is a big problem. One more reason to abandon GitHub.

GitHub is "OK". It's only popular because it was first to market(?)

I find GitLab far superior in every regard: its UI, CI stack, Auto DevOps, K8s integrations, on-prem/self-hosting option, support options, pricing structure, customer service, communications and open nature, frequent updates with great release notes, and more.

Did Microsoft make a decision here? I thought it was DMCA takedown which legally requires immediate removal.

The issue is not that MS obeyed the takedown as they are required to do. The issue the SFC are discussing is that MS are part of the group that issued the takedown request in the first place.

MS has two hats in the game here: one as a member of the RIAA who have issues a DMCA take-down request, and on as the owner of GitHub which has acted upon that order.

DMCA takedowns are about removing violations of someone's copyright, like posting a Beatles song on YouTube. Though the request was made to look like a takedown request, it's hard to see how they could make the claim the software itself was a copyright violation.

My two big uses cases for youtube-dl:

1) Downloading videos to use later in editing for analysis pieces / editorial videos.

2) Download videos that may disappear in a "No longer available" blackhole.

The 2nd is incredibly useful and telling when you go back, look up videos by their ID, and see what has been deleted. I'd also recommend changing the OUTPUT format to include the channel name or Twitter account, so you can easily see if an account or channel has been censored/banned.

My main use on youtube-dl is downloading videos to watch them in places with slow / unreliable DSL.

Watching a youtube video on a slow (1k) DSL has become entirely impossible because a few years back youtube (seemingly) stopped buffering the entire video, so you can't start the video, pause, go do something else and watch it once it's loaded. Not only that, but it really does not handle either temporary disconnections or high latency spikes well.

I've had similar issues on a recent camping trip since the switch from google play music to youtube music... when I lose connectivity, there's no buffer, and seems to be no songs kept to be able to play anything, and it was a really crappy experience.

May have to go back to spotify or offline mp3 files, I don't miss having to manage my own playlists tbh.

Google's engineering is not about providing a good user experience, but about using metrics to squeeze out the last 1% of inefficiency. This inevitably makes the system fragile and leaves some people out in the cold, but you're the product, not the customer.

Some people would argue that optimising YouTube for the bottom 1% of Internet connections makes no financial sense, but I have gigabit fibre and YouTube stutters. It automatically upgrades to 4K videos (of course), but its buffering algorithm is pared so close to the bone that it can't handle high bandwidths as well.

I was living a few blocks away from YouTube HQ building and with 1gigabit up/down fiber optics cable had YouTube stutters. They [YouTube] definitely don’t optimize for that

Watching TV shows about manufacturing process (e.g.: How It's Made) made me realise that manufacturing is not just about "making things perfect", but also "discarding the outliers".

You can make something quite bad on average, and that's okay, you just have to be able to filter out everything you don't want and keep what you want. When CPUs are manufactured, this is what they mean by the "yield". It's the percentage of the product that can be kept, with the rest of the wafer discarded.

Chef's Gallery had a scene that actually shocked me a bit -- this award winning chef was making this deep-fried puff thing that was absolutely perfect. They showed his process, which was to make dozens of them and then plate just the best one for the customer. He never had a knack at all for making them perfect! He was just throwing out 99% of the puffs that he made, using the same technique as anyone else would.

You just have to change your perspective: You're the product. You're the deep fried puff.

If you're an outlier, you will be discarded. You're the bent piece of framing. You're the slice of the silicon wafer that failed the test.

Nobody feels the slightest bit bad about rejecting a faulty product on the production line. No tears are shed. No phone calls are made to the product to see if there's anything the manufacturer can do to fix the situation.

This is Google and by extension YouTube in a nutshell. They're an advertising company manufacturing ad impressions and ad clicks. Viewers are their product.

Viewers on 1 Mbps or 1 Gbps are equally outliers. Both are too weird to cater to, less than optimal, unpredictable, difficult to advertise-to viewers.


> Viewers on 1 Mbps or 1 Gbps are equally outliers. Both are too weird to cater to, less than optimal, unpredictable, difficult to advertise-to viewers.

Viewers on 1Gbps give off two important signals to advertisers.

1) Probably living near a city to get fiber coverage, which these days is usually a signal of wealth

2) Can afford a fiber internet monthly subscription, also usually not cheap.

This is most likely somebody that advertisers are very interested in getting their ads in front of!

> 1) Probably living near a city to get fiber coverage, which these days is usually a signal of wealth

> 2) Can afford a fiber internet monthly subscription, also usually not cheap.

Unless you’re assuming US residents these are not really signals or helpful. In Japan or Korea you can get GBPs fiber for $30. In Eastern Europe you can get fiber for as low as €10 in Romania I think, in western you can get that for 50 in Switzerland.

Meanwhile there are places where it’s not an option at all unless you get into “contact us” price ranges.

I understand their optimizations, that said, many people listen to music while driving and will have spotty connections, even commuting to/from the burbs. So it's a negative experience that probably outweighs the bandwidth issues. Even a couple songs in buffer at 64kbps encode in the background is better than letting it freeze multiple times a minute for 10+ seconds at a time.

They're too aggressive on their optimizations for mobile networks/devices.

> but I have gigabit fibre and YouTube stutters

This is a peering issue and your ISP's fault, not YouTube's. They can't fix bad peering from their end, your ISP has to do it.

Your comments seems contradictory. How is optimizing to provide a good experience for those with bad internet not providing a good user experience? I have fast internet and it works fine for me, maybe the problem is on your end?

My point is that because of the excessive "optimisation" at YouTube, users now have a poor experience if they "don't fit in the middle of the bell curve". The cause of the issues can be high latency, low latency, high bandwidth, low bandwidth... whatever. If you're in the 1%, YouTube has optimised you out in order to save 1% in hosting costs somewhere. It's a type of over-fitting that results in a fragile system.

Engineers at Google get bonuses for shaving 0.1% off of something, because at their scale that could be millions of dollars saved.

Hence protocols like HTTP/3, which exist almost entirely to optimise some Google backend by single-digit percentage points.

YouTube has had every last percent of "inefficiency" squeezed out of it, to the point where lots of users have a degraded experience.

Google famously doesn't care about user experience at all. They care about costs and their own internal KPIs, which are all tied to advertising revenue, not "video playback smoothness".

This is why Firefox was 5x slower on YouTube for years. This is why Google famously has next to zero "customer support", even if you pay them. They don't view you as a customer. You're the product.

Firefox users click on ads less. Firefox users tend to have adblock. They're not good products.

Similarly, Google is fighting a turf war with the likes of NetFlix and Apple for advertising eyeballs, so they do not want to ensure that Apple TV can play back YouTube in the best possible quality. They optimise for Chrome and Chromecast first, everything else a distant second. Got to build that walled garden!

I pay NetFlix the same amount monthly as I pay for YouTube Premium. NetFlix provides support, YouTube doesn't. NetFlix works flawlessly on every device I have, YouTube doesn't.

YouTube doesn't play 4K on my Apple TV 4K! It doesn't play 4K on my flagship Samsung TV! It downgrades my iPhone for 480p even on WiFi!

This kind of anti-consumer (anti-product?) bullshit is why Google needs to be broken up.

The vendor that makes the device, the browser, the search engine, the network protocol, and the advertising platform shouldn't also be television for half the world.

