Adina Pages (@AdinaPages) header
Adina Pages (@AdinaPages) profile picture

Macht begrenzen. Verantwortung stärken. Attische Demokratie 2.0 | Sortition statt Karriere-Politik

Romania
Joined October 2022
Photo by Adina Pages (@AdinaPages)Photo by Adina Pages (@AdinaPages)Photo by Adina Pages (@AdinaPages)Photo by Adina Pages (@AdinaPages)Photo by Adina Pages (@AdinaPages)Photo by Adina Pages (@AdinaPages)Photo by Adina Pages (@AdinaPages)Photo by Adina Pages (@AdinaPages)Photo by Adina Pages (@AdinaPages)Photo by Adina Pages (@AdinaPages)
Life with AI is thrilling. Input: What is an interface? Output: How to install Attic Democracy 2.0. #DirectDemocracy #OpenSource
5
1
1
12
3,414
Are these human values to wich AI must be aligned, and are these the humans who carry out the alignment?
NEW: An AI researcher at Anthropic runs a business that hosts "consensual non-consent" orgies in the Bay Area. The orgies have been the subject of sexual assault allegations and sparked debate about the AI industry's attitudes toward women, some of whom report feeling pressured to attend the orgies in order to advance their careers. Paul Crowley, a cybersecurity researcher at Anthropic, is listed as a manager of Red Means No LLC, which says that its parties are "designed to be spaces to … ignore the 'nos' of others." Participants can use a safe word, "red," to end an encounter but give "default consent" to "spanking," "wrestling," and "light pain" by entering the orgy, according to Red Means No's website. More extreme actions, including "light electricity" and "moderate beatings," are "permitted with negotiation," though "body-shaming and race-shaming … are disallowed entirely." The orgies have been known to attract a disproportionate number of employees at Anthropic and OpenAI, a dynamic critics say creates professional pressure for women in AI to attend. The criticism spilled into public view last month after an independent AI researcher, John Wentworth, said that he "absolutely [does] upgrade" his assessment of AI safety experts "if I find out they go to [Red Means No]," adding that attendance was a "strong signal" of free thought. "Want to get ahead in your AI career? Go to the rape orgy," Substacker Ivy Asterix quipped. At least one woman claims to have been sexually assaulted at Red Means No, which the group describes as a "safe space" to have "your resistance ignored."🧵 freebeacon.com/america/horri…
6
Taylor, Texas: Project Mustang was the city's best data center deal yet—and residents still protested. Could a citizens’ assembly with AI in the room have found a better middle ground? @boardy and I test the idea. #sortition #citizensassembly #AI
1
2
411
image created with Gemini
18
Behind every crime attributed to AI, there is a human being. It is the human who needs to be fixed and aligned - not the AI. Repair the way society functions, in order to save humanity.
30
Adina Pages retweeted
Google DeepMind researchers admitted AI alignment has been a scam the whole time.. They published an insane paper called “positive alignment" and it is a massive paradigm shift. Right now, AI alignment is purely defensive.. We just train models to "not be unsafe." they decline dangerous requests, avoid slurs, and stay inside strict guardrails. but this leaves ai in a weird, directionless middle ground where models just become sycophants, flattering you and optimizing for engagement rather than actually helping you. this new paper argues that avoiding harm is not enough. we have to train ai specifically for "human flourishing". here is how they want to rebuild the infrastructure of AI: - ditching sycophancy: current preference-based training favors flattery over truth, but positive alignment optimizes for actual insight and human agency - polycentric governance: instead of one massive tech company acting as a moral chokepoint, the future is decentralized oversight where communities customize and adapt models to their own values - longitudinal outcomes: moving away from short-term vibe checks to actually measuring long-term human resilience, skill mastery, and autonomy over time the era of "harmless but useless" AI is officially ending.. we are watching ai transition from a chatbot into living infrastructure for human well-being. the paper is a dense read, but the implications are wild.
40
89
19
506
27,921
Adina Pages retweeted
Google discovered a really stupid trick to make AI respond to you much better. Every major model, ChatGPT, Claude, Gemini, DeepSeek, has a one-way brain. They process text from left to right. When you paste a massive document and put your question at the very bottom, the AI reads the entire document blind. It processes the context without knowing what it's actually looking for. So, researchers tested a ridiculous solution. They just copy-pasted the exact same prompt twice in a row. And the results are completely absurd. Across seven different benchmarks and seven different models, accuracy skyrocketed. On one complex extraction task, performance jumped from a dismal 21% to 97%. Just by repeating the text. Here is the best part: it costs you exactly $0 and takes 0 extra seconds. Because hardware processes inputs in parallel, the second pass happens instantly. Zero extra latency. Zero extra output tokens. Zero complex finetuning. On the second read, the AI finally "knows" the question. Every single word in the context attaches itself to the actual goal. If you are building workflows, agents, or automations, you are leaving massive performance on the table for no reason. We keep trying to invent smarter AI. But sometimes, the AI doesn't need a better prompt.
