#theinternetisdying

11 minutes estimated reading time

Perhaps due to the febrile nature of the times we live in, the tech-savvy community seems to have become aware of various trends, leading them to conclude that #theinternetisdying. This term itself is imprecise. The internet is a series of abstractions, ranging from physical infrastructure to software and functions upon which communication, messaging, video streaming, app data, transactions, and web pages operate.

So technically, #theinternetisdying is actually #thewebisdying.

So why now?

I believe the most intriguing question regarding #theinternetisdying is this: why now? What has occurred is more about internet users awakening from their ‘comfort zone’ and abruptly realising how rapidly things have evolved? I believe that a number of inciting incidents are the cause of this sudden wakefulness:

Google as regressive ‘tax’ rather than marketing channel exposed.
Continued deterioration of Google web search.
Link rot.
Online media businesses look to make money by selling proprietary and user-generated content to LLMs as a revenue stream.
Realisation that a lot of web content is adapted or created using LLMs. The non-English web has been expanded by machine translation of English language content. Secondly, LLMs have been used to create a lot of good enough content in English for publications like Sports Illustrated.

The reality

In reality, what’s happened to the web as netizens knew it has happened over time. To use a vintage web phenomenon as an analogy. It’s like the vintage Joe Cartoon interactive Adobe Flash animation Frog in a Blender from the late 1990s.

A cartoon frog sits in a blender and admonishes the viewer, claiming that they wouldn’t dare to to blend him. The blender has settings from 1 to 10. 1 is mildly agitated water, 10 is instant blended frog. Silicon Valley has slowly upped the power of the blender and netizens realise that things have got weird.

Google tax bias

Google search has been on the shit list of websites as engineering documents from inside Google were leaked. They revealed some aspects about how search actually worked that Google had been denying for decades. A few of the key findings were:

Google search learns from ‘external devices’, so things like Chromecast dongles. Data from the Chrome browser is used in a similar manner, despite Google repeatedly claiming that it wasn’t in the past.
Google values quality, relevant sites, BUT, that value is caveated by restrictions it puts on small quality sites and the benefits it provides to large platforms detailed below.
Popular sites receive higher search levels through the ‘Navboost’ system. This reinforces platforms and established sites. Smaller sites would need to spend proportionately more on Google advertising to match their larger competitors traffic funnel. This designed imbalance is the digital equivalent of John Pierpoint Morgan’s nefarious involvement in railroad transportation, or the Vanderbilt and Rockefeller agreement on oil transport in the 19th century, which drove much of the subsequent anti-trust regulation in the US – this is the Google tax. It’s a regressive tax that is levied on smaller businesses and the ‘free web’.
It is deliberately set up to hold back small sites, many of whom have seen their traffic drop by up to 91 percent. This adversely affects sights that might have deep domain knowledge, specialist retailers and netizens who host personal sites and blogs like this one.

Google has even lied in court and in parliaments to hide these facts. Disclosure of these details have rippled through the search engine marketing industry and strongly discouraged numerous web businesses once the truth came out – for a lot of businesses #theinternetisdying.

For Google the timing couldn’t be worse:

It is seen to have dropped the ball on LLMs, despite having developed most of the key technology powering the likes of OpenAI and Anthrophic.
Google’s local business advertising for the likes of coffee shops or nail clinics have suffered due to the cost-of-living crisis post-COVID and consumer behaviour changes in various countries.
Traditional Google search advertisements for e-commerce are being rapidly eroded by retail media. That is ‘search style’ adverts on the likes of Amazon, Tesco and eBay.

Google is perceived as having set itself up as the ‘start page’ to the open web, while all the time sticking the proverbial knife in all of which adds an inevitability of the feeling of #theinternetisdying.

Decline of Google search

Back in June 2022, The Atlantic complained about the declining utility of Google. This echoed similar themes on discussions that had happened earlier in the year on Hacker News and Reddit. The consensus was that they searched Reddit, StackOverflow, Hacker News or StackExchange as it provided a richer, more relevant base of search results.

I have been using social bookmarking service Pinboard and photo service flickr for search for similar reasons for the past few decades.

Pinboard allows me to search 65,000+ web pages that I have found over that time for something that might be useful. My act of saving the page link in pinboard allowed me to categorise the page saved and implied a certain ‘good enough’ quality to it. I also get to search the public links of other netizens that do a similar thing. Pinboard is insufficiently popular to reward spammers, so the quality quotient is relatively high. Reddit offers a more expansive corpus of links and information, without the same level of quality control.

The reason why Google’s web search has degraded has its roots in Google’s pivot to mobile two decades ago. Google abandoned key areas of interest to web users:

Boolean search terms, which would have been harder to do on early mobile devices.
Blog search because it was non-mobile content.
Google News and RSS, in favour of nascent mobile social platforms that it lost out to.

