Erasmus

Sometimes I think we collectively underscope the impact of bots on society A couple of examples that have been living rent free in my head lately

It's not new, but I do think we shrug it off a little more than we should

EU–INC is the single best thing Europe could do to catch-up in the AI race

A simple unified pan-European startup structure, with modern employee ownership and simple access to capital, able to tap into Europe’s full talent pool.

‼️ but it’s at high risk of not seeing the light of day. You can help👇

The Electric Slide The history, 99% decline, and future of the Electric Stack with Sam D'Amico

The future is electric (drones, robots, EVs). #Chips, #batteries & #magnets will be the components on which it's built. Demand will drive #innovation, so the US turning back to fossils is a future-sized foot gun.

Deep dive on the "Electric Stack" (well worth 4hrs)
www.notboring.co/p/the-electr...

Predicting our own demise Creating society-destabilizing incentives for fun and profit!

"The absence of natural hedgers in #predictionmarkets ends up being its Achilles heel. Everyone in the market is either a noise trader or a sharp [...] [They sustain themselves] on the back of new gamblers willing to lose money to the sharps."

reducibleerrors.com/prediction-m...

Just like the äppärät in Gary Shteyngart’s Super Sad True Love Story

Your Review: Dating Men In The Bay Area Finalist #8 in the Review Contest

#SF is probably something like early-XX century #Vienna: a locus for a furious, anxious, and madly creative energy. Much of our future is conjured in the minds of the men living there, and, it turns out, they are not alright... www.astralcodexten.com/p/your-revie...

YouTube video by Sabine Hossenfelder GPT-5: Have We Finally Hit The AI Scaling Wall?

Seems true:
1. This generation of AIs got us something that feels 99% right, but the last 1% (incl. error correction) will take OOMs more effort: de facto, a wall.
2. Pattern matching on language doesn't touch the world-as-is, so LLMs can never truly reason.
www.youtube.com/watch?v=mjB6...

Right now, I don't have any serious alternatives. That's why I hope to flag the issue and start a discussion, so we hopefully can find some ways of mitigating the "Moloch" risks inherent in the #ATProto account portability.

"ATProto's frictionless, cross-platform portability contains within it the threat of creating a negative competitive environment, a classic Moloch, that could drive the online conversation into more extreme trenches."

erasmus.github.io/2025/bluesky...

@noupside.bsky.social @jay.bsky.team thoughts?

In this blog post, I argue #ATProto and its "platform fungibility" may inadvertently create a commercial environment, where we'll "race to the bottom of the brain stem".

➡️ erasmus.github.io/2025/bluesky...

Please tell me all the ways in which I'm wrong.

a set of bubbles summarizing things people were saying all week in Berlin during Local First Conf and User * Agents * Berlin about the intractability of making free software that gives users agency, while also paying the rent.

convened a group of people working on open source software that is intentionally designed to give people more power.

i took notes.

Minimal and shade
“Road of lightness”
#photography #streetphotography
#cinematic #japan #streetphoto
#shadow #dark #tokyo #minimalism #silhouette #street #fineart #fineartphotography

This report from Lund University's Psychological Defence Research Institute lays bare the structure of a large-scale disinformation campaign.

All the tricks of the trade: meme floods, lookalike websites, Telegram influencers, etc.

Wild.

www.psychologicaldefence.lu.se/article/beyo...

Nadar on the art of photography, which somehow feels pertinent to this day:

"Photography is a marvelous discovery, a science that has attracted the greatest intellects, an art that excites the most astute minds -- and one that can be practiced by any imbecile."

There Are People Who Can See and Others Who Cannot Even Look The greatest early photographer's portraits of Hugo, Manet, Dumas, and more

Wonderful portraits by balloonist/photographer #Nadar of #Liszt, #Hugo, #Dumas, #Sand, #Baudelaire, and some samurai.

worldhistory.substack.com/p/there-are-...

Continuous Thought Machines Introducing Continuous Thought Machines: a new kind of neural network model that unfolds and uses neural dynamics as a powerful representation for thought.

A plane is not a bird, but the search for methods to mimic our wetware thought producer in silico is probably worth the while. And stimulating. pub.sakana.ai/ctm/

“Debt: The First 5,000 Years” might be worth a try instead

Copenhagen morning mist

USER * AGENTS * BERLIN · Luma The User & Agents community invites you to come hang out and make things together for TWO WHOLE DAYS following Local-First Conf 2025 in Berlin! We'll be coming…

If you're around Berlin after the conference, check out this fun community-organised event: User Agents Berlin - two days of space for hanging and hacking with friends you met at Local-First Conf! Register here: lu.ma/v457jxp2

this pitch for an open protocol for payments on the web using HTTP 402 is smart in being quiet about the crypto part, the word Coinbase doesn't appear anywhere

www.x402.org

Just out in @politiken.dk: My interview on why Europe must urgently reclaim democratic control over its critical digital infrastructures from tech oligarchs. We need real digital sovereignty, not dependency.
#Eurostack #DigitalSovereignty

GitHub - roostorg/awesome-safety-tools: A curated collection of open source tools for online safety A curated collection of open source tools for online safety - roostorg/awesome-safety-tools

*bird noises*

We heard from many developers and organizations that even finding open source tools that can be used for online safety is a challenge. Our first baby step is putting together this curated list of awesome safety tools.

We welcome contributions and feedback!
github.com/roostorg/awe...

why bluesky "feeds" don't work, a thread

Google Is Winning on Every AI Front Neither OpenAI nor Anthropic have a chance at this point

#Google seems to be getting their shit together. AI was, of course, theirs to lose: they had talent, TensorFlow, data, etc.

The foot guns went off due to anxiety.

$70 Million in 60 Seconds: How Insider Information Helped Someone 28x Their Money On April 9, 2025, someone risked about $2.5 million—and walked away with more than $70 million in under an hour. The trade was simple, but bold: buy a specific kind of option tied to SPY, the exchang...

"If there was interest, the government could know exactly who placed these trades. But the call (options) are coming from inside the house."

Top comment on the Hacker News post: news.ycombinator.com/item?id=4366...

data-and-politics.ghost.io/70-million-i...

Tracing the thoughts of a large language model Anthropic's latest interpretability research: a new microscope to understand Claude's internal mechanisms

For the parrot case: #Anthropic's recent research gave a good look into the "brain" of an LLM.

The LLM does math by text prediction. No mathematical reasoning at all. When you ask it how it arrived at a result, it will make up an answer that mathematically makes sense.

AI 2027 A research-backed AI scenario forecast.

For the superhuman case: Daniel Kokotajlo (
conscientious objector from OpenAI) & co. published a month-by-month scenario for how the techno-political journey towards #ASI could play out

These are strange days. One day, it seems that superhuman AI is just around the corner, the next, that LLM-based AIs are indeed just stochastic parrots.

The Great Social Media Diaspora | NOEMA What happens when sprawling online communities fracture into politically homogenous, self-governing communities?

...but the idea of a massive global town square with top-down centralized control didn't work. It can't.

But: this is a new opportunity for users to shape their experience! What I love about the profile is Jay's emphasis on user agency & anti-authoritarian values.

www.noemamag.com/the-great-de...