Response to Google open letter • Hacker News Board

Re: Response to Google open letter

#241

Post by mthoms » Mon, Aug 17, 2020, 6:23 PM UTC

dodobirdlord wrote: post #234

corty wrote: post #120

Earlier quoted context omitted.

The "worse results" probably stem from the part of the law about changes in pagerank algorithms: If search engines are required to notify about changes in algorithm and provide mitigations for those changes, the easiest way to provide this would be: Keep the algorithm as it is, never change it again (for Australia) and accept the degradation that will result from that.

Also, what does this section even mean in the context of ML models? How do you communicate that you trained an interest-estimation model on more data and will now be using the upgraded version? I’m sure Google uses many such models for producing personalized results, and probably several are updated in production every day. Does Google have to share the inner workings of their search stack with these news companies?…

These are excellent points about the ML aspect that I haven't seen mentioned so far.

Re: Response to Google open letter

#242

Post by mthoms » Mon, Aug 17, 2020, 6:27 PM UTC

arethuza wrote: post #183

cutemonster wrote: post #140

Earlier quoted context omitted.

Similar to "strawman" in online forums, right

More "message discipline" I suspect: https://en.wikipedia.org/wiki/Message_discipline

Interesting. I wonder who coined that phrase? It's a curiously positive sounding phrase for something that's actually quite nefarious.

Re: Response to Google open letter

#243

Post by burkaman » Mon, Aug 17, 2020, 6:42 PM UTC

asdfasgasdgasdg wrote: post #227

BurningFrog wrote: post #219

Earlier quoted context omitted.

> that in no other case provides that level of autonomy. I, like all FB users, can delete comments to my own FB posts.

So can media organizations. They just can't pre-moderate the comments.

Where are you seeing a requirement to pre-moderate?

Re: Response to Google open letter

#244

Post by harryf » Mon, Aug 17, 2020, 6:49 PM UTC

JoshTriplett wrote: post #200

harryf wrote: post #146

Earlier quoted context omitted.

I'm well aware of NewsCorp, having grown up in the UK and seen them basically decide every election since perhaps Thatcher. Despite that, this particular point in the draft legislation in an interesting idea, at least to discuss. "Why Not?" was short for "Why not require Google and Facebook to inform affected businesses when they're making changes to their ranking algorithms?" ... that's a discussion that can be had…

Because the ranking algorithms are likely changed regularly, in opaque ways. Because many of the changes will be things like "reran the training algorithm on a new data set". Because it's a foot in the door to making them justify such changes. Because the purpose of a good search algorithm is to serve the people searching and not the people writing sites. Because it's an inherently adversarial process between search…

I agree with all the points you raise. But...

These have are all fundamental assumptions anyone working on the web makes and we’re established as fundamental at the point Google beat out its competitors, in the early 2000’s

At that time Google was a different company and didn’t have a monopoly, and the web was a different place. Now Google does have a monopoly and users browsing habits have narrowed to a few sites like Facebook and Reddit where they get “everything” content wise

So perhaps it’s time to question those assumptions and ask if they’re really still true in today’s context, or whether other things matter more now?

Re: Response to Google open letter

#245

Post by kmonsen » Mon, Aug 17, 2020, 6:57 PM UTC

ImaCake wrote: post #16

cblconfederate wrote: post #3

I feel like this time it's different , google (and big tech it represents) will get little sympathy on reddit and other watering holes. Even if the law seems restricting, it's Google who broke the social contract of the internet and turned from a benevolent rent-seeker to a feudal lord. (E.g. check this: https://www.google.com/search?q=money+for+nothing , google just pastes the lyrics and there is NO link to read the…

>conspiracy theories that too many paid commenters are among the crowd here I think HN just has a large pro-google crowd. I remember it used to be everyone loved google. So it shouldn't be a surprise that one of the remaining bastions of google public support would be a silicon valley startup community. I suspect a lot of people outside Australia are unaware of how respected an institution the ACCC is. Maybe that is…

I think it is just a lot of Google employees on HN, I used to be one of them.

Re: Response to Google open letter

#246

Post by screye » Mon, Aug 17, 2020, 6:57 PM UTC

specialist wrote: post #239

screye wrote: post #221

Earlier quoted context omitted.

2 reasons. 1. the quality of the recommendations is the real competitive advantage. It's like saying that all the sports scouting agencies need to fully reveal their algorithm to finding top talent to ensure fairness, when that exact algorithm is the thing that makes them a top scouting agency. 2. The company itself doesn't really know how the recommender works.(source: I built explainable ML models for a living). Th…

It's great that you're conscientious, seriously working on ML, big data, adjacent policy issues. Thank you. In the spirit of "yes, and...": Can't manage what we can't (or won't measure. So policy wise, we should not delegate judgement, and default to hippocratic until we know more. ML & big data is following the fad cycle, just like optimization, expert systems, etc. I'm 95% certain that >95% ML and big data is hooey…

> Thank you

I don't think I deserve it. I still work for a big conglomerate and most of the rest of my time goes into building the same algorithms that try to eek out that extra bit of accuracy. Although, we do have a lot of responsible AI, compliance, legal and interpret-ability committees we need to satisfy to make sure we aren't causing unintended consequences.