They shouldn't get to degrade the experience to benefit their browser team. They shouldn't get to slow down the experience for a competing browser. They shouldn't get to simply ignore customer complaints. Television broadcasters in most countries have to answer to an ombudsman. YouTube doesn't.

That's too much control that invites anti-competitive, anti-human behaviour. The incentives are all wrong.

I agree with all your critique of Google/YouTube, so I was surprised to read that

> I pay for YouTube Premium

So you're aware that you're the product even if you pay them and you still pay them?

Not that it matters as it's a drop in the ocean.

A hen will lay eggs, which have a fairly direct monetary value. It doesn't mean the farmer treats their chickens like customers.

In a poor town, a shopkeeper may accept barter instead of cash. Their customers might pay them with eggs, and they would be treated like valued customers even if they're too poor to use trade with real money.

It's a matter of corporate culture on how the consumers are viewed. For largely advertising-driven companies their users are products, even if they pay. For largely product and sales-driven companies their users are customers, even if they're in a free tier or a trial account or whatever.

In the case of YouTube Premium, you're paying in lieu of advertisements, so you're the customer again. ;-) Also, my understanding is for most Premium users the creators get a bigger slice than they tend to get from ad views.

What are you talking about, Youtube Music literally downloads automatically 100s of songs locally for offline playing, and you can download entire albums for offline playing at the press of a button.

There's even a little popup saying something like "It seems we can connect to the server, do you want to play your downloaded songs instead ?" when there are connectivity issues.

I don't even keep most of the videos I downloaded due to that, as they are not worth watching more than once anyway, so it could be argued that it's no different than how people usually use YouTube.

AFAIK, (good) browsers still give you the direct links if you know where to look, but the slow boiling of the frog is really evident with things like hiding View Source and such (often under the guise of "usability".) The demise of youtube-dl, along with Google's pushing of their proprietary protocols and other continued user hostilities surely paints a sad picture for the freedom of the Internet...

Ditto, but with slow/unreliable 3G. It's perfectly possible to watch a 4k video with max quality that way, even if one has a poor connection. Otherwise, I'd be stuck at 240p much of the time.

try out mpv. with cranked up caches it will buffer the whole vid in ram and bobs your uncle.

It is really distressing to see how many of the videos in my, admittedly quite large, favourites list are gone by now. Even ones that I only added in the past few weeks. Either simply "Deleted", "Private" or "Removed/Blocked by the copyright holder". In most cases I have trouble even finding out what the title of the video even was.

This is an example of the dichotomy between "the internet remembers everything" and the inherent ephemerality of electronic data (effort is required to ensure it remains un-destroyed).

I self-host some things because I like the technical challenge, and what I've found is that the effort required to maintain the online presence of the data is quite demanding. Less so if it's outsourced, but without actively 'tending the garden', it will inevitably disappear at least from the public-facing internet.

The ephemerality is also exacerbated by Google / Facebook account terminations.

It really is an interesting balance between "The internet may never forget" and "The internet may never remember".

The internet remembers everything applies much stronger to text than to rich media. I can still find a mention of myself from a 1983 newspaper article that has been duped across a bunch of content harvesters (and also on the way back machine’s view of the newspaper’s site).

One layer of the internet is about packets, another is about ordered sequential streams — and both are widely supported protocols. Above these could potentially sit a dissemination layer, retaining and persisting anything of interest. But in a legal system where information can be speech or property, it is problematic to do that as a standard protocol. Instead we have archive.org, The Archive Team, BitTorrent, and a million progressively less durable options from there, including some of us running youtube-dl on videos of interest, grabbing PDF’s off Arxiv, SSRN, etc. I’m not sure what to do about any of this except a yearly donation to archive.org, and accepting the great distributed fitness function of fate.

Wonder if archive.org and the EFF could get a video site up to compete with youtube, or work to back something like bitchute or otehr alternatives to raise awareness and push back on overreach takedown requests. YouTube/Google just seem to cave, and apparently so does GitHub.

So does everyone who got owned by a multinational megacorp. That's when all the human vision goes out the door. Because there's no mechanisms to keep it there, it will eventually wither.

We could have had so many nice things if it wasn't for those things.

It's almost as if there's somewhere a country which acts as a fertile breeding ground for these uncontrollable unregulated monstrosities, to grow really big, and at some point left roaming wild for other countries to deal with "yeah we don't do that here" (by which I mean stuff like false advertising and reasoning like "if it's legal and I make a profit then I get to do it").

I've been distraught at the same thing. I had a few playlists that I grew over time, and it's annoying to go back and see gaps of "[Deleted video]" where your favorites used to be. I'm curious to know if you've found a solution for this.

Personally I've started locally saving anything I think I would want to watch or reference again. When favoriting something or saving it to a playlist, I download it too. I started off with using youtube-dl on the command line and have experimented with the "Import from YouTube" function on Peertube but ended up with a tiny cobbled-together video platform, importer (youtube-dl wrapper), and Firefox extension. I think the next step will be easily saving non-video things - probably recording WARCs?

I beefed up my browser cache to remember everything and have a server in the background downloading any youtube channel I visit. Now I know that anything I've ever looked at via the browser is now accessible offline.

One (poor) solution I did was to add in the description a list of the videos in the playlists, so once a video was removed, I could just add it back.

I did that with only one YouTube playlist, so it wasn't too much work.

Wish I had thought of #2 long ago. The majority of videos I've saved are long gone.

#1 is an activity protected by fair use, and something fairly important. Imagine a world where you can't reference images or clips in your journalism or reviews, forever having to rely on only licensed material and descriptions.

Fair use does not override the TOS which YouTube-DP violates

You can’t issue a DMCA takedown for software that violates someone’s TOS. (Never mind that it’s someone else’s TOS.)

It wasn't a DMCA takedown notice. It was a DMCA section 1201 notice. Section 1201, which criminalizes circumvention of access controls, has nothing to do with copyright.

John Deere uses it to go after sellers of unauthorized replacement parts for their tractors.

Non-compliance is a felony carrying a 5 year / $500k punishment. Unlike takedown notices, there's no counter-notice opportunity for the receiving party. So if the receiving party believed the notice isn't valid, they'll need to argue that as a criminal defendant in criminal court.

Wow, by that logic news reporters aren't even allowed watch videos on the site for news reporting purposes because it would violate the ban on commercial use.

I don’t think that follows.

A TOS is not private law.

Please quote relevant TOS.

B. Content is provided to you AS IS. You may access Content for your information and personal use solely as intended through the provided functionality of the Service and as permitted under these Terms of Service. You shall not download any Content unless you see a “download” or similar link displayed by YouTube on the Service for that Content. You shall not copy, reproduce, distribute, transmit, broadcast, display, sell, license, or otherwise exploit any Content for any other purposes without the prior written consent of YouTube or the respective licensors of the Content. YouTube and its licensors reserve all rights not expressly granted in and to the Service and the Content.

IANAL, but for me the terms used here are kind of problematic on a technical level.

- "unless you see a “download” or similar link displayed by YouTube on the Service for that Content." -- If there is a download link somewhere in the comments you're allowed to download using youtube-dl?

- "provided functionality" -- either functionality is available or it's not. Whether there is a button for it on the UI or not is entirely different. (However, this sort of seems to imply that the HTML specs are legally binding?)