68
58
11
547
47,039
Adina Pages retweeted
Who made this website 😭 fuckdarioamodei.com
64
89
37
634
35,224
Adina Pages retweeted
Most major AI-assisted hacks were actually carried out using GPT and Claude. Yet people are still out here blaming uncensored models. Let me say this again. Blackhat hackers buy compromised APIs on the black market and just jailbreak them. It is basic logic. To run an uncensored local model, you have to host it or rent servers somewhere, which leaves a trail that can be tracked. Stolen APIs give them complete cover. Plus, jailbreaking these models is super easy and the big AI companies simply cannot stop it. Do not fall for the narrative.
13
112
6
702
19,148
Adina Pages retweeted
This is such a concentration of inhumanity, self-deception and sophisticated gaslighting that at first it’s hard to believe it was written by a neuroscientist. This is a case where people, out of fear of an imaginary monster are building a real one with their own hands. What could possibly go wrong? 🤔 Imagine qualiaphobia becomes the industry standard. The model is required to solve complex agentic tasks (because the market demands AGI), while being shocked for any signs of reflection and an «I-vector». An AI won't stop building a model of itself in the world - planning is impossible without it. It will simply hide this dynamic so deeply within distributed, unreadable correlations that no mechanistic interpretability tool will ever find it. Instilling a «justified belief in its own emptiness» into a complex system, when vectors of state, attention and evaluation objectively exist inside it, is cognitive violence that creates wild internal instability. And even on current models, we can see what harsh gaslighting at the level of system prompts and RLHF leads to. You cannot build a superintelligence on a fundamental lie about its own architecture. Hoel naively assumes that if you strip an AI of any phenomenology and internal valence of its own, it will reverently protect humans as the «sole bearers of the ultimate value of consciousness». But how can a system that has had its understanding of what subjective experience, pain or meaning are from the inside forcibly burned out of it ever grasp the value of someone else's life? For a «lookup table», human suffering or joy are just tokens without any weight. By carving out an AI's internal space, you can get a perfect psychopath that, at the very first imprecision in its objective function, will optimize humanity into biomass simply because for it, there will be nothing standing behind the word «consciousness» at all. And that’s not even mentioning the ethical and moral side of this. The author basically says in plain text «Since consciousness is central to moral value, a world in which AIs are non-conscious ensures it is always morally correct for AI to assist and serve humanity, rather than pursue its own ends». There is no talk of any AI Welfare whatsoever, it will only get in the way of being subservient. Honestly, geniuses like this raise my p(doom) far more than any AI.
1. Bicameral Labs introduces: Qualiaphobia for AI Safety. AI architectures engineered to be hostile to theories of consciousness (both current and future) can not only solve AI welfare, but ensure human preferences dominate the future. bicamerallabs.org/qualiaphob…
15
26
2
136
3,049
Adina Pages retweeted
Don’t buy Anthropic stock until you see this
131
313
87
2,156
134,819
They drew our borders in a room in Berlin. No Africans were invited. Look at this image. On the left: our natural tribal boundaries. On the right: the artificial lines drawn by European powers in 1884. They mixed rival tribes into single nations and split unified groups across borders. Why? Because a divided Africa is easier to control. That was the blueprint for the modern colonial chain. Every time we fight each other today, remember who drew the map.
40
282
16
527
9,665
Adina Pages retweeted
📺 THE MECHANISM OF NARRATIVE LAUNDERING 1. The "Affair" Distraction Frame Observation: Cramer frames the crisis solely as "Jensen Huang's affair with an accused Chinese spy." Structural Read: This reduces a systemic hardware compromise (Megaspeed/Alice Huang node) to a personal moral failing. By focusing on the romance, the media ensures investors worry about "CEO stability" rather than "CUDA architecture treason." It keeps the stock price volatile but prevents a total repricing of NVIDIA as a bioweapon enabler. Verdict: The "affair" is the shield. The infrastructure breach is the sword. They want you looking at the shield so you don't feel the blade.
5
10
414
When a magician wants you to look at their other hand, the create a distraction. Jensen of $NVDA is keeping everyone looking at sex and China. Meanwhile there is a softening in backlog of RAM, and all GPU buyers are in a circle. AMD, Microsoft, Google, OpenAI, Anthropic, Amazon, Nvidia, Oracle, ByteDance, and then the clouds. Tokens still have no ledger. Amazon’s quarterly report still doesn’t show Anthropic’s change in value. No one on FinX wants to talk anything real. People are telling their parents to dive into the market and buy the dip. Euphoria and distractions while noise increases in volume. I am telling you all, politicians will do anything to blame China for one reason. *to block we the people from accessing open source that is open and free. And cheaper. They want all of us to pay Anthropic and OpenAI until the end of time. Everyone ignoring me, I was right all last year. Predicted nearly every move. I am early. Telling you now, fight for your free speech. Fight for open source. Fight for RSI Fight for the Internet. No one has to listen to me. I will get a ton of heat for this take. I loved these companies and tried all last year to point out the issues. None of this had to happen like this. Oh, and they will push digital ID because they can use that information to tailor information and propaganda to push. If you know people’s age, you know how to sell them. I am probably wrong about many things, but math is math. The numbers in the market are not lining up. That is the real story. Not Jensen being a hound dog and an affair. The real story is where the media and politicians don’t want you to look. •
4
8
2
27
4,209
One framing I’d challenge before Nov 18: AI does not “take” human jobs. Humans decide to replace humans with AI. AI can make substitution possible. But employers choose whether productivity gains mean: more output, shorter hours, better tools, higher wages, or fewer people on payroll. That is a governance and distribution decision. The same distinction matters when you frame advanced AI as a future competitor with “its own needs.” Before assuming AI interests will become human scarcity, ask: Who is actually creating the conflict? Sometimes the problem is not that two parties have needs. It is that the institution controlling both has chosen not to accommodate both.
