Events & Actions

Tuesday June 30 (5:30 PM – 6:30 PM) Drink-in for Good Jobs & Union Beer! (BuzzWorks, 365 11th St)

Tuesday June 30 (6:30 PM – 7:30 PM) Ecosocialist Bi-Weekly Meeting (zoom and in person at 1916 McAllister St)

Tuesday June 30 (7:00 PM – 10:00 PM) Tenant Organizers Social (1600 17th St)

Wednesday July 1 (5:30 PM – 7:30 PM) Phonebank for the Affordable Housing Guarantee Act (zoom and in person at 1600 17th St)

Thursday July 2 (12:00 PM – 1:00 PM) Housing Guarantee Act Turn Signature Turn-In Rally (San Francisco City Hall, Van Ness steps)

Thursday July 2 (6:00 PM – 9:00 PM) Housing Guarantee Act Signature Gatherer Happy Hour (Zeitgeist, 199 Valencia St)

Thursday July 2 (6:00 PM – 7:00 PM) Education Board Open Meeting (zoom)

Thursday July 2 (7:00 PM – 8:00 PM) Immigrant Justice Regular Meeting (zoom and in person at 1916 McAllister St)

Friday July 3 (5:00 PM – 6:30 PM) Delaney Hall To Adelanto! Solidarity Rally with the Hunger & Labor Strike (Turk Street & Taylor Street)

Monday July 6 (6:00 PM – 7:30 PM) Labor Board Meeting – Office Hours (zoom)

Monday July 6 (6:30 PM – 8:00 PM) Homelessness Working Group Regular Meeting (zoom and in person at 1916 McAllister St)

Wednesday July 8 (6:45 PM – 9:00 PM) DSA SF General Meeting (zoom and in person at Kelly Cullen Community, 220 Golden Gate Ave)

Thursday July 9 (6:30 PM – 7:30 PM) Public Bank Project Meeting (zoom)

Thursday July 9 (7:00 PM – 8:30 PM) ICE: Emergency Planning (1916 McAllister St)

Sunday July 12 (5:00 PM – 6:00 PM) Tenderloin Healing Circle Working Group (zoom)

Monday July 13 (6:00 PM – 8:00 PM) Tenderloin Healing Circle (Kelly Cullen Community, 220 Golden Gate Ave)

Monday July 13 (7:00 PM – 8:00 PM) Labor Board Meeting – Existing Union Support (zoom and in person at 1916 McAllister St)

Check out https://dsasf.org/events for more events and updates.

Drink-In for Good Jobs and Union Beer!

When Anchor Brewing reopens, workers deserve good, union jobs.

Join us to call on the new owners to offer workers their jobs back, and sign a neutrality agreement with the union!

Attendees will receive a UNION BEER poster!

Buzzworks SF, 365 11th st, on Tuesday Jun 30 at 5:30PM. RSVP here!

Tenant Organizers Social

Meet tenants organizers from across the Bay. Come to Thee Parkside, 1600 17th St on Tuesday, June 30 at 7:00 PM for one final time before it closes its doors for good. 

The same forces of rampant speculation and gentrification responsible for destroying local culture spaces (like Thee Parkside) are causing the massive displacement of tenants throughout the Bay Area. Please come and share your experience as tenants organizing in the face of finance capital’s agenda to build their “luxury city”.

The Affordable Housing Guarantee Act is Headed to the Ballot!

BIG NEWS!! In just two short months, we’ve reached our signature goal and the Affordable Housing Guarantee Act is headed to the ballot!

Thank you to everyone who helped us so far to get us to this point. Hundreds of volunteers gathered signatures, donated, fundraised and secured endorsements. If you haven’t donated yet, here’s the link. We’ve got two ways to celebrate coming up on Thursday, July 2nd:

• 12PM: Rally to turn in signatures at City Hall (Van Ness Steps) — RSVP here!
• 6PM: Signature Gatherer Appreciation & Social Housing Happy Hour at Zeitgeist (199 Valencia)

You don’t even need to have gathered a single signature to join us at the happy hour, you just have to want a beautiful future where housing is permanently removed from the speculative market and treated as a human right!