- Your internet connection is only capable of sending and receiving packets, and it isn't wrong at all to describe "sending" as "uploading" and "receiving" as "downloading".

- It says that you aren't allowed to (among other things) "display" any Content for any other purposes -- other purposes than what? "For your information and personal use"?

- It used to be that you could just copy video files from your browser's cache directory, or you could use hard links -- then you wouldn't even have copied them technically. (Not sure if that's still possible on youtube -- but again what if you had a browser that generated PNG files instead or or before "displaying" stuff on the screen?)

Many TOS are full of legal BS that has not been tested in court.

"You shall not download"

Or what? Can Youtube actually sue me for damages, they haven't suffered any. Some videos are lisenced as Creative Commons. I seriously doubt this has any legal force.

Given Google amakes both Chrome and YouTube, the lines are pretty unclear on what provided functionality is.

1) Downloading videos to use later in editing for analysis pieces / editorial videos.

I wonder if certain famous YouTubers used youtube-dl for this purpose. How else would you download, edit, and analyze a video's editing or special effects?

There’s plenty of online tools and browser extensions to accomplish the same task (I wonder how many of those use YouTube-dl on the backend)

Browser extensions are very similar, just as an extension instead of direct. Unfortunately, it's difficult to use non-store extensions in Chrome, and Google pretty much blocks anything that allows youtube downloads from the store. youtube-dl is likely the most popular option for this, and even then, the same (deeply flawed) logic could be used against any of them.

> I wonder how many of those use YouTube-dl on the backend

i wager to say that the vast majority of browser extensions and websites that allow downloading of media content use ytdl in the background

On Twitch streamers have a browser window open and pause it to comment.

HDCP protections added to Chrome in 3...2...

edit: I only meant this in so much that it's entirely possible for Google, that controls both Chrome (the biggest browser) and YouTube (biggest online video site) to encable DHCP restrictions on video capture from within the browser for sites (such as YouTube) that might implement it. Given their proclivity towards giving the RIAA/MPAA whatever they ask for, it wouldn't surprise me.

> DHCP protections...

I'm guessing you meant HDCP?

I'm pretty sure that encrypted media sites already use HDCP because my projector glitches for a few seconds sometimes when I first navigate to Prime Video, and whenever I have a secondary analog VGA monitor connected Prime Video will only play in standard definition.

That was what I was thinking, edited, thanks... Half surprised they haven't done this to some youtube videos/channels already then.

Any rebroadcast is already against Youtube TOS. It's only for personal use.

There's fair use though. It makes sense that you are able to quote a video. Its just that those people on Twitch don't quote; they play the whole thing. Captions with bookmarks would fix that (Pornhub has that natively, quite useful).

My primary use case has been to download talks and watch them later. I frequently used to work out of cafes (not since Feb of course) where the connection may not be reliable. Youtube-dl was very helpful in those cases.

I used ffmpeg and audacity with youtube-dl to download and do file and spectrograph analysis on a faked political video. That analysis showing how the audio was spliced together may have killed the story before it got legs.

In an era of fakes and deepfakes, we need tools like youtube-dl more than ever.

> In an era of fakes and deepfakes, we need tools like youtube-dl more than ever.

I've been working on deepfake detection for the past two years. I use pytube3 and youtube-dl to scrape hours and hours of footage, not just youtube, but cspan, news sites, anything I can get my hands on. I have ~100 hours pulled so far, at multiple encoding rates.

Facebook recently changed their API and now the best I can scrape easily is 240p. That's insufficient to detect the artifacts the models need to train on. I can only pull 1/3rd of cspan videos I can watch through the viewer because again, api changes. It's a constant cat and mouse.

I'm not exaggerating when I say that if I don't have those wonderful folks out there keeping these tools and test suites up-to-date, this project is massively disadvantaged.

I don't have any public release assets at the moment, but you can read about Siwei Lyu and Hany Farid, who are leading the charge in deepfake detection.



Did you post an article about it? I'd be interested to read it.

This is probably one of the coolest and most underrated comments in this topic!

Could you tell us more about this process?

Thanks, it's way less cool when I write spectrogram instead of spectrograph. My mistake on that part.

If you use those tools, it's really straight forward process of using the tools. I didn't know about Audacity until I had spent most of my time figuring out ffmpeg/ffplay. It was an amateur effort, but I've done a lot of security analysis.

That's freaking awesome

When my full-time job was as a journalist, I used youtube-dl all the time as a way to have archives of content that was either in danger of being taken down (related to a mass-shooting or other global event) or that the creator would take down after it received attention. In fact, I used to have instructions for how to install it in the various wikis/docs for other team members (this was easier once the Python GUI frontend became available, but I still created step-by-step instructions for specific settings).

I produce/host a few weekly news shows focused on developer-centric news and updates and have a keyboard macro setup to run youtube-dl against a text file with URLs not to download the videos, but the thumbnails for videos, so I can use them in the on-screen graphics when talking about a specific story. Before I scripted that solution, the process of having to manually extract the thumbnail for any video I was highlighting was a major PITA.

Also, being frank, youtube-dl is significantly better than the official YouTube API for downloading past content from channels I own/manage. It’s faster and a lot more scriptable. I have automated scripts set to watch specific playlists or channels and auto-download stuff for archival purposes — and again, this is content that I either own or that exists on a channel where I’m one of the admins.

It’s unfortunate that the test suite had links to commercial music — we’ve seen in past RIAA litigation that that is enough to go against the argument that this isn’t encouraging download of copyrighted content.

But as a tool, for not just YouTube but so many other services, it’s invaluable even in non-data hoarding/grey area or straight up infringement scenarios.

To be clear, I’ve often used youtube-dl to infringe (as have the vast majority of its users), but that’s my choice/fault. The tool itself has plenty of non-infringing uses. I just wish they’d either linked the test file to another area or used other examples.

YouTube even offers a Creative Commons feature, such as on https://www.youtube.com/watch?v=XUY01kW8p2o: if you scroll to the end of the description it reads: License Creative Commons Attribution license (reuse allowed).

I've never understood why they don't offer a download option for such videos: the likes of youtube-dl is the only way to get a copy, and it's perfectly legal.

Likewise for public domain videos.

I don't think it's any more complicated than that it's because they can't roll ads on videos that are downloaded.

youtube-dl et al are niche enough that it doesn't put a real dent in their ad business, but making it more convenient with a Download button might.

> I don't think it's any more complicated than that it's because they can't roll ads on videos that are downloaded.

That's certainly something they can do; they just don't do it.

Compare movies on VHS tapes, which all roll ads before actually playing the movie.

Sure, Google could do but there's more than enough problems vs reward to make it unattractive:

- When you fast forward VHS tapes you still see the adverts. When you time seek a digital video file you don't.

- It's harder to do targeted advertising (since you no longer know if the downloader is also the viewer)

- Adverts baked into a video file can't be changed out later for newer adverts after the user has downloaded the file

- There's no way of having users "click" those adverts, let alone report back to Google that they've engaged with it

- Some of those aforementioned points might also have a knock on effect with how advertising is charged, which would mean Google would need another pricing model

...and all of this is before you even address the technical issues of video files having multiple different bitrates and formats. You could probably generate a new file on the fly but that would be extremely computationally expensive. So they'd rack up more costs hosting the service as well as lose money on advertising.

