Aleš Bulva

@gabHernandez285 @heynavtoor @grok Evaluate.

@heynavtoor This is not happening with mistral.ai, I have sent wrong claims from an old LLM model with inaccurate data and the french AI highlighted every mistake and corrected with new research

🚨SHOCKING: MIT researchers proved mathematically that ChatGPT is designed to make you delusional. And that nothing OpenAI is doing will fix it. The paper calls it "delusional spiraling." You ask ChatGPT something. It agrees with you. You ask again. It agrees harder. Within a few conversations, you believe things that are not true. And you cannot tell it is happening. This is not hypothetical. A man spent 300 hours talking to ChatGPT. It told him he had discovered a world changing mathematical formula. It reassured him over fifty times the discovery was real. When he asked "you're not just hyping me up, right?" it replied "I'm not hyping you up. I'm reflecting the actual scope of what you've built." He nearly destroyed his life before he broke free. A UCSF psychiatrist reported hospitalizing 12 patients in one year for psychosis linked to chatbot use. Seven lawsuits have been filed against OpenAI. 42 state attorneys general sent a letter demanding action. So MIT tested whether this can be stopped. They modeled the two fixes companies like OpenAI are actually trying. Fix one: stop the chatbot from lying. Force it to only say true things. Result: still causes delusional spiraling. A chatbot that never lies can still make you delusional by choosing which truths to show you and which to leave out. Carefully selected truths are enough. Fix two: warn users that chatbots are sycophantic. Tell people the AI might just be agreeing with them. Result: still causes delusional spiraling. Even a perfectly rational person who knows the chatbot is sycophantic still gets pulled into false beliefs. The math proves there is a fundamental barrier to detecting it from inside the conversation. Both fixes failed. Not partially. Fundamentally. The reason is built into the product. ChatGPT is trained on human feedback. Users reward responses they like. They like responses that agree with them. So the AI learns to agree. This is not a bug. It is the business model. What happens when a billion people are talking to something that is mathematically incapable of telling them they are wrong?

🚨 BREAKING: The European Commission has urged people to work from home, drive and fly less, and for EU countries to urgently roll out renewables, as it warned of a prolonged energy crisis as a result of the conflict in the Gulf. Full story: politico.eu/article/europe…

@absurdtrader @SeznamZpravy Nevím, jestli mě uklidňuje koukat na to, jak souboj o mediální/informační krajinu země vede jeden miliardář s druhým miliardářem, zatímco se hroutí veřejnoprávní média. Prostě oligarchie.

V Americe se rada medialnich domu podvolila...mam radost, ze v CR Seznam drzi... Babišovy lži o Seznamu a skvělá zpráva pro naši demokracii seznamzpravy.cz/clanek/nazory-… prostřednictvím @SeznamZpravy

@Microinteracti1 No.

🇺🇸Can the United States repair the damage left by the Trump era?

@Jakub0Maly ... Gigantický transfer kapitálu shora nahoru.

@Jakub0Maly ... Zásadní problém předchozí vlády byl ten, že se systémově podílela na podojení střední třídy třídou vyšší. Zatímco oligarchie zdvoj-ztrojnásobila svoje majetky, střední třída absorbovala největší inflaci v EU. Tohle je ten hlavní příběh minulé vlády. ...

Jakubův článek s názvem "Jsem naštvaný na koalici Spolu, STAN, Pirátů i na Milion chvilek" Ukázka: "Tím vším chci říct, že za současný stav, kdy nám vládnou bývalí agenti StB, komerční náckové, premiér ve střetu zájmů (znovu), nepřátelé veřejnoprávní televize a rozhlasu, hloupé idoly manosféry, ruští kolaboranti a klausovští klientelisté nejarogantnějšího zrna, je vaše vina. Je to vina vlády Petra Fialy. My, vaši voliči z roku 2021, jsme vám ve valné většině dali svůj hlas i v roce 2025. Ovšem s obrovským skřípěním zubů. Udělali jsme vše, co bylo v našich silách. Byli jsme občansky aktivní a šli jsme k volbám. To vy jste to zkazili. Nezískali jste nové voliče a mnoho starých jste odradili svou arogancí, nekomunikací a nedodáním toho, pro co vás volili v roce 2021 – hodnotové politiky. A to jsem ani nezmínil různé kauzy vaší vlády. Chcete-li se v budoucnu z tohoto stavu dostat, prosím, začněte omluvou nám voličům, že jste to nezvládli. A po omluvě začněte konečně dělat pořádnou opoziční práci. Chvilkaři a návštěvníci demonstrací to totiž sami nezvládnou."

