Sometimes. logo

Sometimes.

Archives
2026-02-25

Fell For It Again

Going back to post-it notes.

You are viewing a web archive of an occassional newsletter. If you like anything here, please consider subscribing for free at sometimes.email or at the top of this page.

If you’d prefer, there is an RSS feed too, though email is the easiest way to keep track if anything ever changes.

A little over a year ago, awash in unread tabs, I created an account on a new bookmarking app by known internet person ftrain hosted on his new company’s, Aboard, domain. I moved a bunch of links there, and into “stacks”. In late January I went through and organized it a little more with tags. You can guess what happened next.

It was discontinued February 1st.

I didn’t know it was being discontinued – likely due to the fact that my account was an Apple relay account and opted out of stuff – and only discovered days after discontinuation when I got stuck on a failing login page. I’d login and it would refresh. I opened a network console and saw one of the requests was loading a page that was displaying an end of service message. The login page itself said nothing.

They were kind enough to get me a downloadable archive of what I had saved, after an email, but still: I fell for it again. 

I was a very early, and prolific, user of del.icio.us, the original “social bookmarking site” with flat tagging categorization. Since it was social it had a popular page and its look and feel would go on and influence the creation of reddit. del.icio.us sold to Yahoo!, when Yahoo! was still a thing, barely, and incorporated some bad web 2.0 designs, yahoo logins, and faded out of usage until it was eventually bought (ignominiously, as an archive) by social bookmarking site pinboard, which has its own popular page, and is run by another known Internet person: one I stopped following long ago for some fairly mediocre neoliberal opinions about something long forgotten.

So what now? 

Now I go back to unorganized static files on my hard drive. Text files, HTML, JSON files. This is what a personal site should be used for, as awkward as it can be. Laziness and convenience have wrought the worst aspects of tech. Sometimes friction is good.

It got me thinking about the trust we place on our tech providers, whether they are single person code repos or large cloud computing clusters. If my trust in simple public bookmarking services is shot, where does that leave my sense of trust for the services I use that contain more of my information? What of those that contain my life?

I send and post this on Buttondown. If it goes down, I can save my contact list and still send emails. However, I write drafts and organize links in Google Docs, in an account in which I have years and years worth of correspondences and order histories and complaints and job discussions. If it goes down I can go elsewhere. It would be a pain, a huge pain, but it can be done. Friction is good.

Now I think it should be done. The failure of trust is much greater with these giant corporations because the threat is not that they will fall but rather that they will betray you. If someone, or some entity, wanted to find dirt on me, intimidate me, harass me, Google has everything they could ever want. And Google, standing arm and arm with fascists, is complying with an adversarial state. It needs to be treated as an enemy. It can’t be “I can go elsewhere”, it has to be “I must go elsewhere, and I must salt the earth behind me and leave nothing.”

Google Fulfilled ICE Subpoena Demanding Student Journalist’s Bank and Credit Card Numbers | Homeland Security is targeting Americans with this secretive legal weapon | Reddit, Meta, and Google Voluntarily Gave DHS Info of Anti-ICE Users, Report Says | Google is stifling anti-ICE speech in the workplace as 1,200 employees call on the company to cut ties | Three Silicon Valley engineers charged with stealing Google trade secrets and sending data to Iran

And yes, deleted is never really deleted – How Google played a key role in recovering the video from Nancy Guthrie’s cameras | CNN Business – so I do mean salt the earth. 


Further updates in mash surveillance:

Amazon Ring Cashes in on Techno-Authoritarianism and Mass Surveillance | Electronic Frontier Foundation | Exclusive: ICE Masks Up in More Ways Than One | Local police aid ICE by tapping school cameras amid Trump’s immigration crackdown

This App Warns You if Someone Is Wearing Smart Glasses Nearby 


Meanwhile in Slopville:


Speaking of ftrain, as he was shutting down the bookmarking tool I was using he was writing AI apologia (apologAI?) for the propaganda paper: The A.I. Disruption We’ve Been Waiting for Has Arrived. I have not been waiting for this. 

The usual proponents have been consistently obnoxious, though in 2026 I’ve seen an uptick in AI exculpation from those that maybe should know better. Or those that hesitated trying it now. Or if not exculpation, confused defence that glazes over the systemic issues or glazes the black box in a way that favours the systems powering it.

What Is Claude? Anthropic Doesn’t Know, Either | The New Yorker

Rusty Foster, at Today in Tabs, wrote a refute of the above (and others): A.I. Isn't People.

And over at tante.cc, Jürgen Geuter, has a critique of Cory Doctorow (who’s normally been good for stuff like this but increasingly he’s coming from a neoliberal if not libertarian perspective which doesn’t jive with some of the actual ethical concerns): Acting ethically in an imperfect world.

In a way this framing shows more about Cory’s thinking than about that of the people he criticises: Cory is focused on markets and market dynamics and in that world it’s about purchasing. But moral choices only sometimes relate to markets.

I like to think that my opposition to AI comes from the ethical point of view, I try, but it’s also worth noting that it also very materially sucks shit:

Semantic ablation: Why AI writing is boring and dangerous • The Register | AI Slop Will End Canadian Culture as We Know It | The Walrus | How AI Destroys Institutions | I don't care how well your "AI" works - fiona fokus | AI Is still making code worse: A new CMU study confirms | Rob Bowley

Even Anthropic’s own research, where it actively uses a model that is meant to promote and protect American interests (again, this is an adversarial state) has a high rate of blackmail. Seems good. This shit can’t act ethically because it doesn’t have ethics, it’s just statistically cosplaying based on shitty crime novels written by people.


LLM Slop Will Make Us Antisocial - Ruminations on IFComp 2025:

It took me a long time to recognize how vibe coding and the culture of pro-LLM people are repudiating the very idea of community, and I find this to be the most convincing argument against them. Even if people have figured out how to make LLMs not spread misinformation, found a way to decrease energy costs, etc., I still believe these technologies as they are currently implemented should be rejected for inculcating rent-seeking mentalities.

Anyway, some good news: A cup of coffee for depression treatment has better results than microdosing.


Listen to this

Simo CellLiving Emojis & Abdullah Miniawy

Off of Dying Is The Internet by Simo Cell & Abdullah Miniawy, coming March 12th.


o7,

sometimes

Don't miss what's next. Subscribe to Sometimes.:
Share this email:
Share via email Share on Mastodon Share on Bluesky
Powered by Buttondown, the easiest way to start and grow your newsletter.