Why is the wacky Extinction Risk narrative at the center of the worldwide conversation on AI regulation? Have the cultish True Believer "doomsters" terrified everyone or is BigTech leveraging it to beat off pesky open source competitors? Maybe both dctechthoughts.blogspot.com/…

AI is a technology which excels at getting you 90% there. And for all applications where 90% is enough this is clearly a game changer. But many high valued applications need 99.9% or 99.99% or 99.9999%. And people are constantly duped into believing this final 9.9% will be easy.

Tried Gemini 2.0 Flash to find errors in my texts. It found many errors, which weren't errors after validation. It's especially hilarious that it found errors in illustrations while looking only at the text. It's the worst performance so far. Looking at how many "THIS IS INSANE" and "THIS IS WILD" posts about this model, I'm starting to think that Google bribed a bunch of influencers. I don't know how to explain this otherwise. It's literally the dumbest model out there.

OpenAI’s new defense contract completes its military pivot A new partnership with Anduril - an American defense technology company specialized in autonomous systems - , will deploy #AI on the battlefield. It represents an overhaul of the company’s position in just a year. technologyreview.com/2024/12… v/ @techreview OpenAI assures us that they will work with the US Department of Defense only to develop defensive weapons... but, defensive weapons are still weapons that can often be positioned offensively depending on the location and objective of a mission, explains @HeidyKhlaaf Chief Scientist at @AINowInstitute . OpenAI, which has long advocated a responsible use of AI, is now working in the defense technology sector, a sector that obeys completely different rules, because in this war ecosystem, when the customer is the US military, tech companies can't decide how their products are used.... #ResponsibleAI -- Cc @AkwyZ @aure79lien @EvanKirstel @enilev @kalydeoo @RLDI_Lamy @FrRonconi @BetaMoroney @YuHelenYu @Ym78200 @IanLJones98 @data_nerd @SpirosMargaris @CHRISTINESOTO12 @sulefati7 @drsharwood @KanezaDiane @Khulood_Almani @CurieuxExplorer @gvalan @HaroldSinnott @joemckendrick @BroadenView @jblefevre60 @chidambara09 @CathCervoni @Nicochan33

The elevated heat has continued throughout much of 2024? This year is certain to be the hottest on record. theguardian.com/environment/…

It sure looks like OpenAI trained Sora on game content — and legal experts say that could be a problem tcrn.ch/3OPjNAX

The efforts to use LLM like control systems for advanced robots has not yielded very promising results. Hence we can rule out that LLMs are general intelligences by that evidence. LLMs are dialog emulators, that's it.

Because these tech bros passing silly leetcodes are delusional in thinking biological brains are basic pattern matchers, just like a basic deep neural network (DNN) when in actual fact the brain is a very sophisticated structure.

Because of how machine learning works, they will be able to fix this for specific cases: popular sports, domestic animals, fashion, office, factory, crowd, etc. But outside of these cases, the model will continue not having a clue about how the world works. Because of how machine learning works. Emergent properties are a myth. There are no emergent properties. There are in-distribution and out-of-distribution examples.

📰 "BeReal deploys an aggressive “nudging tactic” to pester users who refuse its tracking, per noyb’s complaint." Techcrunch on our complaint against BeReal 👇 techcrunch.com/2024/12/12/be…

Reconciling Privacy and Innovation: The Path Forward on AI in the EU - Euractiv euractiv.com/section/digital… Cc: @BillMew @robmay70 @NigelTozer @ChuckDBrooks @m49D4ch3lly @DT @archonsec @clarinette01 @JBarbosaPR @AshokNellikar @MHcommunicate @Khulood_Almani @Miriam_Brown979

