Pulling my site from Google over AI training – Tracy Durnell
I’m not down with Google swallowing everything posted on the internet to train their generative AI models.
I’m not down with Google swallowing everything posted on the internet to train their generative AI models.
This would mean a lot more if it happened before the wholesale harvesting of everyone’s work.
But I’m sure Google will put a mighty fine lock on that stable door that the horse bolted from.
I really hope that Betteridge’s Law doesn’t apply to this headline.
While the dream of “personalized” ads has turned out to be mostly a nightmare, adtech has built some of the wealthiest companies in the world based on tracking us. It’s no surprise to me that as Members of the European Parliament contemplate tackling these many harms, Big Tech is throwing millions of Euros behind a “necessary evil” PR defense for its business model.
But tracking is an unnecessary evil.
Even in today’s tracking-obsessed digital ecosystem it’s perfectly possible to target ads successfully without placing people under surveillance. In fact right now, some of the most effective and highly valued online advertising is contextual — based on search terms, other non-tracking based data, and the context of websites rather than intrusive, dangerous surveillance.
Let’s be clear. Advertising is essential for small and medium size businesses, but tracking is not.
Rather than creating advertising that is more relevant, more timely and more likable we are creating advertising that is more annoying, more disliked, and more avoided.
I promise you, the minute tracking is outlawed, Facebook, Google and the rest of the adtech giants will claim that their new targeting mechanisms (whatever they turn out to be) are superior to tracking.
Behavioral ads are only more profitable than context ads if all the costs of surveillance – the emotional burden of being watched; the risk of breach, identity-theft and fraud; the potential for government seizure of surveillance data – is pushed onto internet users. If companies have to bear those costs, behavioral ads are a total failure, because no one in the history of the human race would actually grant consent to all the things that gets done with our data.
Google and the entire tracking industry relies on IAB Europe’s consent system, which has now been found to be illegal.
Now this is a feature request I can get behind!
A user must provide permission to enable geolocation, or notifications, or camera access, so why not also require permission for megabytes of JavaScript that will block the main thread?
Without limits, there is no incentive for a JavaScript developer to keep their codebase small and dependencies minimal. It’s easy to add another framework, and that framework adds another framework, and the next thing you know you’re loading tens of megabytes of data just to display a couple hundred kilobytes of content.
I’m serious about this. It’s is an excellent proposal for WebKit, similar to the never-slow mode proposed by Alex for Chromium.
Remember when I wrote about using push without notifications? Sebastiaan has written up the details of the experiment he conducted at Indie Web Camp Berlin.
Sally takes a long hard look at permissions on the web. It’s a fascinating topic because of all the parties involved—browsers, developers, and users.
In order to do permissions well, I think there are two key areas to think about - what’s actually being requested, and how it’s being requested.
Is a site being intrusive with what they can potentially learn about me (say, wanting my precise location when it’s unnecessary)? Or is it being intrusive in terms of how they interact with me (popping up a lot of notifications and preventing me from quickly completing my intended task)? If one of those angles doesn’t work well, then regardless of whether the other is acceptable to someone, they’re likely to start opting out and harbouring negative feelings.
In this excerpt from his forthcoming book, Cennydd gives an overview of what GDPR will bring to the web. This legislation is like a charter of user’s rights, and things don’t look good for the surveillance kings of online advertising:
The black box will be forced open, and people will find it’s full of snakes.
A collection of interface patterns for granting or denying permissions.
Software is politics, because software is power.
The transcript of a tremendous talk by Richard Pope.
This a great proposal: well-researched and explained, it tackles the tricky subject of balancing security and access to native APIs.
Far too many ideas around installable websites focus on imitating native behaviour in a cargo-cult kind of way, whereas this acknowledges addressability (with URLs) as a killer feature of the web …a beautiful baby that we definitely don’t want to throw out with the bathwater.
I wish to cover the entire Brighton Pavilion in Bakelite for my own amusement.
Let’s be polite. Especially when starting relationships.
Andy sounds a cautionary note: the password anti-pattern may be dying, but OAuth permission-granting shouldn’t be blasé. This is why granular permissions are so important.
A one-stop-shop with links to the authentication settings of various online services. Take the time to do a little Spring cleaning.
Starbucks has opened a branch in Brighton by disregarding planning permission, ignoring planning laws, and by asserting is not in fact a café or coffee shop but a retail outlet.