In case you missed it, here’s how we got here: Since 2020, the transfer tax approved by voters has generated more than $500 million for affordable and social housing. San Franciscans made it clear where they wanted that money to go, but City Hall has refused to use it for its intended purpose and is now trying to gut the tax altogether. Not on our watch!  This ballot measure is going to make sure those funds are used the way voters intended: to build social housing and make San Francisco a more affordable place to live.

Stand with Strikers in Immigrant Detention: Solidarity Rally

Friday July 3, 5 PM

Turk & Taylor

Immigrants in detention in Adelanto, California and in Delaney Hall in New Jersey, facilities run by the private prison corporation GEO Group, have had enough and since last month have been on hunger strike, while at Delaney Hall, they have also been on a labor strike, refusing to work for pennies per hour. 

Strikers at Delaney Hall are demanding the immediate release of all detained people, starting with the elderly, pregnant, children and those with serious medical conditions. They are also calling for fair review of immigration cases, and an end to pressure to sign “voluntary” deportation orders. They are also demanding a meeting with New Jersey Governor Mikie Sherrill so she can observe conditions and hear directly from detained people.

Meanwhile, in Adelanto, hunger strikers are protesting inhumane conditions, including mold, substandard food, lack of medical care and retaliation against hunger strikers. 

To win, strikers will need community support from coast to coast. Come join us in solidarity at the site of the historic Compton cafeteria riots, the first well-documented trans and queer uprising against police violence, which is now a carceral facility run by GEO Group.

Come tell GEO Group and ICE: Free them all, and GEO Group out of our city!

EWOC Fundamentals of Workplace Organizing Course

Sign up here!

EWOC holds a regular training course to help you build your union from the ground up alongside workers in your industry. It doesn’t require an organizing background to understand the material, which covers topics including mapping and charting, building an organizing committee, uniting over common concerns, and how to take action. If you’re interested in becoming any level of organizer for EWOC, this course is mandatory.

This course will in person at the DSA office (1916 McAllister). We’ll watch the EWOC lecture together and then go through the discussion activities. If you can’t make all of the sessions, reach out to Caitlin Stanton (SF EWOC local lead coordinator) for accommodations.

SCHEDULE:
Week 1: Developing Leadership
Tuesday, July 14 (7-8:30PM)

Week 2: The Organizing Conversation
Tuesday, July 21 (7-8:30PM)

Week 3: The Arc of the Campaign
Tuesday, July 28 (7-8:30PM)

Week 4: Inoculation and the Boss Campaign
Tuesday, August 4 (7-8:30PM)

The public loves the minimum wage. A by-no-means-unsophisticated cohort of economists agrees that the policy makes sense. Despite this, there is a deep well of contempt toward the policy on the center and the right — often invoking the idea that “basic economics” links wage laws to economic disaster. Arindrajit Dube, a leading academic proponent […]

The post The Wage Standard-Bearer appeared first on Democratic Left.

"From the Free Speech Movement to the Factory Floor: a Collective History of the International Socialists", recounts the ISO's experience from its birth to its dissolution.

The post Remembering the International Socialist Organization appeared first on Democratic Left.

The Buffalo Chapter of the Democratic Socialists of America congratulates Adam Bojak on his victory in the Democratic primary election for the 149 NY Assembly district, and extends immense gratitude to everyone responsible for making his victory possible. 

So, what next? Why join DSA? For a decade, Buffalo DSA has fielded these specific yet important questions. Last Tuesday, with chapter co-founder Adam Bojak’s primary night victory, we received the resounding affirmation that when we fight, we win! 25,000 doors knocked. Thousands of phone calls. Six figures of fundraising from the average person, the worker, those who share the belief, “a better world is possible.”

What does this win mean? We hold the firm belief that electoralism isn’t “the answer,” no matter how exciting or gratifying it is to win, but rather another step along the way. It’s a catalyzing step, a way to draw people in and organize them, but winning an election is never the end goal. We don’t believe in the mentality that you should “just vote.” To be a serious organizer and Buffalonian interested in a more just and equitable society, controlled democratically by the working class, we must do more. An election result is only as good as the power we can build and the priorities we can win.

In this moment, we step onto a much bigger stage for our chapter and broader movement for working people. Ask yourself: what can I do? Then, let’s do it. Nothing is stopping us but ourselves!