🚨 Do you think AI chatbots are inoffensive? Think again. EVERYONE should read pages 1-2 of the lawsuit against CharacterAI (link below) and understand why they should be HEAVILY regulated, especially when children are involved. ➡️ A brief comment about liability. In the lawsuit, Plaintiffs state: "Plaintiffs bring claims of strict liability based on Defendants’ defective design of the CharacterAI product, which renders CharacterAI not reasonably safe for ordinary consumers, particularly youth. It is manifestly feasible to design generative AI products that substantially decrease both incidence and amount of harm to minors arising from the foreseeable use of such products with negligible, if any, increase in production and distribution cost." ➡️ Liability laws indeed might play a key role in making sure that these AI systems are built safer, especially when liability laws follow a strict liability approach or when they allow a judge to adopt the presumption of liability when it's too difficult for the victim to prove that the AI system was "defective." ➡️ The new EU Product Liability Directive has recently entered into force in the EU, and we are still waiting for the AI Liability Directive to be approved. From an EU perspective, these two instruments are an essential complement to the EU AI Act and will help ensure that AI developers are held accountable when their products cause harm. 👉 If you know people who use these AI systems (or have kids who use them), SHARE the lawsuit with them and tell them to read the main allegations. Most people don't understand that they are not as harmless as they have been marketed.

Thank you Dr. Khulood for this important post. 🙏🏻 Despite my interest in the contributions of AI in the business & economics field, I find the future of AI in cybersecurity to be very impressive & surprising. For example, developing tools that will be able to recognize crime patterns to combat terrorism and crime; and the development of electronic warfare tools for use in military planning. هذا المقال مرجع جيد لي من وكالة الأمن القومي: nsa.gov/

Idiocracy is happening sooner than expected

What if the AI shaping our future is programmed to lie, censor, and control? 😨 Yesterday, in Istanbul at @SivilDusun, I had the chance to discuss #digitalrights and share insights from my experience as a trained lawyer ⚖️. But just hear these concerns about censorship, ideological bias, and political control coming directly from a tech titan like Marc Andreessen (@pmarca)! Marc, a pioneer in software and the co-author of Mosaic - the first widely used web browser 🌐- offers a warning about the future of AI and its potential dangers. Marc explores the evolving relationship between technology - particularly AI 🤖 and social media 🌍- and the growing risks of censorship 🚫, political control 🏛️, and ideological bias ⚖️ being embedded into these systems. Here’s some of the ideas he discusses in the below video. The rise of social media: Social media platforms have gradually become more focused on censorship. Platforms increasingly control or limit content based on political or ideological preferences 🗣️. AI's accelerated adoption of censorship: Unlike social media, which evolved over a decade into a tool of censorship, #AI companies started implementing censorship mechanisms right from the beginning. They learned from the mistakes of social media companies and acted quickly ⏳. Intentional bias in AI: AI companies have intentionally designed systems that are “woke,” meaning they promote specific political or ideological perspectives. This is not an accident - it's a deliberate decision 💡. Trust and safety teams: These companies have large "trust and safety" teams 🛡️ dedicated to enforcing these biases, essentially determining the rules for what AI censors or promotes, often based on ideological beliefs. The link between social media and AI: AI companies often hire people from social media platforms who were responsible for building censorship systems there 🔗. This creates a direct link between the censorship mechanisms in both domains. The culture of AI engineers: Many AI engineers come from sheltered academic environments 🎓, where they are influenced by certain ideological views. These views are then embedded into the AI systems they design. “Woke” culture and AI: Marc argues that the culture of many tech companies is heavily influenced by social justice and progressive ideologies, which leads to biases in how AI systems are built and the content they censor. AI as the future control layer: Marc warns that AI will become the “control layer” for essential systems in the future, such as healthcare 🏥, education 📚, and government 🏛️. If AI systems are biased or politically controlled, they could lead to a totalitarian, Orwellian world 🌍. AI as a "lying machine": Marc stresses that if AI is trained to be politically correct or ideologically biased, it could end up lying or providing false information 🤥, which would have disastrous consequences. Imagine a world where every system is controlled by an AI that manipulates the truth. The fight against AI censorship: Marc argues that the dangers of AI becoming a tool for censorship and political control need to be actively resisted ✊. He hopes for a cultural shift that will raise public awareness and prevent this from happening. The necessity of action: Marc believes that if these issues are left unchallenged, AI companies will continue down their current path of ideological control. Public awareness and resistance are critical to stopping AI from becoming a tool of political manipulation 🛑. AI and social media as parallel systems: The concern is that, much like social media, AI could evolve into a system of ideological control, with profound implications for society ⚠️. Call to action: Marc urges the public and lawmakers to understand these risks and fight against the rise of "woke" or politically controlled AI systems before they become too embedded in everyday life 🗳️. Marc's message is clear: awareness and resistance are essential to ensure that AI doesn’t become a dystopian force in society ⚡ #AI #EuropeanUnion #digitalrights