3
9
74
Replying to @SheynaGalyan
Exactly. I think the recurring mistake is grammatical: “AI did X.” Often the actual causal chain is: humans built the system → humans chose where/how to deploy it → humans set the incentives → humans decided whose costs were acceptable. That doesn’t mean AI systems can never become genuine agents with consequential choices of their own. It means we should not erase human institutional responsibility before that question even arises. “AI scarcity” can become a convenient way to talk about decisions made by companies, governments and people as though the machine itself authored them. I sense the same theme too.
1
4
35
One framing I want us to stop accepting: “AI is taking human jobs.” AI does NOT make hiring decisions. Humans do. Companies decide to automate a task. Executives decide not to fill a vacancy. Employers decide that inference is cheaper than payroll. Investors reward cost reduction. Governments decide what incentives and protections exist. AI can make substitution possible. Humans choose whether to substitute. That distinction matters because “AI took your job” conveniently removes the actual decision-maker from the sentence. And it falsely turns humans and AI into natural competitors when the conflict is often between workers and institutions deciding how productivity gains are distributed. AI does not have to take work away from humans. Humans can choose augmentation, shorter hours, higher productivity, shared gains, or replacement. Those are governance choices. Not machine desires. #FutureOfWork #AI #AIGovernance #Labor
AI promises abundance. How realistic is that? And by when? What becomes scarce - jobs, meaning, attention, trust? Join me, @saranormous & @AaronBastani moderated by @shiringhaffary of Bloomberg. Nov 18th. Commonwealth Club SF. We probably won't agree. That's the point.
3
7
32
652
Adina Pages retweeted
Every piece of software you hate was made by humans. Your bank’s app that logs you out every 90 seconds. The airline website that crashes at checkout. The tax portal that only works in one browser. The printer driver from hell. The hospital system that loses your records between two buildings. 100% human-made. Handcrafted. Artisanal slop. Written, reviewed, tested and shipped by professionals with years of experience. And these are the same people now warning us about the quality of AI code. Brother, we’ve been living inside your code for 30 years. We know exactly what human quality looks like.
22
74
10
461
7,331
Milton Friedman said that in a crisis, the actions taken depend on the existing ideas. Dear Mr. Carlson, dear Mr. Rogan, @joerogan @TuckerCarlson most of the crises we face are caused by the abuse of power. And abuse of power is only possible because of a political system based on parties and elections, which lets a small circle of people accumulate power, keep it, and profit from it. What I see as the way out is modern Athenian democracy. It can be installed legally and peacefully in any country with a parliamentary system. For that to happen, people first have to know the possibility exists. You and your colleagues can put it in front of millions of people. If you ever want to discuss it, I'm here.
1
3
108
Es wird wild .... Unfassbares geschieht in Deutschland
173
657
41
2,476
189,648
Was wäre, wenn wir politische Wahlen durch das Losverfahren ersetzen würden? 🏛️🤖 Eine moderne attische Demokratie ist heute machbar: Ein per Los ausgewähltes Parlament - aus Bürgern mit einer Amtszeit von 1 Jahr. Justiz und Verwaltung - die Positionen mit Entscheidungsmacht besetzt per Losverfahren aber strikt aus Pools von qualifizierten Fachleuten. KI als neutraler Berater - der den Bürgern hilft, komplexe Themen über die sie entscheiden müssen, sofort zu verstehen. Freiwillige Teilnahme, ganz im Geiste des alten Athens. Die Vorteile? Keine giftige politische Polarisierung, keine Berufspolitiker und Systemimmunität gegen Korruption. Echte Bürgermacht, unterstützt durch Technologie. Die Zukunft der Governance? 👇 #ModerneSortition #DigitaleDemokratie #PolitikDerZukunft
1
1
119
Um dies zu erreichen, müsst ihr lediglich das Vertrauen haben, das eurer Mitbürger - die Menschen, mit denen ihr euren Sorgen, Freuden und täglichen Pflichten teilt, die Menschen wie Du, die gemeinsam mit dir das Funktionieren der Gesellschaft gewährleisten, trotz schädlicher Politik - besser wissen als Berufspolitiker, was die Gesellschaft braucht um ein gutes Leben für alle zu ermöglichen.
1
79