The "Quora Wedge"

The promise of the internet plus the World Wide Web was an open, free network of hyperlinked pages, filled with all of the worlds knowledge. For years, the terms "internet" and "world wide web" were almost interchangeable/synonymous.

30 years on... It has issues.

As a 'telecoms' platform (for transmitting 'raw' packets of data), the internet is 'free' and 'open'. Anyone can connnect to it and use it, without permission from any sort of centralised body, or having to pay any sort of royalty fee for using someone's intellectual property.

As a 'media' platform (for transmitting content for/by people), the world wide web sits on top of the internet, and is also 'free' and 'open'. Anyone can write a web page, run a server to host it (using free/open-source software), and 'publish' it online. If you want a domain, so your page can be found at a friendly and memorable address like "somerandomnerd.net", rather than a less friendly IP address like 198.185.159.145 (remember when we were able to remember long numbers?), then that does cost something - but typically, you're looking at something like £10-20 for a year. And actually having a connection to the network will generally mean paying for an always-on, reliable service.

However, although commercial organisations in the "media" and "communications" business - especially those producing 'content' of their own - benefit from 'free' and 'open' platforms, they often don't particularly want their own content to be 'free' or 'open'; their content is their intellectual property, and the attention of their audience is their business model.

The 'traditional' example; Google benefit from being able to freely access websites and read their content to help them build a better search engine; they read websites and build an index, which powers their search engine product. Their search engine helps website owners attract an audience. But Google has no interest in making their own index 'free' and 'open' - why would they want to help a competitor build on their work to make a better search engine?

A more '2024' example; OpenAI benefit from being able to freely access websites and read their content to help them build a better large language model; they read websites and build weights for a model, which powers their ChatGPT product. But OpenAI has no interest in making their own weights 'free' and 'open'.

One of them is seen as aligned with many (but not all) website owners' business model. The other is currently seen by many (but not all) as an existential threat to their business model.

Since the web was established, everyone got smartphones, and smartphones have apps. Everyone with a smartphone seemed to want websites to be apps, so pretty much everyone with a website went ahead and made an app.

If you've got an app, then you're probably incentivised to want to encourage people to use the app instead of the website. First, you need to justify the expense. Second - its (generally) a lot easier to keep and track user data through an app than it is through a website.

If your business model relies on your website but online advertising isn't delivering what your business model requires, then you need to reconsider how your website is designed. That might mean limiting the availability of 'free' content.

Which is what leads to the 'Quora wedge' - a term coined (I think) by Dan Hon;

The Quora wedge is... well, it's not a wedge. It's like a splinter. It's what happens when you search for something on Google and, if Google hasn't rewritten your query to be something slightly more profitable for them, the top result is some Quora question/answer. You go click/tap/press your nose on the glass to read it and surprise! You can't read it, because it's behind a Quora registration wall. All shitty all around for everyone, definitely not a great seamless user experience.
Substack (the second of the three hosts for this newsletter) is doing a Quora, and I'm relatively sure that their Quora wedge is going to get a lot worse. When you hit a link to a Substack newsletter, the first thing that pops up is the modal asking you to subscribe, along with a little "nah, just let me in anyway".
There is no way this isn't going to get worse

So - its not a wedge, and its not just Quora. Substack seem to be on the path - but to me, the worst offender is Medium. Like Quora and Substack, Medium often insists that you log in to access the article that you've clicked on. Unlike Quora and Substack though, its something of a coin toss with Medium - sometimes you can read the page once you've logged in as a registered 'member', but sometimes it isnt' just for 'members' (in the sense of "people who have signed up and joined Medium"), but only for paying members.

One consequence is that people like me get into the habit of opening Google search links in separate tabs, so I can go back to the search results and try a different one if the first doesn't deliver on what I expected. Which then leads to opening the first three links in new tabs - opening a bunch of web pages that I don't look at, meaning getting served adverts that I never see. (Bad for publishers, bad for advertisers.) The experience of web searches gets worse - and the potential for generative AI to be a better 'front page of the internet' increases. (And thats even if AI technology doesn't get better - which is something that is pretty much inevitable.) Which, similarly, is bad news for publishers.

The bit about "if Google hasn't rewritten your query to be something slightly more profitable for them" is a reference to a recent news story about how Google adjusts users' search terms in a way that can mean that while the original search term would not have been something advertisers were bidding on, the adjusted search term would be something that qualified for the advertising auctions - and therefore make Google some more money. Advertisers paid to advertise against the words people typed into Google, but Google would change the words people typed in to the words advertisers had paid to advertise against.

That link is to the Wired article, which now reads;

EDITOR’S NOTE 10/6/2023: After careful review of the op-ed, "How Google Alters Search Queries to Get at Your Wallet," and relevant material provided to us following its publication, WIRED editorial leadership has determined that the story does not meet our editorial standards. It has been removed.

Its frustrating that the article has been removed entirely, meaning that you can read the conversation or allusions to the article, but not read the article itself.

The thing I see more prominently on the page is this message;

This is your last free article. See the future here first with 1 year of unlimited access.

Again - there is content on the World Wide Web, but you might not be able to see it. Maybe you need to subscribe to a 'platform' like Quora/Medium. Or maybe you need to subscribe to a 'publication' like Wired.

These are slightly different situations; Quora, Medium and Substack started out as 'platforms' - hosting services that allow you to write content and have it hosted for you - that have then shifted their business model to 'publishers' - services that pay writers for creating content.


I'm trying to figure out what the 'best' way to write for the open web is right now - should a page be a reflection of a particular point in time, maintained for the sake of an archival reference? How much work should maintaining a page involve - editing to remove things like typographical mistakes or broken links seems like a good thing to do, but what if you're editing to reflect a change in your own opinion or point of view? What does that mean for someone who posts a link on social media to an article that they agree with, if that article gets changed to something they don't? (Or, if you don't think you 'owe' some random nerd on Twitter anything - what if something you've written becomes a citation link on Wikipedia?)

And if 'publisher' content gets shut out of future LLM training data, what sort of responsibility do we have to each other if the collective work of our online content is to form the training data for a future shared electronic intelligence?