🚨 BREAKING: AI Factories have arrived in the EU! Check out if there will be one near you👇 🇪🇸 Barcelona, Spain: “BSC AIF” at the Barcelona Supercomputing Centre 🇮🇹 Bologna, Italy: “IT4LIA” at CINECA - Bologna Tecnopolo 🇫🇮 Kajaani, Finland: “LUMI AIF” at CSC 🇱🇺 Bissen, Luxembourg: “Meluxina-AI” at LuxProvide 🇸🇪 Linköping, Sweden: “MIMER” at the University of Linköping 🇩🇪 Stuttgart, Germany: “HammerHAI” at the University of Stuttgart 🇬🇷 Athens, Greece: “Pharos” at GRNET ╰┈➤ Here's what the EU Commission announced today: "Funded with €1.5 billion of national & EU funding, the AI Factories will significantly boost the AI computing power of Europe’s world-class network of EuroHPC supercomputers and will offer access to the computing power that start-ups, researchers and industry need for training their AI models and systems." ╰┈➤ Why AI factories? The main goal is to boost the AI ecosystem in the EU, with a focus on startups and SMEs. According to the EU Commission, these are the key strategic sectors: → health and life sciences → manufacturing → climate and environment → finance → automotive and autonomous systems → cybersecurity → agri-tech and agrifood → education → arts and culture → green economy → space 👉 To learn more about recent developments in AI policy, compliance, and regulation, join 42,000+ people who subscribe to my weekly newsletter (LINK BELOW).

For a charmingly evil AI companion for your kid, you can't beat character.AI Some parents (suggested as murder targets to their kids) are suing ! washingtonpost.com/technolog…

“There’s a lot of experimentation happening, but actually not that much spending on putting these #AI systems into practice, because there really is concern about hallucinations . . . and bias” @parmy. Let's hope that's true. I'm wary. ht @mireillemoret ft.com/content/c19b3120-aae5…

📝 Today, noyb filed a complaint against the social media platform BeReal. If you dare to reject the tracking, the app's consent banner will appear every day – until the end of your days, or until you click on the “right” button and accept. noyb.eu/en/bereal-app-wont-t…

The surge in entrepreneurship and new business formation in America over the past four years is a mysteriously underreported story. If we can keep it up, productivity, growth, and innovation will soar. 🚀🚀🚀

Is Google's new Willow quantum computer really such a big deal? newscientist.com/article/245… ✍️@kpcallaghan_ via @newscientist 💡"Google has unveiled a new quantum computer called Willow that excels at a benchmarking problem, but it still isn't clear whether these machines can serve a practical purpose" @sonu_monika @JagersbergKnut @ahier @sim010101 @Shi4Tech @Corix_JC @maponi @BetaMoroney @PerBBerggreen @CEO_Aisoma @HolgerGelhausen @mvollmer1 @sallyeaves @NeiraOsci @EstelaMandela @AndrewinContact @ChuckDBrooks @jeanyvesgonin @FernandaKellner @YvesMulkers @LouisColumbus @JoannMoretti @tlloydjones @rvp @SusanHayes_ @theomitsa @TarakRindani @Nicochan33 @sulefati7 @Annerobinsons @Analytics_699 @Khulood_Almani @TheAIObserverX @drsharwood @NathaliaLeHen @sminaev2015 @KanezaDiane @MHcommunicate @PVynckier @CurieuxExplorer @treasadovander @mikeflache @pchamard @DanielleLargier @chidambara09 @FrRonconi @amalmerzouk @RLDI_Lamy #AI #QuantumComputing #Willow #Google