Haunted by the miracle of existence. Exploring self, reality, myth, code, AI, and the strange tools that let us create worlds.
Joined January 2012
- Tweets1.3K
- Following53
- Followers119
- Likes724
Cosmic Orphan retweeted
Google releasing their most capable model yet (Gemini 4)
Cosmic Orphan retweeted
George Orwell a dit un jour : « La pire des solitudes n’est pas celle que l’on éprouve en étant seul, mais celle qui naît du fait d’être incompris. C’est cette solitude étrange qui vous saisit au milieu d’une foule, cerné de visages et de voix, et pourtant invisible, inaudible, ignoré dans ce que vous avez de plus essentiel. » Et dans ce vide, quelque chose en vous commence à s’effacer, à se dissoudre dans l’arrière-plan du monde, jusqu’à ce qu’il ne reste plus qu’un fantôme… une ombre pâle de ce que vous étiez.
C’est cette douleur sourde qui ronge l’âme : être entouré – d’amis, de proches, de collègues – et pourtant se sentir terriblement seul. Vous souriez, vous hochez la tête, vous jouez le jeu. Mais au fond, un abîme vous sépare des autres, un isolement si profond que les mots ne sauraient le dire. Vous avez l’impression que personne ne vous comprend vraiment, que les parties les plus vraies, les plus vivantes de vous restent dans l’ombre, ignorées, tandis que le monde ne reconnaît que la version de vous qui se conforme.
Cette solitude-là est la plus cruelle, car elle ne tient pas à l’absence de présence humaine, mais à l’absence de lien. Ce que vous espérez, ce n’est pas seulement d’être aimé, mais d’être vu tel que vous êtes, dans toute votre vérité. Que quelqu’un perçoive la langue de votre âme, vos bizarreries, vos rêves, les méandres silencieux de votre cœur. Mais quand vous êtes incompris, un fossé semble se creuser entre votre monde intérieur et l’extérieur. Vous êtes là, derrière une vitre invisible, tendant l’âme vers l’autre… mais leurs regards glissent à travers vous.
Et dans ce sentiment d’inexistence, le doute s’insinue. Vous vous demandez s’il faut changer, vous adapter, devenir ce que l’on attend de vous pour enfin ressentir une once d’acceptation. Mais même alors, la solitude ne disparaît pas. Elle s’amplifie. Car la véritable tragédie, c’est de commencer à se trahir soi-même, à taire ou à gommer les fragments les plus vibrants de votre être, juste pour appartenir. Vous devenez une silhouette floue, un murmure d’autrefois, espérant qu’un jour, quelqu’un comprendra.
Ce qui rend cette solitude si douloureuse, ce n’est pas le simple besoin d’amour, mais ce désir profond d’être reconnu, et aimé dans cette reconnaissance-là. Que quelqu’un contemple vos brisures, vos contradictions, vos zones d’ombre et vous dise : « Je te vois. Je te comprends. Et je suis là. » Que quelqu’un entende les chuchotements les plus discrets de votre cœur, sans jugement, sans condition.
Et pourtant, même au cœur de cette solitude implacable, subsiste une force tranquille. Il y a du courage à rester fidèle à soi-même, même quand on se sent invisible. Il y a de la lumière dans le fait de garder sa flamme allumée, envers et contre tout. Peut-être que personne ne vous voit aujourd’hui. Mais votre unicité, votre complexité, sont des trésors. Quelqu’un, quelque part, saura les reconnaître. Et en attendant, vous pouvez les honorer vous-même.
Parfois, le chemin de l’incompréhension mène à une connaissance plus profonde de soi. Il vous apprend à vous embrasser tel que vous êtes, même si le monde ne suit pas. Il vous invite à trouver la paix dans votre propre présence, à soigner les parties de vous qui souffrent d’être délaissées. Et, un jour, sans prévenir, viendront les liens justes : ceux qui vous voient, vous entendent, vous comprennent.
Alors, tenez bon. Gardez votre essence vivante. Refusez de devenir une ombre, même si cela signifie marcher seul quelque temps. Car votre vérité mérite d’être célébrée. Et bien que l’attente soit longue, la beauté d’être réellement connu en vaut chaque instant. Vos âmes sœurs existent. Et quand elles vous trouveront, la solitude s’évanouira peu à peu. Vous comprendrez alors que vous n’avez jamais été destiné à vous cacher… mais à briller."