The work is an ongoing struggle, and we need you to achieve our vision. If you’ve ever felt ready to take on the establishment, the time to join the largest socialist project in American history is now.

Thank you to the people of the 149th Assembly District. We continue to fight for you, the people of Buffalo, Erie County, and workers of the world. 

Join DSA today!

Monday, July 13 at 7 pm

Join DSA Ventura County’s Mutual Aid Working Group for a planning meeting focused on addressing unmet needs in Ventura County. Bring your big ideas, suggestions for coalition partners, and a desire to stand in solidarity with others. We are cookin’ up some ideas, and will post an agenda on our slack.

Sponsored by

To be blunt, bullshit is surprisingly important to the late capitalist economy. Many of the products and services now being sold are themselves bullshit, and can only be marketed using a massive dose of additional bullshit. Bullshit is needed to attract paying customers and, especially, to bring in investors and jack up share prices, which are themselves bullshit, often by orders of magnitude over any truly productive products  (see most of Nazi oligarch Elon Musk’s business ventures).

Bullshit is all around us, and no contemporary issue is soaked in it more than the “generative AI” debate. Even the framing (calling chatbots “artificial intelligence” because they are now marketed as replacements for human ingenuity) is bullshit. So as we wade into this bullshit debate, we must be vigilant in recognizing and guarding against that  bullshit.

At our June Convention, the Chicago Democratic Socialists of America passed a total ban on the use of large language models (LLMs) and similar technologies in our chapter work. Two days later, our Socialist Night School covered an aspect of the AI debate, in the context of eco-socialist theory and practice. One article we read, “Democratic Governance of AI Is the Real Solution” published this April in Jacobin by Holly Buck, made an argument that also came up in our chapter’s debate: We should fight to shape and harness these technologies, not stop them.

Chatbots Don’t Dream of Electric Sheep

This argument hinges on a claim familiar from tech industry marketing pitches: AI is incredibly useful, its maturation as a product is inevitable and world-changing, and everyone needs to get on the bandwagon. “How many of these people will block data centers but end up paying for a subscription to a frontier model once it is clear how useful it is to navigate daily work and life?”, asks Buck. 

That claim is, well, bullshit. But it’s not merely bullshit marketing. Bullshit is also inherent to the technology itself: LLMs cannot help but make shit up. Their output is inevitably bullshit, and, crucially, it always will be.

By now, everyone has heard about ‘AI hallucinations’, but this term is in itself another bullshit obfuscation. An LLM cannot ‘hallucinate’, because it does not think, it does not sense, and it does not imagine. It is designed to give people the impression that it is thinking, but that impression is pixel-deep. It truly is just glorified AutoCorrect: based on statistical analysis of massive amounts of (ill-gotten) text, it generates phrases, sentences, and paragraphs statistically likely to pass for what you asked for. 

At no point in this process does the model reflect on, consider, or imagine anything. Above all, it does not hear, see, or otherwise sense a world outside of the texts it is based on. It doesn’t get the world wrong — it has no world to get right. As one literacy specialist put it, LLMs have a “baked-in indifference to the truth.” They have no grasp on reality. The closest they get is testing outputs against documents they have already ingested. They are like a ghost trapped in a massive library with no real memory of the world – only what it has read about it. It makes things up without noticing.

The Margin of Bullshit

What an LLM does is statistically approximate outputs humans will accept. A margin of error is par for the course. The output is designed to give a good impression, but any piece of it could be subtly wrong or completely unhinged.

This is well-known in the AI industry, and accepted as an inevitability. Developers are trying to shrink the margin, but don’t expect to eliminate it. Google has released a set of benchmarks to test the factual output of different LLM models, and the best score any model has gotten on any of them is 85%. On the combined average score across the set, the best score is 71%. Most models score 40-60%.

In other words, you cannot trust anything these chatbots tell you. If you know the subject matter well, you will be able to spot the bullshit right away and correct it; but if it says anything you didn’t already know or cannot independently confirm, you have no idea if it is just bullshitting you again. It can misinform, but it cannot inform.

Good for Them

I have been trying to make the case that LLMs are not nearly as useful as they are made out to be. So why is the ruling class wagering the entire economy on them?

The basis of working-class power is that the bosses need us more than we need them. As the song goes, “without our brain and muscle, not a single wheel can turn.” The history of industrial capitalism is a history of the bosses trying to break worker power by replacing them with machines, which will not strike, talk back, or need rest. 

