25/10/2023
When Twitter changed hands and Mastodon had its 15 minutes of fame, on October 2022, it seemed that we were on the verge of a digital revolution based on ActivityPub, the open protocol that powers Mastodon and other apps in the so called fediverse.
A lot of people got excited. Among them, Matt Mullenweg, CEO of Automattic. In a conversation on Twitter, Matt announced that Tumblr, owned by his company, would get ActivityPub compatibility.
Almost a year later, nobody talks about it anymore and it seems Matt’s promise succumbed to project issues and lack of interest.
In July, someone replied to Matt’s original post and asked if Automattic was still working on integrating ActivityPub with Tumblr.
“It’s a little depressing to think Facebook of all companies might beat you to it”, he said, referring to Mark Zuckerberg’s promise — also not yet fulfilled — to bring ActivityPub to Threads.
“Yeah 😕”, Matt replied. “They are orders of magnitude bigger and better capitalized than us. Not an excuse, just reality.”
The reality, however, may be a little worse than Matt’s answer implied.
A few days after this exchange, Javier Álvarez López, a software engineer from Spain, former Automattic employee, suggested on Mastodon that the project to bring ActivityPub to Tumblr had been shelved.
According to his LinkedIn profile, Javier left the team responsible for Tumblr in May and, in September, left Automattic. Today, he holds a position at Yoast, a popular SEO plugin for WordPress.
Then, Javier echoed a post by Cyle Gage, a Tumblr product person, answering a reader’s question on his blog.
“Matt [Mullenweg] announced it impromptu in a single tweet and never talked about it again”, they wrote. “I just assumed it’s been canned now, or at least, significantly delayed, is there a reason activitypub integration was never properly announced on tumblr?”
Cyle’s complete answer was:
it’s been delayed, but it’s something in our list for @labs and we’re evaluating it.
my concern, re: activitypub, is that while federation is a really great idea, it’s never had a good product. it doesn’t really fulfill a widespread need, it’s very niche. nobody is joining threads or bluesky because it will federate someday. so i don’t think it would help grow tumblr at all. actually i think it would likely end up costing more than it makes, which is a concern.
but stay tuned!
In another post on Mastodon, Javier revealed the behind the scenes of Matt’s untimely announcement:
I was working for tumblr when the Twitter thing happened and we scrambled to put together a project. I was the one supposed to do all the web UI work.
Once my backend counterpart did some initial investigation and scoping, and it became obvious it was a complex task and that it would become a cost center without any chance to generate any cashflow, the project was quietly put on long-term hold.
In 2023, Automattic acquired an ActivityPub plugin then in development for WordPress, hired its developer, Matthias Pfefferle, and together they have already released a 1.0/stable version.
In October, they released the integration with ActivityPub to all WordPress.com blogs, Automattic’s commercial WordPress hosting arm.
Tumblr users, on the other hand, are still waiting for the ability to talk to Mastodon and other fediverse users.
11/10/2023
Last Sunday (8), Elon Musk recommended two “good” war news handles for his 150 million followers on Twitter to get news from the conflict between Israel and Hamas.
The billionaire’s two recommendations are notorious sources of misinformation. In May, they spread the lie that the White House had been bombed, for instance.
Upon noting the gaffe, Musk deleted the post. Before that, it had accumulated +11 million views.
It was only a matter of time — and a dramatic event — for Twitter’s decay to reveal itself in the worst possible way. Over the course of almost a year, Musk’s wrong incentives and disastrous serial decisions turned Twitter into one of the worst places to get trustful information.
Activists and OSINT experts are wasting precious time debunking video game images and old videos, reposted on Twitter to direct narratives and/or earn money with the (bad) revenue sharing program implemented by Musk.
It’s not that digital misinformation has emerged now or is it exclusive to Twitter. It’s just that, there, it’s out of control.
In almost a year, Musk fired ~75% of Twitter employees, dismissed all the thousands of contractors who moderated content, trashed the press, empowered extremist speeches, and created the worst incentives for misinformation to flourish on the platform.
The situation is so serious and peculiar that Thierry Breton, Commissioner of the European Union, published an “urgent letter”, in a harsh tone, pointing out Twitter violations of the Digital Services Act and demanding action from Musk within 24 hours.
Twitter, nowadays, is what all alt/extremist social media — Gab, Truth Social, Parler — have always dreamed of: a space frequented by millions of people, controlled by an extremist himself, and where money and truculence speak louder in a supposed “cultural war” that would be underway.