And the list goes on, I am less sure why it has suddenly surfaced into the public consciousness now?

Link rot

A month or so go my friend Matt in his newsletter recommended a website that allowed you to search Google to find out the oldest mention of a term. So I put my own name in, and nothing came up prior to 2004.

That meant all records of my early agency work had been expunged from the web. Work that included big brands:

BHP-Billiton
Ericsson
MTV Networks
Palm
Sony
Verizon

Alongside startup brands that fizzled out almost as quickly as they had started. Maybe there is still some traces locked somewhere in behind LexisNexis or Haymarket Media paywalled databases.

Author and veteran member of the digerati Cory Doctorow wrote about link rot this year, partly prompted by research from the Pew Research Center. Pew found that 38 percent of content surveyed disappeared over a ten-year period.

But link rot isn’t a new concern. Interest in link rot seems to have peaked 20 years ago.

Link rot is a subset of a wider concern called bit rot, where digital media degrades over time, or can no longer be read due to issues with software file compatibility. Bit rot as an issue was explored in a series of short stories by Canadian author Douglas Coupland in a book of the same name back in 2016.

Web of data to walled gardens

Of all #internetisdying factors, this one surprised me as much as link rot. Closing of Twitter API access was considered to be a defining moment for #theinternetisdying. However it fails to acknowledge that the high point of the web of data was web 2.0 and the comparatively free access to APIs. Facebook with its closed wall by design set the standard for subsequent services like TikTok and Instagram. Like link rot, the awareness timeline feels a decade too late. The closure of Google Reader is an equally big impact back in 2013, stopping mainstream adoption of RSS in its tracks.

LLMs

Journalist Steven Levy has been chronicling Silicon Valley for decades. He wrote a few of my favourite non-fiction books including Insanely Great, Crypto and Hackers. In the summer of 2023, he wrote an article for Wired magazine: What OpenAI wants. This became a cover story for the September 2023 issue of the magazine under the header ‘Dear AI Overlords, Don’t Fuck This Up’. Less than a year later, the consensus from netizens seems to be that they already have.

Several things have happened, here are three of them:

Imitation became mainstreamed. OpenAI used a female voice that was apparently a copy of Scarlett Johansson’s voice due to Sam Altman’s infatuation with the premise of the Spike Jonze film Her. A Ukrainian YouTuber found her likeness being used as an avatar to sell Russian goods to Chinese consumers.
Misinformation had everything from the Pope wearing a designer down jacket and fake black supporters for presidential candidate Donald Trump.
Scott Galloway talked about ‘corporate ozempic’ where AI being good enough to reduce human tasks allowing for corporate headcounts to disappear. The CEO of Klarna freely admitted that they used AI to replace 700 employees in customer service roles. AdVon was used to write articles for SportsIllustrated driving anger and anxiety in readers and journalists.

Automation has eaten blue collar roles for decades, but it has taken the automation of white collar roles to create the panic and sense that #theinternetisdying and AI is killing it.

The Sky Is Falling In

As a child I fell in love the Asterix The Gaul books. In them was the Gaulish village chief Vitalstatistix – who is portrayed as mostly reasonable, well-informed, fearless, (comparatively) even-tempered and unambitious. Vitalstatistix was known for his irrational fear that the sky may fall on his head tomorrow. I was thinking about Vitalstatistix as I wrote this post on #theinternetisdying.

Back in the late 2000s, Dr Ira Wolfe wrote a book that discussed how online behaviour and Google services were creating irreparable damage in the workplace and beyond. His book was merely the latest in a series of panics about societal destruction:

Emmanuel Macron’s ‘decivilisation‘ facilitated through social media connections and content.
Rock music and the record industry created concern about societal degradation in the 1950s.
Alleged over-stimulation due to violent TV and film violence.
Video gaming had some negative effects associated with it in the 2010s according to academic research.
Television faced similar accusations, but there was only evidence of eye strain with long watching sessions.

Back in March 1997, Wired magazine had their own version of the #theinternetisdying, they believed that web browsing (or web surfing as it was termed back then) was about to be killed off by ‘push technology‘. This episode of The Computer Chronicles gives a good overview of push technology at the time.

You may already be using push technology without realising it, such as receiving mobile notifications for breaking news or localised weather alerts.

In conclusion, #theinternetisdying? really?

Previous technological shifts introduced new challenges, but we adapted and progressed. There’s no reason to think the current ‘#theinternetisdying’ phase is any different from those before. Perhaps in 15 years, I’ll be writing about how people feel the ‘metaverse’ has become closed or some other futuristic concern.

Posts on related content can be found here.