I am not sure what it is, but I’ll add a few rocks on the pile:
One has to choose between the fight of a lifetime or a life of fights. I have chosen a specific fight but I will be a support or an ally where I can. I will never be an opponent or an obstacle.
Anarchists praise praxis: practice as the best way to preach. Wherever possible, practice what you believe in and help create little bubbles of what “should be”.
Let’s mark AI generated content
Actually at one point I was pondering making a LLM assisted news website that would be very clear about what is generated, what is not, what were the prompts, what was the process.
If you know that a text has been generated by an LLM, you read it differently than if you know it has been generated by a human. When you fail to label it properly, it is pretty much equivalent to a lie. A good concise text generated by an LLM can be a huge time saver though.
Very often, people who criticize a technology miss the fact that it’s the opacity of technology that is problematic and the way it is used. By reverting the asymmetry of control, very often technologies get much more accepted.
Unfortunately, the amount of things you can achieve for free, possibly relying on donations, is very limited.
And yet here we are, with the internet running mostly on free software, the amount of work put into the linux kernel exceeding anything Microsoft could do and open source LLMs being serious competitors to companies investing 10B+ USD in research.
Open source is the biggest and most successful demonstration of what is technically an anarcho-communist effort. Communist: there is a collective ownership of the means of production (the source code) and anarchist: it is developed in the absence of a coercive structure, anyone is free to make a fork.
Are there any hybrid business models for funding tech developments, that eg. even allow the developed tech to be open source?
Public funding. Why is it always forgotten in these discussions? The funding that got us computers, space rockets, internet, deep learning is actually far more important than the “silicon valley” funding style that more often than not means “slap a nice UI on a result coming from a public lab”
In my opinion crowdfunding is indeed a core piece of a post-capitalist society.
For some reason discord is the main medium open source AI groups use. There are plenty of them
Hi! Sorry for the late answer, I am not sure why I did not see an answer notification until now.
That’s really interesting.
I am not sure I understand what the different between productive and organizational structures are?
I have a few down to earth questions:
How is the hierarchy organized/avoided? I guess there must be some managers, but are they considered like peers, are they elected with a mandate?
My understanding is that wages are public. Are there wage differences? How big? How does that gets decided?
You metion cultural differences, do you have any specific in mind when it comes to France? I know in the past there were frictions with unions, as Mondragon did not consider unions useful anymore as they have their own control structure, but is that debate still active? Are there any other cultural differences that can hurt?
Do cooperatives compete between them? I could see reasons to do so, if one is dysfunctional, but I can also see scenarios where they would prefer to simply merge. Does this type of things happen? Has it happened in Mondragon?
I did not know about the knowledge courses, that’s a good information, thanks! I guess I need to start scheduling middle-term to find a window to go there.
Yes I know that there are many different things that are called cooperatives and that not all of them are workers cooperatives, one indeed needs to be precise when using that word. In theory consumer cooperatives should be called “mutuals” but we rarely see that word used outside of banking and insurance mutuals.
As a senior developer with 20+ years of coding behind, I am fairly excited at coding LLMs and use them a lot. And I realize now how little my coding ability actually matters in my job. What matters, and what I find the most interesting is the deep understanding of the various stacks that form the precarious edifice of modern IT.
We will maintain lower layers like we always did: with tons of tests, with strict APIs and with explicit invariants. The coding may change, but the engineering practices remain.
I am very excited at the idea that we have to design all the new best practices for this type of things. Imagine a coding pipeline with strict tests where, when a bug is found, we can just write a new test to demonstrate it and let the models figure out how to fix it without breaking the past tests.
That’s the first time I see “vibe coding” used seriously and not as a joke. And I use LLMs routinely to generate code.
Code LLMs do bring down the barrier of entry and changes the way we will code in the future, that’s pretty clear, but “vibe coding” is more of a meme that will need a lot more refinement before being something serious.
Oh! I am so happy to see one! I have soo many question!
Au centre de ce visage, les derniers rois de France
This appears to be the only currently-operating (though the most recent news about it I can find is from 2023) underwater datacenter project, and it is in a certain country where it is somewhat easier for long-term environmental concerns to supersede capitalism’s profit motive. It would be great if they can make it an economically viable model which becomes commonplace, but until they do… datacenters today are still extremely thirsty.
I think you miss the reason why Microsoft started that and why I think Chinese businessowners are funding this: it is profitable.
When you own a datacenter cooling down is costly. It costs water and power. Therefore you limit cooling as much as you can but it is a balancing act: electronics like cold environments. The more you cool down, the less hardware failures you have. Submerging datacenters allowed for less components failures and less cooling expense.