A lot of good people are still on Twitter, among other (few) reasons, “to get informed”. I’m sorry to say, but the old Twitter no longer exists and what’s left in its place is not good for that.
With information from the Associated Press and Wired.
9/10/2023
Wired’s Kate Knibbs discovered a bizarre phenomenon on YouTube: channels that read obituaries of ordinary people, in large volumes.
In the investigation, Kate found that the channels do this with looking for advertising revenue that Google shares with youtubers.
Despite the bad taste, at first there is nothing illegal in it — at least in the United States. People do weird things to earn easy money.
YouTube’s revenue sharing program has been around for a long time and is a success. It helped to launch small audiovisual empires and establish strong names in the digital influence market.
More than that, it is the realization of something that many activists ask: that the work done by ordinary people on social media benefit all parties, not only companies.
It’s a fair claim, even if it’s not universal. And there may be the mistake of some (all?) Silicon Valley entrepreneurs: believing that money is a universal motivator, the only one that matters.
The fact that YouTube is a de facto monopoly, the only viable place to distribute videos of medium duration in landscape format, should be considered, because it attracts those who do not care or even know about the revenue split.
In other places, this may not be the case.
At the end of September, Reddit announced a similar program, in which users buy credits to pass on to others who post cool things. Along the way, Reddit gets +50% of the money.
Reddit was one of the few places where the motivation to exchange ideas and knowledge was genuine. Not by chance, people started to add “reddit” in Google search queries to get away from posts made for SEO (that is, encouraged by money) and read the disinterested opinions of ordinary people.
Will things continue like this, once there is a financial incentive? Hard to say.
In X, former Twitter, there is now a brazen effort to keep users within the platform, with the promotion of direct content publishing and the implosion of link cards’ layout.
The lure for content creators, journalists, and newspapers to do this? Money. X is giving part of its waning advertising revenue to some paying users with a large audience — without transparent criteria or logic.
Commercial social networks, already dominated by advertising content and by people who have become brands, are lacking in genuine interactions. They lost space for the controversies that go viral and that give money on X, for the “finds” that convert generous commissions from stores eager to reduce customer acquisition costs, to the big business of the internet.
When the place we go to becomes a big mall, we can only play the role of consumers.
People feel that. Hence the search for refuge in messaging apps, where it is still possible to have some intimacy without a flashing ad (or a scam attempt) trying to win or deceive us.
This breather shouldn’t last long. Meta has already felt the blood in the water and has begun to instrumentalize its messaging apps, including WhatsApp.
30/9/2023
The arrival of OpenAI’s ChatGPT in November 2022 was the beginning of a new gold rush.
Companies at the forefront of what is conventionally called “artificial intelligence” (AI), such as Google and Meta, until then did not risk putting unreliable chatbots and super powerful editing features in the hands of anyone.
OpenAI gambled, and ingrained itself into the popular imagination, forcing a change of course in much larger companies.
Google felt the heat and jumped the gun with Bard. The presentation of the chatbot, made in a hurry in last February, was a fiasco.
Since then, Google itself, Meta, Amazon, large companies and new startups have started to pour big money into the creation of AI-based commercial products.
In September, almost a year after ChatGPT replaced web3 and NFTs with AI as the new obsession in Silicon Valley, many of these products began to show up:
- Bard reached Google’s hegemonic search engine and other company services, such as Gmail.
- At Microsoft, Windows 11 and Surface computers became secondary to Copilot, the brand that signals AI features in household products.
- Amazon has updated Alexa to speak more and better, thanks to generative artificial intelligence, and invested USD 4 billion in Anthropic.
- Meta launched a rival of ChatGPT and Bard, the unimaginative Meta AI, and hired celebrities to give faces to AI-based personas that appear as contacts on WhatsApp, Instagram and Facebook.
- Meanwhile, ChatGPT gained a new version and learned to “see, hear, and speak”, as noted by the announcement on the OpenAI blog — which, by the way, could be that of a bad science fiction and/or horror movie title.
With so many companies that control our lives inebriated by AI, eager to push it to the public, it’s a good time to stop and think about why.
In a pragmatic sense, AIs help in brainstorming sessions, with first drafts, organizing and cleaning data, creating images.
Many of these applications raise relevant issues, such as the unethical use of other people’s material for training, the consumption of natural resources by the powerful servers used to train and respond to user requests, and the impacts on the labor market.