> foolsgold pathogens

I've flip flopped on this opinion, but over the last year or so, my opinion has solidified to being, "There is very real, massive untapped potential and value in the market, but most companies use it for marketing rather than solving the real technical problems at the core of it". I'd like to think we are the former, that solve real ML problems....but well, I'm exactly an unbiased judge of my own self.

I read this really nice paper that suggested grounding ideas of AI ethics in UN fundamental human rights, rather than hand wavy ideas like accountability and transparency. Since UN Human rights are much discussed, well understood and have universal-ish acceptance, it is far easier to take policy written from that POV and actually apply it to ML/AI.

I'd like to see something similar to the ACL2020 best paper "checklist", that creates a whole bunch of narrow statistical tests to validate certain traits of ML models. It'd be nice to have something like that, but for AI/ML ethics.

An openly available, standardized suite of narrow tests that your ML pipeline has to pass as an audit, before being certified as "compliant" and deployed into the web. That way, the regulators get to control the test, and the companies get to control their secrets. It also ensures a level playing field and each component can be changed without an explicit dependency on the other.

Re: Response to Google open letter

#247

Post by thu2111 » Mon, Aug 17, 2020, 7:12 PM UTC

yorwba wrote: post #157

thu2111 wrote: post #21

I'm very curious how this law wouldn't violate various free trade deals Australia has struck. Surely a tax that explicitly names only foreign companies is the opposite of free trade? Sure, China has driven a truck through the whole concept of the WTO over decades and so I doubt anything will happen. Free trade deals are very much toothless, on the assumption that they are supposed to make sense on their own terms so…

The draft does not explicitly name foreign companies, it only refers to generic "digital platforms", so it's not a free trade issue. Free trade agreements tend to only be concerned with equal treatment, e.g. Most Favored Nation status. If, say, a country requires mining companies to implement certain minimum safety standards, that's totally acceptable from a free trade perspective if it applies to all companies, even…

That's interesting. Other posters on this topic have asserted that the directive names Google and Facebook specifically, and nobody else. That's why I say it looks like a clear cut WTO sanctions case. The law would not apply even to Twitter at the moment, let alone Australian services.

BTW are you saying joint venture requirements no longer exist in China? Can non Chinese citizens get IP addresses now?

Re: Response to Google open letter

#248

Post by asdfasgasdgasdg » Mon, Aug 17, 2020, 7:35 PM UTC

burkaman wrote: post #243

asdfasgasdgasdg wrote: post #227

Earlier quoted context omitted.

So can media organizations. They just can't pre-moderate the comments.

Where are you seeing a requirement to pre-moderate?

There is no such requirement. But that is the capability they want that they do not currently have. They already have the capability to delete comments, just like you can delete comments on your own posts.

Re: Response to Google open letter

#249

Post by renaicirc » Mon, Aug 17, 2020, 7:36 PM UTC

BurningFrog wrote: post #219

zentiggr wrote: post #193

Earlier quoted context omitted.

That still carves them out a protected niche in what otherwise was an independent service, that in no other case provides that level of autonomy. If I were in FAANG's place, I'd tell the AU government to get stuffed by turning every response page into a "Tell your government that this bill is ridiculous and we'll be back when it's repealed. So long and thanks for all the fish."

> that in no other case provides that level of autonomy. I, like all FB users, can delete comments to my own FB posts.

Excellent, but why enshrine the ability to do this in law only for sufficiently wealthy media companies? Seems a bit ridiculous, doesn't it?

Re: Response to Google open letter

#250

Post by asdfasgasdgasdg » Mon, Aug 17, 2020, 7:40 PM UTC

clouddrover wrote: post #231

asdfasgasdgasdg wrote: post #226

Earlier quoted context omitted.

If the number of organizations is the concern, the government could fund a larger number of organizations. This doesn't seem fatal to the option I proposed.

Or the government could just make Google pay for what it uses. It's simpler and more direct.

Haha, no it is not. Taxing a specific organization for a specific policy purpose is much more complicated than paying for that policy purpose out of the general fund and paying for it with general taxation.

To see the truth of my claim, consider what's going to happen if Google declines to "use" the Australian news media after this law is passed. I doubt any serious person believes this will be helpful to the news media's bottom line. So you'll have an even more serious problem for news media revenue, and also the search engine most people use will be doing a worse job of helping people get the information they're looking for. All because the government chose a complex policy solution to a simple policy problem. (At least, the problem as stated is simple. I suspect the actual problem the government is trying to solve is a little different than what is publicly claimed.)