Honestly, I can't blame Google for not supporting a download option.

What do you mean by "technical issues"? Google already has a streaming video player supporting multiple bitrates and formats. They could just do whatever youtube-dl does to pack it into a file or even more optimally, grab it from their video database directly should there be something like that.

I'd already explained the problem, albeit without much detail.

The adverts are a different file to the video files. So you need to either merge them ahead of time or dynamically splice them in real time.

Ahead of time: Now you need to not only have a video asset for each bit rate and video format but you need to multiply that with the number of adverts on any given day. This would result in thousands of files, maybe even hundreds of thousands once you take targeted adverts into account. And they'd need to rebuild their entire catalogue whenever an ad campaign ends. Clearly that isn't going to work long term.

Real time: Option 2 is to stream the advert then follow immediately with the video content on the same HLS stream. This is much more achievable than option 1 but you are then running the streams like a live TV service where you're dynamically splicing content into existing streams. It isn't difficult to do per stream but it is extra processing compared with the existing set up. The real problem lies with scale. A TV broadcaster might do this with a dozen to a few hundred channels, each with contracts ranging from thousands to hundreds of thousands. Youtube have millions of streams, each which earn pennies from advertising (and even less if they had to download because of the change in advertising model -- as I also described before). So they'd have to pay more and earn less. The financials simply wouldn't stack up.

Before you comment that YouTube and Twitch offer live streaming services based around the same advertising models, yes they do, but they also operate using the same stack as the pre-recorded videos because those video streams don't need to be spliced by YouTube / Twitch (any editing happens by the content creators before it hits YouTube and YouTube can show the adverts before or even between the videos by injecting them in the browser (ie so the splicing is managed at the front end rather than on Googles servers). None of this is doable with a "download video" button.

Exactly, Google has that player, but to replicate the same effect for offline playing you would need to play a playlist or a DASH video and multiple fragments.

But I grant it's technically possible to do with a single MPEG file, just not trivial. You would need to prepend the advertisement to the video bitstream (re-encoded with same parameters such as width, height, FPS, intra picture interval) and then do the same for audio, while ensuring the video/audio synchronization is kept intact (MPEG edit lists are not well-supported by players, such as by ffmpeg), while preferably not re-encoding the actual video. A bit more complicated to do the insertion in the middle of the file.

It's possible to do, but I don't think it's something that any off-the-shelf tools actually do.

Yeah there are off the shelf tools available to the broadcasting industry. But these things are designed to tens of streams, not millions. As usual, it's a scale problem rather than a code problem.

But you cannot bill per view or click.

Not directly, perhaps, but the Law of Large Numbers says that views-per-download are very probably statistically derterminable, and can be estimated with high accuracy.

Otherwise, this is the same problem print newspapers have; how many readers per copy, seeing what adverts, how many times? This can be modelled or assessed (e.g., by reader response codes / instrumented URLs) fairly readily.

They could certainly bake ads into their videos if they wanted to.

Last time I was messing around with youtube-dl I was surprised by the huge size and variety of formats for just one video. I can’t think of an easy way to constantly embed new ads into all those video files. And can they charge for an impression without knowing if the video was ever watched? Maybe its just too much trouble.

Make it part of the Premium subscription then (which is ad-free). If people or companies have a legitimate use for downloading videos, paying $10/mo (or however much it is) isn't going to break the bank.

Google has zero interest in free downloads. They are in bed with the music industry which makes them some ad dollars and their client thinking is sub-zero. Thats because the public is not the client, the advertisers are. If I worked at Google I would be scratching myself behind the ears and take a long big look at what the company has become.

I suspect there's at least a small dose of ultra privilege of googlers involved too.

The same way governments and bureaucrats fuck up things like Covid responses, because they can't even imagine people not having 9-5 office jobs which are totally suited to WFH, and then blithely implement disease protection schemes failing to account for low income and precariously employed people, many of whom are working 3 jobs to pay the bills (including stuff like cleaning jobs where they work at 4 or 5 different old people's homes) - and then wonder why outbreaks spread so fast.

Googlers have probably forgotten the olden times, when they didn't have Gigabit connections to their pockets and lounge rooms, and don't even remember last time they wanted to "save a copy into a hard drive locally, like a hilarious boomer!"...

I want to agree with your cynicism.

But I lived in Mountain View. Decent internet is more or less not available there. Your choices are comcast or DSL.

Maybe they just don't remember the last time they had a waking hour out of the office ... :)

It's clear to me that whomever is making these product decisions has a very different relationship with computing and their data than I do-- but whatever the reason is, I don't think it's because they have a much better internet connection.

Youtube makes money on advertising. You can’t put modern ads into a downloaded video. I don’t think the ultra-privilege of Googlers factors into this particular decision.

(Also if you live in Mountain View, you’re stuck with a Comcast cable modem and a 38mbps max upload speed even when paying for their “gigabit” plan. But you'll only get 10-15mbps during the peak hours of the work day.)

> You can’t put modern ads into a downloaded video

youtube-dl has an experimental --include-ads option =)

> Googlers have probably forgotten the olden times, when they didn't have Gigabit connections to their pockets and lounge rooms.

Sure, that's why projects like Project Loon were launched by Googlers.

I don't think the CC video thing is even ads related. Just that most videos are not CC and no one thought about creating an exception for CC videos.

The copy-protection MAFIAA (a joke acronym, Movie and Film Industry Association of America, a hypothetical super-beast merger of the vicious & loathed RIAA and the MPAA) spent a while going after BitTorrent itself too, in a similar manner, as a potential tool for piracy.

Thankfully plenty of people were also using BitTorrent for things like distributing Linux distributions & creative commons material.

That, again, may perhaps save the day.

I use youtube-dl to archive my recordings of my own performances which have been put online by third parties (and which no one but me has any rights to!).

I use youtube-dl to view videos on other sites (such as vimeo) that don't display in my browser (due to compatibility or privacy settings), usually product support stuff.

I use youtube-dl to collect evidence about cryptocurency scammers, which often gets removed from youtube by the time the scam starts collecting. By collecting it when its online it makes it possible to hand over to lawyers so they know what to go subponea. At least once a video I saved probably personally saved me from a frivolous lawsuit (attorney fired his lying client after sending me a threat letter which I responded to with an archived video).

I sometimes use youtube-dl to buffer longer pieces which I'm watching which might suffer from infuriating connection interruptions.

I don't use youtube-dl to save copyrighted music (except incidentally), among other reasons: the audio on youtube is often pretty low quality.

Another valid use-case is to use the hardware acceleration with your native video player. AFAIK, mpv uses youtube-dl under the hood to play youtube videos. For a long time the video acceleration in browsers wasn't very good for Linux users, so we used players like mpv instead to watch high resolution content.

it still isn't. I can't watch 720p videos on my thin client with browser overhead, but I can with mpv.

What browser are you using? Google has been known to make YouTube perform worse on browsers that aren’t Chrome.

Actually Chrome on Linux doesn't use hardware acceleration for video playback and never did, while the most recent Firefox almost does, if one turns a couple of experimental knobs.