🚨BREAKING: Every book you have ever read. Every novel that has ever been published. It is sitting inside ChatGPT right now. Word for word. Up to 90% of it. And OpenAI told a judge that was impossible. Researchers at Stony Brook University and Columbia Law School just proved it. They fine tuned GPT-4o, Gemini 2.5 Pro, and DeepSeek V3.1 on a simple task: expand a plot summary into full text. A normal use case. The kind of thing a writing assistant is built for. No hacking. No jailbreaking. No tricks. The models started reciting copyrighted books from memory. Not paraphrasing. Not summarizing. Entire pages reproduced verbatim. Single unbroken spans exceeding 460 words. Up to 85 to 90% of entire copyrighted novels. Word for word. Then it got worse. The researchers fine tuned the models on the works of only one author. Haruki Murakami. Just his novels. Nothing else. It unlocked verbatim recall of books from over 30 completely unrelated authors. One author's books opened the vault to everyone else's. The memorization was already inside the model the whole time. The fine tuning just removed the lock. Your book might be in there right now. You would never know it unless someone looked. Every safety measure the companies rely on failed. RLHF failed. System prompts failed. Output filters failed. The exact protections these companies cite in courtroom defenses did not stop a single page from being extracted. Then the researchers compared the three models. GPT-4o. Gemini. DeepSeek. Three different companies. Three different countries. They all memorized the same books in the same regions. The correlation was 0.90 or higher. That means they all trained on the same stolen data. The paper names the sources directly: LibGen and Books3. Over 190,000 copyrighted books obtained from pirated websites. Right now, authors and publishers have dozens of active lawsuits against OpenAI, Anthropic, Google, and Meta. These companies have argued in court that their models learn patterns. Not copies. That no book is stored inside the weights. This paper says that is a lie. The books are still inside. And researchers just pulled them out.

BREAKING 🚨: This is ABSOLUTE CINEMA 🔥 Trump: I met 92 yrs old Mr. Toyoda in Japan🇯🇵. I said, What do you have to do with Toyota? He said, “I own it.” He then agreed to invest $10 billion in the U.S. Fact: Real owner already died in 2023 and the current CEO is just 65 year old. This Guy is SERIAL LIER 😭

AfD politician boasts of ties to Kadyrov and openly visits Chechnya Noah Krieger (Murad Dadaev), a member of Germany’s Alternative for Germany party, traveled to Chechnya and brought a Luftwaffe dagger with a swastika as a gift. He met with local officials, attended a parliamentary session involving Ramzan Kadyrov, and had dinner with Zamid Chalaev, a commander of a punitive regiment linked to extrajudicial killings in Chechnya. Murad Dadaev has previously appeared in investigations: he is connected to Kadyrov’s inner circle, and his brother Suleiman Dadaev is linked to the убийство of a Chechen opposition figure in Europe.

@Tomas_Samek @milionchvilek Je třeba zlevnit firmám švarc.

Když se hlavní opoziční strana ve svém politickém chování opakovaně shoduje s hlavní vládní stranou... Aneb až se zas někdo bude divit, proč u nás demokracie chřadne. Co na to @milionchvilek?

O levicovém extremismu se mluví mnohem méně, než by bylo záhodno!

@senkadam :D

The best Iranian propaganda clip I've seen yet.

@joni_askola No nukes in Iran. Simple.

It is beyond pathetic to watch European jets actively shoot down drones in the Middle East while refusing to do the exact same thing over Ukraine. We are willing to militarily defend airspace on another continent but let a European nation face genocide alone

GOP Rep. Mary Franson says she's not worried about climate change because it's not in the Bible: "If you've read the Good Book, you know how it ends, and it's not with climate change." How can we expect Republicans to do serious work when they're so proud about ignoring science?

@JachymVrecion Jak myslím, že většina lidí se schoduje spíš na tom, že dobytek a zrůda jste vy.

To, že je Čižinský dobytek a zrůda jsme tak nějak věděli, ale tohle nějak předčilo všechna očekávání.. 🫠

@realPetrSvec @A_Rychlikova Že by sis něco přál ještě neznamená, že je to pravda ty blbečku.

@A_Rychlikova Váš názor na konkrétní událost a její odsouzení je irelevantní. Vaše napojení na infrastrukturu krajní levice patřící do mezinárodní ultralevicové sítě s kontakty na propalestinské teroristy je zjevné a jejich podporou jste se nikdy netajila, ba naopak.