Hacker News Reader

Comment by WarOnPrivacy | Hacker News Reader

I'd like to recap the issue here:

An Australian law was based on the premise that it was unfair for Google to reap benefits from sending traffic to AU media sites. The law also forced search engines to firehose user data (w/o implementing privacy protections [1]) to Australian media interests (eg: Rupert Murdoch).

Google said, "Ok. We'll comply with the law, forgo those benefits and stop sending traffic." Oddly, this did not satisfy the interests behind the law.

Microsoft stepped in and said "We're totally down with this law" - including, apparently, the egregious gift of user data to AU media interests.

[1] https://www.techdirt.com/articles/20200817/22554345133/googl...

View on HN

Self-responding - adding my position.

I find most of these arguments against Google to be reasonable, inc complaints of monopoly. Personally I strongly mistrust Google and have mostly divested myself from their offerings.

What I see Google did here was to vet the law's claims of fairness (Google gets money from our content so we should too) by stopping that arrangement. Once done, that particular unfairness was ended and the purported aim of the law was satisfied.

The grumpyness following Google's action confirms the law's intention is to create a new stream of tax revenue that AU Gov would grant (for now, anyway) exclusively to AU media entities.

I could potentially be behind a tax->news. Maybe. If it were constructed with more wisdom than this law is. I can not support this law because of it's shoddy construction (firehosing user data? srsly?) and because it's supplied purpose (ending unpaid usage of content) isn't truthful.

A larger issue may be that snippet laws seem to serve to further entrench Google's position by making news orgs dependent on Google for their operating revenue. This seems highly problematic.

It also grants new control over news funding to the AU Gov. If that control can be misused - to forward (AU Gov or 'partner' Gov) agendas - modern history tells us it will be.

Lastly, once this is normalized, who should pay next?

Another boneheaded provision on the law is the requirement to give advance notice on upcoming changes to the algorithm to the news publishers. This is broken for both the reason of not being able to deploy 'algorithm changes' without delaying them 14 days (PDF page 23), which might not even be possible given machine learning's always-learning algorithm, and for the reason of news outlets for some reason being entitled to information on algorithm changes before everyone else in order to get a leg up in search results over anyone not in cahoots with the search partner.

Law text: https://parlinfo.aph.gov.au/parlInfo/download/legislation/bi...

Sorry, what kind of machine learning is always-learning?

That can be how regular programming works (using live statistics), but ML is all about very expensive offline training, the model doesn't change as you execute it.

I think op might be referring to online learning [1].

[1]: https://en.m.wikipedia.org/wiki/Online_machine_learning

Right, but all the methods in that article are old. Deep learning is typically not online at all, although you can do some things to personalize it.

The monopoly powers Google holds over search have given them such a strong position, that Google is more than reaping benefits. It is in the public's interest to prevent such powers to destroy the 4th column of balance in democracies that journalism should strive to be.

This is not to say that Murdoch's news empire isn't bad as well, but no country should sit back and watch while Google and Facebook siphon money out of the countries.

The traditional way to do this would be something like building a domestic search engine and levying steep taxes on foreign ad buys, rather than this bizarre attempt at backhanded protectionism.

I don't know if you can say "traditional" so much as "historical". It's extremely difficult for software services to differentiate a domestic offering in the same way as physical goods - let alone when both countries share a language and the service has little cultural content per se.

Has a second-mover service ever overtaken a first-mover via on this strategy within in the Anglosphere? Or even within "the west"? (Xing vs. LinkedIn maybe? But everyone I know has both...)

Domestically building something like this and maintaining it is almost impossible. People have freedom and they will switch to Google (or a better site when it happens) irrespective of what Govt desires.

Journalism is a joke.

Nobody disputes that journalism as it currently exists has a lot of issues. But do you really think that the world would be better off without journalists?

If you said that the brakes on your car were "a joke" I would not understand you to be advocating for the abolition of brakes. Likewise, I'd assume people complaining about journalists want better journalists, not zero journalists.

I want better journalists for example - although I also don't have a good solution in mind for the economics problem that the internet gives journalism.

Since the world refuses to pay for journalism, it does seem that almost no one sees value in journalism.

I would say that 80-95% of journalism is working as intended to report what happening in the world and be investigative/critical.

If people want to watch Fox/read yellow press, that is not the journalists fault.

80-90% of journalism is either fluff or capital propaganda.

It's crass but youtuber friendlyjordies shows how terrible the aussie press is on a weekly basis

In the UK the royal family is put on the front page the same day the home sec is rulled to have broken the law with corrupt contacts

In the US no one reports Cuomo's covid failure until he got me too'd

"4th column"

https://i.imgur.com/o7VI2PA.jpg

You left out the part where Google changed their minds and decided to pay for news in Australia. The whole "Ok. We'll comply with the law, forgo those benefits and stop sending traffic." bit was an empty threat.

https://www.nytimes.com/2021/02/17/business/media/australia-...

Google's action did what it was intended to do, expose that the stated reason for the law (using content w/o compensation) was a lie.

It was unfair for Google to reap benefits from: (1) Copying story summaries from AU media sites, putting them on a Google site, capturing visitor data from that site (2) Sending traffic to AU media sites

It's more (1) than (2)

> Copying story summaries from AU media sites

I could be wrong, but aren't these summaries published voluntarily as semantic metadata or a feed from the media sites?

They are, media sites are also free to opt out of google indexing.

Indexing in Google's search engine to be included in response to user queries should be decoupled from being excerpted in Google's pretend newspaper front page.

I disagree about news.google.com being a 'newspaper frontpage'. If I wanted newspapers I'd individually visit each 'paper' online. What I value is a broad summary of the different news topics that exist at any given moment and links to more details. Through that generalized (and if I provide keywords, more specific) lens on the world Google's News offering provides value not found in any single publisher's biased view of the world.

If this was the intended use case Google would present it as a tool to enter a news article and it would show you all articles it believed were equivalent coverage. (i.e. what is now hidden behind "View Full Coverage" and without judgement about "Top Coverage" etc.)

Instead it is structured almost exactly like a newspaper front page, right down to the damn weather report.

“but weren’t they willingly paying the mob?”

In a situation where if you stop paying then the mob responds by ignoring you.

What is firehose user data? I've become an expert on 11,000kPa-pressure hoses used to extinguish fires, but still have no idea what you mean!

Typically used to mean "sending you everything", e.g. the Twitter firehose used to be an API that sent you all of Twitter live. I suppose this would just mean your analytics get every news view.

'Firehose' is meant to illustrate an enormous volume of data, in this case the personal data of millions of users. That's a huge amount and it would be delivered continually.

This is an absurdly biased and fact-free recap.

Any framing that posits that Google is benevolently "sending traffic" betrays that it should be discarded at the outset. Google (like Facebook and most other "make money on your content" empires) desperately tries to send as little traffic as possible to anyone, which is why they have AMP, scrape content and provide a dubious "expert engine", and provided a news dashboard where they summarized every news site.

I see this same ludicrous argument made re: Facebook. When Facebook temporarily blocked news, downloads of news site apps rose dramatically. It turns out that they weren't helpful whatsoever to anyone. They are a desperate suckerfish just trying to live off of the creation of other people.

Terrible recap.

The idea that Google is the one who gets to know who you are despite not creating the content that they are exposing to you is absolute nonsense.

Google is not the gatekeeper to the internet.

How does that relate to this specific law? Something can be bad, and the cure as bad or worse. Can we all agree on at least that?

Reply on news.ycombinator.com