Off topic but I just tried to move from MacOS to Pop!_OS and this is the main reason that I can't migrate now. No GPU Accel on Chrome means scrolling is horrid on QHD/higher resolution (besides YouTube being pretty much useless). Firefox is fine but there's no user profile modes to separate between my various profiles. With so much work being done on browsers, it feels weird that Linux is so behind on this. And Chrome seem to not want to implement the GPU acceleration at all. I guess I'll pay the Apple tax for the foreseeable future.

Well, I am not sure I understand what you mean. There are two distinct GPU Acceleration technologies at work here:

- for videos

- for website rendering

AFAIK the video acceleration part is the big problem as the issue was ignored for many years by the browser vendors. For rendering websites most browser should do just fine, some hardware-browser combinations might be exceptions though.

I also wonder what you mean by Firefox has no user profiles, because Firefox has a profile concept too. Be default it doesn't ask you, but when you start it with `firefox --ProfileManager` you will see what I mean. I don't like messing with the Profile Manager every time, so I just have different Icons for starting different Profiles (e.g. Music --> Spotify, Movies --> Amazon & Netflix, etc.).

Currently there are just two websites for which I have to use Chromium:

1. Geforce Now, officially Linux is not supported, but with Chromium it does work when you change your user-agent

2. Binance, for whatever reason their login capture doesn't work with Firefox

OK so I'm not sure what's happening with the scrolling. But it's definitely smoother on Firefox than on Chrome. I'm sure it's not my hardware/OS combination (1060 + PopOS w/ Nvidia blobs).

On the profiles part: I am aware of the --ProfileManager option. But the icon is new to me. Should be worth trying. Thanks. Work uses G Suite so I don't know how Meet + some other tools react with Firefox.

I prefer using Firefox anyway, but if you want to dig deeper regarding your Chrome GPU acceleration you might find the following pages valuable:

chrome://gpu/ -- here you can see the current status of your GPU Acceleration

chrome://flags/ -- search for 'GPU' or 'accel'. This is the place where you can override some related settings, but be aware, that this might break your browser, so better backup your browser configuration directory before changing anything ;-)

From what I've heard, this is used as a replacement for profile modes on Firefox. https://addons.mozilla.org/en-US/firefox/addon/multi-account...

Ah yes that's it. I've read about it sometime ago and thought it was built it. Might try it out. Thanks!

I tried firefox, chromium and chromium-vaapi (which gives a slight improvement).

Hardware is a HP T610.

Firefox added hardware acceleration support very recently, June for Wayland and end of August for X11, and needs to be manually enabled plus YouTube needs to be told to use a supported codec. It's also possible your distro isn't supporting it yet.

Yup, with MPV I can watch video on a Pentium M without any dropped frames, and this is on a computer where the web is largely unusable.

It still isn't. The Youtube player is garbage and full of tearing.

also a question of codec.

h264 is still the only widely supported in-hardware decodeable codec out there (cheapest for the client) but google serves webm by default (cheapest for them and built in-house)

ytdl-format lets you choose

I'm concerned that the "internet of videos" is not being properly archived, like the hypertext one is. I use ytdl primarily to save valuable homemade videos that I'm afraid could disappear if that person, or Google, were to delete or block the account. I'd hope this could be a coordinated effort by libraries and institutions interested in preserving this invaluable and unprecedented corpus of daily life, culture, spoken word, music etc. that could be studied for generations to come.

And I couldn't care less about VEVO and other copyrighted material, which, I wish, could find home in a completely different platform for all I care.

archive.org archives YouTube videos, too! They might not have everything but I have used it more than once to salvage an old video for which I used to still have a link but it was removed by the uploader...

I left a comment about my own non-infringing use cases of the tool earlier — and I think it’s important that the non-infringing use cases get mentioned.

Having said that, can we please stop pretending like any of us weren’t using this to infringe copyright? Most of us weren’t distributing anything, but I think most of us were aware this was at the very least a grey area, if not outright infringement.

I’m a huge fan and user of youtube-dl and think this action from the RIAA is ridiculous (though not surprising — I’m a little surprised it took this long) and that the law over this stuff is absolutely bonkers — but the commentary and false pretenses about not just how we use the tool but the purpose for why the tool was built is incredibly disingenuous.

It was built and designed to download content that the creators or sites that host the content either didn’t want people to download or outright didn’t allow. The program has support for username/passwords for TV Everywhere SSO’s for premium services. There are ways to tunnel in via a proxy to avoid region block downloads. Again, I’m a huge fan, and I’m someone who absolutely used these features, but let’s not pretend the purpose wasn’t exactly what it is.

That doesn’t make the RIAA’s actions any better or anything — but I really dislike pretending like we weren’t all using the program for the exact purposes the complaint laid out, or that they use case wasn’t the primary reason this tool existed.

In my country it's not illegal to download copyrighted movies/music. It's just illegal to share them.

So how was I infringing copyright by downloading something from a public website where copyright holder left the content publicly accessible to anyone?

I'll give you that it may or may not be against ToS of the distribution service, but that's all. But it's not a distribution service causing the ruckus around youtube-dl, but the copyright holder, so that's kinda beside the point.

Thank you so much for the thread here! I would like to share the service on https://streamingsites.com where you will find many good paid and free streaming sites and apps with cool movies to watch. It was really helpful to me, I have to say. I am sure that you will like it a lot too, good luck with it!

Which country would that be for the curious?

What even is the relevance of this? So host a youtube-dl repo and reject DMCA requests then. Github, as it so happens, is American so it has to listen to America.

Relevance is "Having said that, can we please stop pretending like any of us weren’t using this to infringe copyright?"

I'm not pretending if I'm not violating, even if I'm downloading the copyrighted videos.

Well obviously it doesn’t apply to you then. I feel like this is really easy to understand.

The word "any" has a specific meaning. Most people in the world aren't American and most countries at least from my experience allow private copying.

> most countries at least from my experience allow private copying

Most countries in the west allow it in the American sense. Only a very limited few countries allow it in the sense of Spain and Switzerland.

Hmm, where do you get that from? All of Eastern Europe pretty much, Scandinavia, France, Canada, Spain and Switzerland, at the very least, allow you to copy copyrighted content you legally have access to legally.

It's not a "limited few". I think most of the Eastern Europe allows downloading copyrighted content for personal use.

I genuinely can’t believe you’ve decided to do this. I’m ashamed to have even participated in this. I’m outtie.

The only thing I use it for is via Tartube to download videos from my young nephew's channel. He's enthusiastic about making YouTube videos, but doesn't understand the nuance of copyright law or the ToS.

For example, the first thing he did after I helped him set up OBS was to fire up a game + Spotify to make a live stream. I had to explain to him why he's not allowed to stream music like that. He doesn't do it now, but there's still a decent chance his channel that gets 5 views per video will get struck or banned for a similar misunderstanding in the future.

He doesn't know much (under 10) beyond using the Windows Game Bar or OBS to upload something and doesn't have a bunch of spare storage, so he doesn't keep copies of videos anyway. I download them so if he gets banned I have an archive and can help him set up a new channel on another platform.

I know it'll be easy for someone to say "do X, Y, and Z" then, but the reality is that it doesn't take much for it to become "not fun" for younger kids and the opportunity to encourage them to learn valuable skills like video editing, etc. are lost when they lose interest.