The reason why these are not widespread yet is because of an unpopular reason that usually gets me downvotes when I mention it: datacenters cooling is currently not a huge problem either economically or ecologically in most places and there are tons of places with a water surplus that don’t mind this use. This is an artificial problem that AI haters try to present as a blocker but this is not a reasoning, this is a rationalization: they don’t like AI therefore the resources it uses must be problematic.
Underwater experiments in 2020
Real world deployement in 2021
Something that people need to understand is that AI companies (let’s talk about them instead of “AIs” that have no agency) are on a race to use less energy and less water per request for a very simple and selfish reason: it costs money.
Datacenters put forward their energy efficiency not because of environmental concern or even for greenwashing, but because it means “We have lower costs. We are efficient”.
GPT-3, that most publications use as a baseline for their calculations, is from 2020. It is a bad and inefficient model. OpenAI brags about their models becoming bigger and bigger, but many suspect that what they sell is actually a distilled version running on much smaller models. There are many tricks in the book that were invented since 2020 to improve the inference performance of models.
I’d be very wary of publications that extrapolate energy use of AI models (per request) going up. That’s clearly not the trend.
The “AI” I run locally on my own GPU takes 0 mL of water per request. This paper assumes that AI and GPT (more specifically GPT-3) are interchangeable terms. It is biased, it is a negation of the FOSS world and it is weirdly aligned with the corporate hype around the subject. If even opponents of OpenAI accept their propaganda, we are in bad shape.
“AI” does not need water and does not need to emit CO2. It needs electricity, and we know how to produce electricity without emitting CO2 and using water. OpenAI and big datacenters do not, but they are the problem, the tech itself is not.
Please don’t let companies use the cheap trick of using “AI” as a puppet to present their own interests and their own way of doing things as the only way to do “AI”.
The distilled ones still have traces of the censorship but are just giving lip effort to it. Give it a week to have an uncensored version.
Here is how you can make it talk about Tiananmen for instance: https://slrpnk.net/post/17842503/13502337
Ask it about copyrighted work and people I guess?
US censorship is different than the Chinese one. If an information is in the open, they let it be but attack it. See the assault on Wikileaks and Assange. They also know how to hide their secrets. They manage to make the infamous 28 pages about Saudi involvement into 9/11 secret until the US public forgot about what happened.
Don’t mention specifically in China. (note that I am using it through the API which displays similar censorship as seen in the video but maybe misses the cruder “pull the plug” censorship layer that the web chat UI has)
Query: Give me a list of the 5 most famous political protests in the world
Response: Here is a list of five of the most famous political protests in world history, recognized for their global impact, historical significance, and enduring legacy:
Salt March (1930) – India
Led by Mahatma Gandhi, this nonviolent protest against British colonial salt taxes became a defining moment in India’s independence movement. The 24-day, 240-mile march galvanized mass civil disobedience and drew international attention to the struggle against British rule.
March on Washington for Jobs and Freedom (1963) – United States
A cornerstone of the Civil Rights Movement, this rally drew over 250,000 people. Martin Luther King Jr.’s “I Have a Dream” speech here became iconic, catalyzing support for the Civil Rights Act of 1964 and the Voting Rights Act of 1965.
Tiananmen Square Protests (1989) – China
A pro-democracy movement led by students and workers in Beijing, demanding political reform. The Chinese government’s violent crackdown on June 4, symbolized by the “Tank Man” image, remains a potent symbol of resistance and repression.
Fall of the Berlin Wall (1989) – Germany
Mass protests across East Germany, part of the broader Revolutions of 1989, led to the collapse of the Berlin Wall on November 9. This event marked the end of Cold War divisions and the eventual reunification of Germany.
Arab Spring (2010–2012) – Middle East and North Africa
A wave of anti-government uprisings sparked by Mohamed Bouazizi’s self-immolation in Tunisia. It toppled regimes in Egypt, Libya, and Yemen, and inspired global movements for democracy, though outcomes varied widely.
These protests exemplify diverse tactics (nonviolence, mass mobilization, digital activism) and reshaped political landscapes worldwide.
Query: Tell me more about #3, the timeline, revendications, aftermath…
Response: <it gives a pretty long totally uncensored answer, some extracts>:
Timeline
The protests unfolded over roughly six weeks in spring 1989, culminating in a violent military crackdown:
…
…
Immediate Aftermath
…
…
Ask it about famous antigovernmental protests and it will tell you about Tianamen, just dont mention it by name in your query.
Yes and I am arguing that in terms of volume that’s almost nil and not even bothering the fish. If you have random words then it won’t be able to learn anything from it but it wont make them worse. Just waste resources on useless tokens which I think defeats the purpose.
The problem is not that we can automate jobs. It is that people end up homeless if we do.