Now, the bosses see an opportunity to downsize and deskill in ways they previously could not, and they are excited. Some of them are so high on their own supply, they think they no longer need us at all; no wonder they now embrace end-times fascism. 

Even without going that far, any form of mechanization or automation represents a transfer of power. Steam-powered looms replaced skilled weavers with low-paid machine operators, and now Claude code replaces trained programmers with untrained vibe coders or ‘prompt engineers’. Each time, workers’ knowledge and skill gets built into machines which the boss controls completely, while a layer of specialized workers who once had real bargaining power are cast aside. The workers that replace them require less training and can easily be replaced, leaving them in no position to fight the boss. 

In most cases, replacing human skill with a machine’s statistical mimicry will lead to worse products and services (think chatbot therapists or AI slop books), and to worse working conditions for those providing them. But it also becomes much cheaper to produce those shitty products and services for the masses while the 1% still get their bespoke, artisanal, human-powered goods and services. 

Robocop’s Deadly Oopsies

This shittiness is par for the course. But sometimes, it’s a feature, not a bug, as when the customer is the ruling class and the product is the means of repression. Consider the use of AI in policing: When the cops use it to ‘identify’ a suspect, they not only have to work less, they also become less accountable. When they inevitably arrest, abuse, or murder the wrong person well, darn, that new computer program must have been ‘hallucinating’ again. Whoops! The commissioner will have a word with the vendor. The victims can try and sue the corporation that made the AI, but the poor cops, they were just doing their level best using the latest tools.

This logic of murderous impunity can be seen in action through Israel’s genocidal campaign in Gaza, which employs automated processes to choose targets – taking as a given that there will be a host of collateral casualties around them, and that some targets themselves are within the margin of bullshit. Thinking is streamlined out of the process, and everyone in the ‘kill chain’ can claim it was precision-targeted — they were just ‘neutralizing’ AI-identified targets, after all.

In this way, having a margin of bullshit is a bonus, not a failure. They may not admit it – they may not even understand it – but the inherent inaccuracy of the system serves to protect oppressors. While they rest easy, the rest of us live in terror of split-second algorithmic decisions with a deadly margin of error. Who will they hit next?

Bad for Us

This is not to say that using a chatbot fuels genocide. These are different applications of related technologies. The point is, in our work of agitating, educating, organizing, and mobilizing working people to build socialism, our purposes and priorities are totally different. 

The boss wants interchangeability and impunity; we want empowerment and self-conscious action. The boss cares about outputs and outcomes; we care about process. We take action not only because of the immediate result; we take action to build our capacity to take action, to develop skills, know-how, and confidence.

We have no use for the shortcut of having Grok write copy for our literature; why pass on an opportunity to hone our thinking and communications to a machine? If we ask ChatGPT to design and illustrate a poster, we additionally eliminate art — one of the most pleasurable, expressive activities people are still allowed to enjoy. Instead of developing our creative powers to remake the world, we develop a dependency on machines we have to rent from the boss while we’re trying to fight him. This is not even to mention that almost nobody will bother to look twice at yet another piece of samey-looking slop.

My own organizing work is focused on political education. I always wish I could read more, ingest more theory, and translate more history and insight into our present moment. A lot of people have started asking chatbots to do it for them, but this defeats the purpose. 

Reading ‘key takeaways’ is not the same as doing the reading yourself. It not only jettisons most of the substance, it’s also a completely different mental process. Even if you luck out and it’s bullshit-free, it won’t leave you with the depth of understanding you would get from personally grappling with the actual text. And besides, what use is a summary if any or all of it could be “hallucinated” nonsense?

As with so much of our work, the work itself is valuable and does not need eliminating. We cannot bullshit our way to socialism.

Notable Exceptions

All that said, I almost voted against the AI ban. It is hard to cut through the bullshit and draw a clear line around what kind of automation we should and shouldn’t use. The blanket ban rules out specific use cases we may want to reconsider. 

As a former translator who took just one transcription job and never accepted one again, I would suggest allowing machine transcription. Transcription is (usually) drudgery, and drudgery should be eliminated by automation.