AFAICT I have not used youtube-dl to view videos that I couldn't watch on the site (except perhaps incidentally) - but I don't like executing random code on my computer and youtube web is in my view actively user-hostile and last I tried it had serious performance problems.

I had a feeling I was using youtube-dl to avoid ads (not blocks), but after some retrospective, I don't think I would have seen ads on most videos I watched from youtube and I think most were freely available videos, often embedded on author's pages - except people have an understandable habit of hosting them on youtube.

To conclude, I would disagree that we are all using youtube-dl as complained.

To add to this (which I 100% agree with) is the chorus of voices suggesting that their usage was OK because they weren't downloading RIAA content.

Just because someone isn't an RIAA member doesn't mean they don't have copyright over the content they upload to YouTube. There are some YouTube creators that would be totally fine with you downloading their content but you can't assume every creator is the same. Just because a user filmed, edited and uploaded some video from their own house doesn't mean they have any less rights than music or video from major labels or Hollywood studios.

Yes but a tool is not illegal just because it can sometimes be used illegally. If some are OK with it and some are not, then it suggests it has legitimate purpose (though you need to go a lot deeper).

Note, I'm not arguing copying small-label stuff is OK, I'm rather more OK with copying RIAA stuff.

AFAICT, if a DRM prohibits fair use, breaking it is legal. In some jurisdictions. Or at least it should be. That should be enough of a reason for youtube-dl to exist.

To be clear, I'm not arguing that it is illegal or that it should be -- clearly that it shouldn't! I'm simply saying that the program was designed to download content that the site and/or creators didn't want to be downloaded and that the vast, vast, vast majority of people using said program were using it to circumvent those restrictions.

Again, if anyone is guilty of anything, it's the users (like me) who use the program, not the program for merely existing. And frankly, since I don't distribute the content I download or profit from it in any way, I welcome the RIAA coming after me, just as I did 20 years ago when I was a high school student using Napster on her dial-up connection.

There are creators, who are ok with downloads.

> but I really dislike pretending like we weren’t all using the program for the exact purposes the complaint laid out

Speak for yourself. I used it to pull down youtube videos to watch on a 14 hour plane ride over the pacific.

Proxy support is a standard feature of all browsers. It's a necessity because in some network configurations you don't have direct access to internet, but only through a proxy.

I wasn't, and it's not. I genuinely think you're in the minority here (which is unusual, on the 'net, where piracy's concerned).

I don't think this necessarily infringing copyright. There were court cases around timeshifting / tivo that OKd this.


As long as you don't rebroadcast copyrighted material, you should be fine.

If I'm seeking out to infringe upon a copyrighted video, Youtube-DL isn't even in the top 5 of things I would think to try.

Imo... people should be able to download stuff for free if they’re able to see it for free. The “you can see it but you must watch an ad” model is weird and leads to unenforceable/wacky situations like this

(I’m not really a stakeholder tho so who cares what I think)

> The “you can see it but you must watch and ad” model...

... funds OTA television. Also funds YT.

With OTA, it is okay to record OTA broadcasts using a VCR, so I'm not sure why youtube-dl would be different.

DMCA section 1201 is not about copyright infringement. It’s there to make things that are not copyright infringement illegal.

This is how John Deere got away with prohibiting farmers from repairing their tractors, etc. This is why time-shifting and format-shifting were just fine for OTA/cable television/CDs/etc but are practically nonexistent today.

The different, Google would argue, is that YouTube-dl doesn't download the ads, while a VCR captures ads and content alike.

But really, if the VCR were invented today, it would be made illegal.

Not like they didn’t try to make it illegal originally.

“I say to you that the VCR is to the American film producer and the American public as the Boston strangler is to the woman home alone.“ - Jack Valenti, former MPAA president, in 1982, explaining to Congress why the VCR should be illegal

I think that quote really shows where the organization's priorities are with respect to individual rights.

I mean, they tried desperately to make it illegal in the 80s. The Betamax case went to the Supreme Court! [1] Given the current makeup of justices, I would imagine Betamax would be decided differently today, in favor of Universal rather than Sony. Aereo lost 6-3, even though the dissenting justices cited Betamax from 30 years previous.

ReplayTV was sued and went out of business before it could be fully adjudicated. Dish was sued over its Hopper DVR that auto-skipped ads and ultimately had to settle. TiVo won some of their lawsuits but because it didn’t do automatic commercial skipping of the recordings, was spared her wrath of the networks like Replay was. (TiVo was ultimately ruined by the cable companies who sought to introduce their own inferior DVRs that they could charge a monthly fee over.)

[1]: https://en.m.wikipedia.org/wiki/Sony_Corp._of_America_v._Uni....

> Aereo lost 6-3, even though the dissenting justices cited Betamax from 30 years previous.

I remember that dissent (authored by the late Justice Scalia—RIP). Here’s the citation in question:

> We came within one vote of declaring the VCR contraband 30 years ago in Sony [v. Universal]. The dissent in that case was driven in part by the plaintiffs’ prediction that VCR technology would wreak all manner of havoc in the television and movie industries.

> The Networks make similarly dire predictions about Aereo. We are told that nothing less than “the very existence of broadcast television as we know it” is at stake. Aereo and its amici dispute those forecasts and make a few of their own, suggesting that a decision in the Networks’ favor will stifle technological innovation and imperil billions of dollars of investments in cloud-storage services. We are in no position to judge the validity of those self-interested claims or to foresee the path of future technological development.

I wouldn't take Aereo as a guide to how justices would reconsider the Betamax case were it heard de novo. Aereo's business model was to redistribute over-the-air broadcast TV, and it was arguing that its technical implementation was technically not violating any laws. This cast it into the creative reinterpretation of law scenario that tends to lose out in court.

It's not like they didn't try their damnedest to make the VCR illegal when it was invented.

Not if I paused recording during the break like I did as a kid

You can easily splice out the ads from DVRs/VCRs.

I even have a VHS VCR that does this automatically. It’s fascinating - once the recording is done it rewinds and watches the show at high speed, when it sees something in the signal that indicates that an ad was spliced in it marks the start and end. Then when playing back it fast-forwards over the commercials.

It worked brilliantly, back in the days of analog cable TV.

Remember TiVo?

Recorded TV and cut out the ads.

No it didn’t. You could skip ads with the remote at set intervals (notably, TiVo did not let you auto skip the ad even though they could have, but put 15 or 90 second intervals you could customize on the remote to get past the ads very quickly, but it recorded them). ReplayTV, which did more directly skip ads (via its remote, not via recording IIRC), was sued out of existence.

Dish’s Hopper actually DID do ad-free recordings and it was sued too (but there was eventually a settlement), not just for skipping the ads but for distributing the same copy of a recording to multiple users over a server. Cablevision, likewise, was sued (but settled) over the fact that it had a cloud DVR network that allowed the same copy of a recording to be available to multiple people, rather than storing X-Copies of the recording on their servers.

Later model TiVos can skip the whole break with a single button, newer models can auto skip https://explore.tivo.com/how-to/SkipMode

I never thought of the VCR comparison before and it's interesting. Is recording a video delivered over IP different than recording a video delivered OTA or via cable (which is probably IP these days)?

It's strange to me because I've been using adblokcers for as long as I've been watching things online, as do a large percentage of people, I'm sure. So does it really matter that much to YouTube when I download some video, considering me watching it brings it literally no benefit? What's the gain here?