It is also debated whether the changes that the AIs promise (or threaten, depending on the point of view) will take place. It can be a revolution, it can end up going nowhere.
For companies, however, it is almost an “all or nothing” gambling. It’s not enough to use a chatbot to work. The future envisioned by them is one in which AI is present all the time in our lives.
For Meta, for instance, we will soon be able to play RPG with an AI Snoop Dogg or talk about last night’s game with a virtual replica of Tom Brady.
I watched the broadcast of Meta and Microsoft events. In both, I had a deep sense of isolation, an ethereal artificiality, more fictional than that of science fiction movies that strive to deliver what Mark Zuckerberg and Satya Nadella managed to show effortlessly.
We are on the brink of going from a stage in which we are isolated from each other, interacting only by screens, to isolate ourselves once and for all, too busy talking to machines that look more like human beings than our equals.
14/8/2023
SEO, the set of optimization techniques for sites to rank better in Google search results, is a kind of religion for marketers and technocrats.
The god Google writes its crooked lines, with ethereal or banal tips and blurred guidelines, which are interpreted by the prophets — the so-called “SEO experts” — and applied on websites of the faithful, in the hope that this will revert into blessings in the form of good positions in the search engine index.
It is an exercise of faith, because no one can point out, with methodological rigor, the cause and effect relationship between SEO and results.
Believers follow Google’s religious precepts and have an only option — to believe. If they work, it is the definitive proof that SEO exists. If not, the problem was that I wrote 490 words instead of 500 and repeated the keyword five times instead of four; I didn’t believe enough.
The parallel became even stronger last Wednesday (9), when Gizmodo obtained an internal memo from Cnet in which the company warned employees that it was deleting thousands of old posts to “improve SEO”. The news was confirmed to the publication by a Cnet marketing director.
Now, god Google has come to demand sacrifices as a condition to pour his kindness on click-baiting sites.
The reasoning, according to the internal memo, is that deleting old content that does not generate traffic “sends a signal to Google that says Cnet is fresh, relevant and worthy of being placed higher than our competitors in search results”.
On social media, Google refuted the strategy. This does not mean much, because Google does not reveal the ranking algorithm of its search engine and, allegedly, does not even understand it completely. Therefore, it is not possible to rule out that, even if Google discourages the practice, under specific conditions it may have a positive outcome.
Cnet, let’s remember, was caught in early 2023 publishing texts written by ChatGPT with gross errors, only to attract unwary on Google willing to click on lucrative ads for financial loans and credit cards.
The prioritization of SEO is the infamous tail that shakes the dog. Aiming at good positions in the Google search engine should not, under any circumstances, overlap with editorial decisions, let alone justify the destruction of the archive of (supposedly) journalistic outlets.
However, this is what happens when the SEO theocracy, led by Google, takes the web by storm. Technological fundamentalism, robots above human beings, people reduced to clicks on ads.
6/8/2023
The word “cloud” was adopted by the tech industry to refer to the large realms of scalable servers.
Thanks to it, any company, startup or individual entrepreneur no longer needs to bear the high upfront infrastructure costs to launch a service on the internet. The cloud allows you to start small (and spending little) and grow continuously, according to demand, fast or slow.
It’s a brilliant model. No wonder, the industry leaders — Amazon Web Services, Google Cloud and Microsoft Azure — have become huge players and are very profitable.
Like all transformative technology, we are fascinated by the good side of the cloud and forget the risks of market concentration, eventual unavailability, security and collateral costs, hidden by the shadow of the optimism that technological progress impregnates in itself. Rare and/or incipient, but still present, these risks in general reveal the physical nature of the cloud, built with many limited natural resources, such as rare metals, silicon, and water.
This Bloomberg report (no paywall) caught my attention showing regions that suffer from historical droughts and, at the same time, house large data centers of companies such as Meta, Microsoft and Amazon.
These data centers, the physical addresses of the “cloud”, consume huge amounts of water. One of Meta in Talavera de la Reina, Spain, still on paper, is expected to spend 665 million liters per year. At peak times, it will be 195 liters per second to cool machines that support the digital cloud.
Not by chance, they have generated dissatisfaction and dislike of people who live in the places where they are installed or intend to settle. Suddenly, they are forced to share the little water available with computers.
It is ironic that the “cloud” of the titans of technology behaves in the opposite way to that of nature: instead of bringing water, it consumes it. Unless it counts as “water” that rain noise from streaming apps, possible only thanks to the digital cloud.