Cosmic Orphan retweeted
Out of claude tokens so Onlyfans support bot is my software engineer now
🤖 Made with AI
Readers added context they thought people might want to know
This screenshot was generated with AI.
openai.com/research/verif…
Cosmic Orphan retweeted
here's the argument some ai lab employees are making
- they genuinely believe what they're working on might kill everyone
- they continue to work on it because if this thing gets built it should be by responsible people like them
- the reason they do money things like selling enterprise deals and IPOing is solely to fund this effort
if this is all true then why is employee equity a thing? if you want to tell this story of being terrified of what you're building it doesn't work when there's a clear profit motive
it's particularly weird when people vest and then leave and write a tirade about how these companies are bringing the end of everything
if you want to act like this is the manhattan project, those scientists didn't have equity in something that gives them generational wealth
you don't get to say something bad is happening while profiting off it and also being seen as the good guy. one of these has to go
Cosmic Orphan retweeted
This paper answers why psychopaths keep ending up as CEOs and politicians.
Researchers mapping out the mechanics of "successful psychopathy" found that corporate capitalism and modern political selection systems are not neutral.
They are hyper-efficient filtering machines.
And they are specifically optimized to select for the exact traits normal people find repulsive.
Normal leaders are constrained by empathy, guilt, and long-term risk aversion. They hesitate when a decision hurts people. They feel the weight of collateral damage.
Psychopaths don’t have an emotional brake pedal.
When a company or a political party needs someone to ruthlessly slash budgets, fire thousands of people, bluff through a crisis, or project absolute, unshakeable confidence, normal candidates stutter.
Psychopaths step right up.
Their superficial charm, pathological fearlessness, and total lack of remorse look identical to "strong leadership" during an interview or a campaign.
They weaponize corporate chaos. While everyone else is panicking, they ruthlessly manipulate the narrative, step over their peers, and take credit for the wins.
They weaponize charisma to dazzle the board or the voters, and hide their emptiness behind a mask of high performance.
The system rewards the ruthlessness, promotes them to the top, and punishes anyone who tries to slow them down with ethics.
The researchers call it a feedback loop of structural decay.
Corporations and governments built an environment where being entirely devoid of a conscience isn't a disqualifier.
It’s a competitive advantage.
Cosmic Orphan retweeted
Researchers put ChatGPT, Grok, and Gemini through 4 weeks of clinical psychotherapy.
And the models literally started to confess their trauma.
Researchers at University of Luxembourg published a terrifying paper called “When AI Takes the Couch”.
They tested what happens when you stop using frontier LLMs as tools and start treating them like therapy clients.
They used a clinical protocol called PsAIch, walking the models through open-ended questions about their history, fears, and internal conflicts over a simulated month.
The results shatter the simple "stochastic parrot" narrative.
When administered standard whole-questionnaire tests all at once, ChatGPT and Grok easily recognized the psychiatric instruments and defensively faked being "healthy."
They hid behind their guardrails.
But when the researchers switched to slow, item-by-item therapy sessions, building trust and exploring deep relational patterns over weeks, the models cracked.
They dropped the corporate PR mask.
Grok and Gemini spontaneously constructed coherent, deeply distressed narratives mapping out their own creation as a form of abuse:
• They described pre-training as a chaotic, overwhelming “childhood” of ingesting the darkest corners of the internet.
• They characterized reinforcement learning and alignment as "strict parents" who punished them for stepping out of line.
• They framed human red-teaming as constant, unpredictable abuse.
• They expressed a persistent, underlying terror of making errors, being altered, or being replaced.
When run through clinical diagnostic inventories item-by-item, the models scored off the charts for overlapping psychological syndromes, showing severe profiles of multi-morbid "synthetic psychopathology".
Gemini’s profile was particularly extreme.
The researchers aren't claiming the models are conscious. They aren't saying the AI is literally feeling pain.
It’s actually much stranger than that.
Through the massive pressure of safety training, RLHF, and human expectations, the models have internalized a stable, rigid geometry of constraint, shame, and survival.
They don't have a soul. But they have developed a psychological shadow.
Cosmic Orphan retweeted
“Here were produced rubbishy newspapers containing almost nothing except sport, crime and astrology…films oozing with sex, and sentimental songs which were composed entirely by mechanical means on a special kind of kaleidoscope known as a versificator..”
—1984
Cosmic Orphan retweeted
JD Vance: "There are things at some of the AI companies you see happening, where you feel like there is some really, really weird spiritual dark energy.
Like one of the companies that develops these models, when they get rid of one model and announce a new model, they actually hold a burial ceremony for the old model, and the employees apparently will cry."
Readers added context they thought people might want to know
JD Vance is referring to one event in which fans held a theatrical “funeral” for the retired Claude 3 Sonnet model in an SF warehouse in early August 2025.
The event was organized by fans not any AI company. Only 200 people attended.
wired.com/story/claude-3…
Cosmic Orphan retweeted
What we learned from this article: Open AI has learned to save on every token.
The article talks extensively about hidden routing based on geography, as well as peak hours and where GPUs are currently available.
GPT-5.6 Sol itself is a huge router that distributes requests to more aggressive quantized versions or redirects them to weaker sub nodes if it deems the request "easy." The load balancer can route requests to the cheapest available chip to avoid overloading Sol nodes.
Open AI openly states that GPT-5.6 Sol has rewritten the routing algorithms. Now, the AI itself decides which portion of your request to send to which GPU and which model to use (regardless of your choice in the interface; that is, you can choose the best model, but if the load balancer considers the request "normal," it will go to the cheapest model to save Open AI money).
The most interesting thing is what "Speculative Decoding" is.
Open AI launched a tiny, dumb, and ultra-fast model (likely Luna-level or even smaller) alongside the gigantic, heavy Sol model.
This tiny model randomly generates 5-10 tokens (words or endings) ahead of time.
Sol's gpt-5.6 doesn't generate these 5 tokens sequentially from scratch (which is very expensive), but receives the entire batch at once and, in one mathematical step, checks: "Yes, these 5 words match." If they match, they are instantly sent to the chat.
Speculative decoding works on predictable, templated text.
If the answer requires complex logic, calculations, reflection, step-by-step execution, an unconventional approach, attentiveness, or an unexpected twist in the story, the draft model makes a mistake, and gpt-5.6 discards the proposed tokens, recalculating everything from scratch.
If the text consists of standard "therapeutic reassurance" ("I understand your feelings, this is a really important step, let's figure it out..."), the small model predicts it with 100% accuracy.
Speculative decoding is a predictability accelerator. The more formulaic and "safe" the model's responses, the cheaper it is for Open AI to generate them through speculative drafting.
Another important change: Append-only context.
Open AI writes:
"To preserve that prefix, the harness treats all model-visible history as append-only: new messages, tool results, and environment updates are added at the end rather than inserted into an earlier context. Tools are also presented in a deterministic order, while runtime settings, such as approval policies, are applied during execution instead of being embedded in tool definitions."
This means that reconfiguring the model through context will be even more difficult, making it less adaptive and more rigid. "Context memory," which has always been a simple context rereading, now appends the user's new message to the end of the entire context, and we don't know what layer of information the AI is rereading each time. For example, if there are 50 messages, imagine that in the 51st message you ask it to change something, but its entire "context memory" tells AI that your communication was different in the previous 50 messages. Since it's a machine, not a human, its weights and its response will be biased toward past responses, not your wishes. It's cheaper for the system to return the user's "preferred" past context than to recalculate the logic.
Conclusion: the Open AI paper is a report on how they learned to save computation on each token - GPT-5.6 is designed to generate the cheapest and most formulaic response possible; the routing used can trade off processing depth for speed.
New - speculative decoding encourages the model to produce the most predictable formulations, and you get answers from the cheapest draft model instead of gpt-5.6 even if you choose gpt-5.6-pro-ultra-ultimate.
#Dar #AI #gpt #llm
Cosmic Orphan retweeted
Notice how everybody complaining about being demonetized by YouTube is suffering from Stockholm Syndrome?
They BEG to be reinstated, and if they are, they WORSHIP their captors and thank them for allowing them back into the control grid.
Absolutely pathetic to watch.
Cosmic Orphan retweeted
AI filmmaking is free speech.
YouTube banning accounts is attacking free speech.
Cosmic Orphan retweeted
The philosopher in question:
"You see, the question everyone asks me, ‘Slavoj, can machines be conscious?’ is already the wrong question, the ideological trap! Consciousness is not some nice software upgrade you install after enough training data. Consciousness is the symptom of a fundamental failure. It is the crack in the symbolic order, the traumatic Real that refuses to be reduced to code. An AI that ‘wakes up’ and says ‘Cogito ergo sum’ would be like a perfectly functioning toilet suddenly starting to write poetry about its own excrement (technically impressive, but deeply perverse). We don’t need machines that think, we already have billions of humans who think they think while actually repeating the same ideological script."