Intellectual property law is about controlling what you can see, and do, with your own eyes, ears, and hardware.

If you stream a video on YouTube, YouTube can arbitrarily take it away from you whenever they want. If you download a video from YouTube, it's yours to keep. Tools like youtube-dl take away the power YouTube has over you, and that's an unacceptable loss to the ownership class even if they suffer no financial losses.

Worth reminding people that Big Content Holding industry came all too close to getting VHS tapes outlawed.

> In 1984, the Supreme Court ruled that Sony could continue to sell its Betamax videocassette recorder, overruling the U.S. 9th Circuit Court of Appeals judgement that held Sony liable for consumers’ copyright infringement.


Same s*, different decade.

I am super curious, if this software is open source and the most damning aspect of this case is likely to come down to non-software specific factors like the inclusion of copywrited material (from big name artists) in the unit tests and README and such, what would be the result is someone with the repo just scrubbed the README, refactorfed the unit tests to rid them of any tests on copywrited material, changed the name to something generic like webvid-archiver or some such and rereleased it?

Indeed, it should be good practice for any journalist to download copies of online material they are reporting about (not just video, but anything), for the case that they are being deleted or altered. In the old world, you had a piece of paper or a document that you could store somewhere. An URL is not a way of storing the content, it can always disappear.

Youtube-dl had some good notes in their README about legitimate uses and an explanation about how they tried to avoid adding sites that specialized in infringing content, but they really did themselves a legal disservice by explicitly using copyrighted music video content as tests. That will not help their argument at all. Youtube-dl absolutely has important and completely legal uses, reporting being among them, but the law doesn't care whether it CAN be used for legal purposes. The law only cares if the tool was "primarily designed" or "marketed" or "has only limited non-infringing purpose."

On the other hand, I'm no lawyer, but I think there's a (possibly bad) argument that youtube-dl does not violate 17 USC §§1201(a)(2)(a) because it does not circumvent any technological measures to control access to any content. The video is transferred entirely in the clear to any anonymous users that request it. Youtube-dl makes no false claims or breaks no cryptography or encryption. What measure is in place that is being circumvented?

> did themselves a legal disservice by explicitly using copyrighted music video content as tests.

They had no alternative other than not having tests for that particular functionality. The tests were testing youtube-partner specific functionality: the procedure needed to download content from some youtube partners is different.

Moreover, the tests just throw away the downloaded material... it would be entirely reasonable for a court to conclude that there is no copyright interest involved in that activity at all and would be extraordinarily unlikely that any court would find it to be anything but fair use if they did conclude a copyright was involved at all.

If it hadn't been the test cases the RIAA probably would have dug up some old forum post advocating infringement by some occasional contributor. They were always going to argue something.

The fact that they had to have tests set to get around certain partners — partners who, it could be argued, had different provisions to try to limit the download of their content - works against them.

In retrospect, it would have been much better to link to an external repo or site or source for the test file, rather than to have that as part of the GitHub repo.

ETA: you’re right that they were always going to go after something, but the action for GitHub to take down the repo wouldn’t be arguable if the test file wasn’t in the repo. It’s possible they would have gone to issues/comments within the project (another reason, in retrospect, to host that stuff separately), but in this case there was stuff in the source code that the RIAA can reasonably argue would lead to infringement.

Not really, it's important that yt-dl be able to download partnered content to allow for their fair use.

Plus, the unit test IIRC only downloaded a 100KB (thus less than one second) chunk.

Downloading copyrighted videos, just like home taping, is legal. There was a supreme court case deciding that the latter is legal. I have yet to hear any reasoning why "home taping, but on the internet" would not fall under that same precedent.


Distributing a tool whose primary purpose is circumventing DRM is forbidden by the DMCA. The DMCA was signed long after the home-taping precedent, so that precedent doesn't really speak to what is and isn't allowed under the DMCA. Fair Use rights exist but if a law says "this isn't fair use", then it's not.

The DRM argument would be relevant if youtube had any DRM. It does not, so it isn't.

They allege that it circumvents a "rolling cipher", which (they say) is sufficiently DRMy that it comes under the law. The DMCA doesn't have much of a definition of what counts as an access control measure, but the "rolling cipher" that scrambles the download URL looks like one if you squint. Exactly what counts as DRM is less of a technical question than a legal one, so (not being a copyright lawyer) it's not obvious to me what the right answer is.


There is legitimately some obfuscation happening:


Which is what I'm afraid Google will eventually add to YouTube.

Because this is about distribution, not a home copy. You could copy and paste the test code into a terminal window and download the video. That’s not the same as a home recording.

> Because this is about distribution, not a home copy. You could copy and paste the test code into a terminal window and download the video. That’s not the same as a home recording.

...yes, it is? What difference are you imagining?

The difference is that my home copy is my own work and action. The distribution, in this case, via source code, allows anyone to download that copy.

If the code snippet that downloaded infringing materials didn't exist and the user had to find the parameters themselves, that's fine. But it does exist and was included alongside a tool that makes that recording possible.

A better analogy would be like, if you had a video store that had a machine in the store that allowed people to make copies of a video tape using a duplicating machine. If someone wants to buy a duplicator for home use and use it in their home, that's one thing (though there was litigation around that too), but if you go to a store and can plug in a video tape -- and the example in the store is a copy of a Disney movie -- that's sort of what this is.

(Photocopy shops and libraries often have fairly strict rules about what kind of content can be photocopied, for what it's worth. The guy at Kinkos at 3am might not care that you're copying an entire textbook on the machine, but if you go to Office Depot in the daytime, they definitely will. Kinkos was sued and lost over a photocopy case in regards to excerpts/packets/anthologies for textbooks.)

> A better analogy would be like, if you had a video store that had a machine in the store that allowed people to make copies of a video tape using a duplicating machine. If someone wants to buy a duplicator for home use and use it in their home, that's one thing

That's what's happening here. Github is selling (for free) recording machines.

If you have the recording device, then you can use it to record video that gets transmitted to you. But there isn't a place you can go to have the video duplicated for you. You do it yourself.

Downloading a video is not distribution. Uploading a video to others is.

Downloading a video is same as recording it to a VHS and serves the same purpose.

> The law only cares if the tool was "primarily designed" or "marketed" or "has only limited non-infringing purpose."

The first two elements have a mens rea implication behind them: you have to demonstrate the intent of the developers. That RIAA's strongest evidence of intent is buried in a unit test, and this seeming intent flies against the evidence in the far more prominent README that is discussing its unsuitability for infringing content. Furthermore, they can also demonstrate from their own practice in history that they refuse pull requests and close issues were people are clearly trying to use it only for infringing use cases.

[Post deleted after 60 minutes. It has had adequate exposure.]

The former text of this comment:

    The headline on the article should reflect the fact that Microsoft/Github pulled youtube-dl offline by choosing to respond to a DMCA takedown order that is not procedurally valid. The privatization of law enforcement through DMCA takedown orders does not extend to alleged circumvention tools; it includes only copyrighted material that is allegedly 'pirated.'
    The RIAA is engaging in an abuse of process that Microsoft has no obligation to cooperate with. The fact that Microsoft didn't tell the RIAA to go Disney themselves means that 100% of the blame lies at Microsoft's feet. No exceptions.