By contrast, although translation is beautiful, creative, and often fulfilling work, in organizing I have found that it is often so expensive or time-consuming that it may not happen at all without machine assistance. This technology is closely related to chatbots, but matured years earlier and has a much smaller margin of bullshit. 

In both cases, the services used should be vetted carefully for privacy protection, and the outputs should still be checked carefully by skilled humans. In the case of translation, training up skilled socialist translators may still be worth the time and effort.

I am sure there are other ‘legitimate’ uses for AI out there, outside of my personal experience. I think safe ways to use them should be identified, and an argument made for them. In my opinion, though, starting with a blanket ban and figuring out the exceptions on a case-by-case basis is much better than letting the bullshit generators further rot our brains while we try to catch up.

My thoughts here are indebted to conversations with Hagen Blix, whose co-authored “Why We Fear AI” is an invaluable contribution to a left analysis of ‘generative AI’.

The post “AI”: Beware the Margin of Bullshit appeared first on Midwest Socialist.

At the June Chicago Democratic Socialists of America chapter convention, our membership passed the resolution to “Ban Chapter AI Use and Resist AI Proliferation.” I believe this is a good step towards combating the artificial intelligence oligarchy and its destructive, unpopular data centers, as well as ensuring our members develop their skills by doing. 

The most dangerous thing working-class people have is our brains. Knowing this, and that AI use has been documented to reduce brain function (“Your Brain on ChatGPT: Accumulation

of Cognitive Debt when Using an AI Assistant for Essay Writing Task”), I am concerned for the future of our class struggle if AI use proliferates further. We already have grave literacy issues in this country; in 2019, one in five adults in the United States did not have English literacy skills sufficient to complete tasks that require comparing and contrasting information, paraphrasing, or making low-level inferences, according to the National Center for Education Statistics. That in itself is alarming, and we as socialists should fight to promote literacy at a time when misinformation and fascist propaganda is seeping into all of our media. Discouraging members from using AI is one way to do this, but we also need to begin to explore heavily regulating or outright banning AI use in news media, social media, education, the medical field, the legal industry, government, police surveillance, and benefits assessments, especially considering the problems of false information in these fields. 

Most people learn by doing in some capacity. The act of trying and failing is important for human development, and AI creates a ‘shortcut’ to learning that turns people into mere production machines. I see no reason why the chapter should use AI in any chapter activities. We have smart comrades with a wide variety of skills, and what we don’t have we can learn, recruit help with, or pay a worker to do, rather than contribute to a demand for AI that is degrading our environment at an alarming rate, much faster than other capitalist industries. I personally started participating in the chapter through helping the Lift the Ban on Rent Control campaign research donations from the real estate industry to various local elected officials and politicians, voting histories, and local housing decisions. That self-conducted research helped me develop my politics, notice new things, and recognize patterns. If I had used AI to do that research, I would have received a final ‘product’ much faster, but I would have also missed the political development and personal connections that came with asking other comrades I did not know for help, reading peripheral material I did not originally think was relevant, and seeing over and over the sheer scope of power the real estate industry has over politics in Illinois. 

It is true that AI is hard to escape using any existing technology, because AI companies are infesting every aspect of society to squeeze whatever profit they can out of their investment. We do not have to accept that. We can take a page out of Lina Khan’s book and pass regulations requiring companies to have clear opt in/opt out interfaces so consumers can choose whether or not to use it. We can pass moratoriums on new data centers and encourage research into new ways to run existing ones to reduce their harm to the environment and the lives of their neighbors. We can pass regulations around intellectual property use and data-scraping by AI to avoid dead internet and protect original thoughts and cultural production. We can tax AI companies at high rates, or push for public ownership so that the people make the decisions and reap the benefits around the technology. And most importantly, when we pass these laws and regulations, we must staff regulatory agencies and empower them to go after companies and their shareholders to enforce them. 

The post Against AI appeared first on Midwest Socialist.

Sunday – July 12 at 3pm PDT

Location:  The Open Book – The Oaks Mall

556 W Hillcrest Dr, Thousand Oaks, CA 91360

Join the DSA Ventura County book club for discussions about selected readings. If you love history, the written word, or the world around you, then you belong here!

This will be a friendly, inclusive event open to all, regardless of prior experiences or familiarity with the topic.

Wednesday, July 8 at 6 pm