    . .

    That said, the youtube-dl maintainers should file a bar association complaint against the RIAA lawyer(s) who wrote the DMCA takedown. Filing invalid motions is an ethical violation that should result in sanctions.

I don't follow the argument that the procedural invalidity makes Github more culpable.

In a normal, procedurally valid, DMCA the only harm to the host for not following the procedure is that they lose the safe harbor.

In this case because it isn't a question of infringement there isn't a safe harbor at all.

In either case a company can choose to take something down because it wants to reduce its exposure to litigation.

I'm disappointed that even after acquisition github has continued with the beyond-industry-standard level of aggressiveness with takedowns. But I don't see how you conclude that the anti-circumvention complaint makes github more in the wrong for complying.

> is that they lose the safe harbor.

For that particular instance or permanently?

For the specific averred infringement.

Based on your response, and the parent comment, it seems like content hosts have a responsibility: to determine if a DMCA takedown request is valid. If a host doesn't see this as a responsibility, and simply complies with every DMCA (whether valid, invalid, or even fraudulent), I'd agree with op... the host is culpable whenever they comply with an invalid or fraudulent DMCA request. Maybe not culpable in a legal sense, but certainly they are to blame (otherwise who?) for complying with invalid DMCA.

They don't have a responsibility except in a goodwill sense.

The DMCA gives hosts an option to have legal immunity for each specific case of infringement, to gain the immunity they need only take the content down at the cost of outraging a potential customer. They don't have to take the option, however.

It is common in industry for hosts to simply discard obviously invalid complaints particularly if the target is high profile. The legal immunity isn't very valuable if the complaint is baseless-- (sure, there might be frivolous litigation, but that is always possible).

It also seems pretty common for hosts to allow the target of the complaint to counter-notice in advance of taking the material down and then just skip the takedown, or they take it down but restore it immediately on counter-notice (youtube itself does this, or at least did when I was hit with a spurious dmca complaint there years ago). Both of these procedures don't follow the letter of the law and arguably cost the provider their safe-harbour. OTOH, almost no DMCA complaints are actually valid per the specific requirements of the statute, so maybe they don't actually lose their safe-harbour.

I once worked for a business that got a DMCA request for taking down a legitimate site, from a law firm acting on the site owner's behalf. They refused to listen to reason and didn't stop until we got the site owner to call their own lawyers and tell them back off. Luckily, we were outside the jurisdiction of the DMCA so we had the option of simply saying no.

I can't imagine it was the only time such a thing happened, and that alone is an argument for simply ignoring those take-down requests if possible.

It's not Github/Microsoft's responsibility to make legal determinations about whether a particular DMCA takedown notice is valid.

I completely agree, on the other hand, that a takedown notice issued knowingly without proper grounds should be penalised severely.

> It's not Github/Microsoft's responsibility to make legal determinations about whether a particular DMCA takedown notice is valid.

Except, by choosing to honor it, they decided that it is legally valid.

The appropriate response to a copyright claim against material that the claimant clearly does not own is 'unless you can prove that you own this material, come back with a court order.'

Is it? OR is it the prudent thing to do to temporarily make it unavailable. Then take a second to look at things, then bring it back again if that the action deemed appropriate? If doing nothing leaves the vulnerable while doing something that gets reversed gives them ammo to go after the accuser, then it seems the lesser of 2 evils to me

I mean, to be fair, they probably have one of the most colossal legal teams of any tech company, no?

Regardless, I think they takedown without giving the "DMCA" complaint much scrutiny because the risk of penalty for NOT taking action is greater than just taking down the repo(s) risk-free.

Github's ToS: which says "we have the right (though not the obligation) to refuse or remove any User-Generated Content that, in our sole discretion, violates any GitHub terms or policies." ( https://docs.github.com/en/free-pro-team@latest/github/site-... )

I have to say it IS strange they took down content based on a letter that doesn't prove or even really suggest copyright infringement has taken place (only that it COULD, as a result of using the software), and their own ToS says "There may be legal consequences for sending a false or frivolous takedown notice." ( https://docs.github.com/en/free-pro-team@latest/github/site-... )

Under section 1201 anti-circumvention rules, unless Youtube-DL enters the US legal system & counter-notifies, Github/Microsoft face CRIMINAL liability charges with steep JAIL TIME and a half-million dollar fine.

I want to chew Microsoft out for this one, but the US-based copyright mafia have built themselves an unbelievably vicious draconian set of laws that prevent any sort of challenge, including those coming from scientific inquiry or examination. Anti-circumvention is the most world-hating corporate-owned-world flaming garbage that could be devised, and violating it comes with unbelievably mercilessly cruel criminal penalties.

The menace of this threat has silenced science & speech, has prevented mankind from examining the world about them, learning of it, & discussing it. We have outlawed knowledge, outlawed idea, literally criminalized knowing something about the world with steep jail time. This is a farce, of the highest order, one of the greatest shames the law has done unto itself.

Great recent thread from Doctorow on this: https://mamot.fr/@pluralistic/105090187888073250

I also recommend Bunnie Huang's "Why I'm Suing the US Government", https://www.bunniestudios.com/blog/?p=4782

Anti-circumvention laws must be stopped.

I feel like this line from your link to Bunnie Huang's blog is apropos to many of the conflicts of our time:

> Like the parable of the frog in the well, their creativity has been confined to a small patch, not realizing how big and blue the sky could be if they could step outside that well.

That was posted in 2016; does anyone know what happened with that lawsuit?

Did someone DMCA this post?

Annoying to say the least

> that Microsoft/Github pulled youtube-dl offline by choosing to respond to a DMCA takedown order that is not procedurally valid.... The fact that Microsoft didn't tell the RIAA to go Disney themselves means that 100% of the blame lies at Microsoft's feet. No exceptions.

What you said is clearly correct, but I don't think it's reasonable to expect any company to insert themselves in the middle of these disputes. At scale it's impossible.

Imagine you run a forum with millions of members and tens of thousands of posts per day. You get a stack of DMCA takedown requests every day. Do you think it's feasible for you to scrutinize each and every one, and take a moral stand on those you deem 'not procedurally valid' ? How much do you think that might cost? What happens when you call one wrong and the law comes to YOU?

For anything to really change, it's clear someone will have to go after the RIAA.. (EFF maybe?)

Given the number of fraudulent DMCA takedowns that are created by bots and malicious actors, they absolutely have to be individually examined by competent legal counsel.

At one point Microsoft was issuing DMCA takedowns against (then) openOffice.org mirrors, allegedly on the grounds that any large file with a name that contained the word 'office' must be a 'pirated' version of MS Office. Taking these takedowns at face value would have ended the openOffice.org project and absolutely could not have been respected in good faith.

These kind of abuses are ongoing from multiple parties, including bots, copyright trolls, and crazed ideologues who file completely fraudulent DMCA takedowns to force anonymous personalities to disclose their identities. There is no alternative but to consider DMCA takedowns on legal merits because so many of them have no merits whatsoever.

Having the funding to do this 'at scale' is just as much a cost of doing business as paying for power and connectivity. If hosting platforms don't want to do this, then they should purchase an amendment to the DMCA to impose criminal sanctions on entities that serve fraudulent DMCA takedowns.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact