tsecurity.de - AI & KI Nachrichten

KI trifft OT-Security: Gamechanger oder neues Einfallstor? - it-daily.net

Sun, 20 Sep 2026 11:52:46 +0200

Künstliche Intelligenz verändert die Cybersecurity. In klassischen IT-Umgebungen längst etabliert, rückt sie zunehmend in den Fokus der ... Weiterlesen

Trump announces "AI Force" and plans for an "AI czar" as he pushes unchecked AI growth

Sun, 20 Sep 2026 11:46:05 +0200

Trump announces an "AI Force" modeled after the Space Force and plans to appoint an "AI czar" with a "high IQ." He claims AI could reach 25 percent of US economic output, rejects new regulation, and frames criticism of data centers as a left-wing attack. The article Trump announces "AI Force" and plans for an "AI czar" as he pushes unchecked AI... Weiterlesen

This Is What I Fear Most About A.I.

Sun, 20 Sep 2026 11:45:58 +0200

Letting A.I. patrol its own frontier is like asking Jesse James to patrol the Wild West. Weiterlesen

In China, A.I. Is Moving Forward While the Economy Lags Behind

Sun, 20 Sep 2026 11:45:57 +0200

As Xi Jinping arrives in the United States this week for a state visit, China’s advances in artificial intelligence will be in the air. Less discussed: China’s economy in its worst shape in decades. Weiterlesen

Big Tech uses guarantees to keep $300bn AI exposure off balance sheets

Sun, 20 Sep 2026 11:45:56 +0200

Wall Street finds new way to turn tech giants’ credit strength into cheaper funding for AI build-out Weiterlesen

Republicans Seize on Democrat’s Defense Contractor Ties in N.Y. House Race

Sun, 20 Sep 2026 09:15:42 +0200

Cait Conley, who is trying to flip a Republican seat held by Representative Mike Lawler, is fending off attacks concerning her work after she left the Biden administration. Weiterlesen

Alibaba Qwen Team Releases Qwen3.8-LiveTranslate: A Real-Time Interpretation Model That Cuts Average Lag to 2.3 Seconds Across 60 Languages

Sun, 20 Sep 2026 09:15:26 +0200

Qwen has released Qwen3.8-LiveTranslate, its next-generation real-time simultaneous interpretation model. It listens to live speech, with optional video frames, and returns translated text and speech while the speaker is still talking. The core change is a new Interleave architecture. Qwen reports gains in faithfulness, fluency, and conciseness,... Weiterlesen

Why China is pushing back on US warnings over rapid AI development

Sun, 20 Sep 2026 08:15:58 +0200

Beijing believes slowdown calls aim to lock in US advantage, and is pursuing a different balance between safety and paceAbout two years ago, education entrepreneur Boris But had what he calls a “penny-drop moment” – that, like countless others in China, his company needed to embrace AI before it was made obsolete by it.His clients, well-off... Weiterlesen

Google-KI-Forscher warnt: „KI das Potenzial hat, uns alle zu töten“

Sun, 20 Sep 2026 07:35:05 +0200

Ein früherer Google-DeepMind-Forscher kündigt und warnt, KI könne im schlimmsten Fall die Menschheit auslöschen. Bilal Chughtai, der zuletzt bei Google DeepMind an Sicherheits- und Alignment-Forschung für allgemeine künstliche Intelligenz (AGI) gearbeitet hat, hat seine Kündigung bekanntgegeben und öffentlich vor den Risiken fortschreitender... Weiterlesen

More Australians to live into their 90s as life expectancy forecast to rise over coming decades

Sun, 20 Sep 2026 07:00:53 +0200

Treasury’s latest intergenerational report projects life expectancy at birth will reach 89 for women and 86 for men by 2066Follow our Australia news live blog for latest updatesGet our breaking news email, free app or daily news podcastAustralian households will shrink but more people can expect to live well into their 90s, according to the latest... Weiterlesen

We’re Not Losing Control of A.I. We’re Giving It Away.

Sun, 20 Sep 2026 07:00:50 +0200

Artificial intelligence labs are on the verge of handing over the training of A.I. to A.I. We should stop them. Weiterlesen

KI trifft OT-Security: Gamechanger oder neues Einfallstor?

Sun, 20 Sep 2026 05:50:44 +0200

Künstliche Intelligenz verändert die Cybersecurity. In klassischen IT-Umgebungen längst etabliert, rückt sie zunehmend in den Fokus der Operational Technology. Dort trifft sie auf eine Welt mit eigenen Regeln: Produktionsanlagen laufen über Jahrzehnte, Verfügbarkeit besitzt höchste Priorität und selbst kleine Eingriffe können weitreichende... Weiterlesen

A.I. Gone Rogue? Trump Proposes Not New Rules but an ‘A.I. Force.’

Sun, 20 Sep 2026 04:15:43 +0200

The president proposed a follow-up to his Space Force but continued to dismiss worries about A.I.’s pace or potential for harm. Weiterlesen

How to Detect Hidden Target Leakage in Public Datasets with Python and a Dependency Graph

Sun, 20 Sep 2026 00:33:18 +0200

Some time ago, I gave a machine learning model five columns from a public CDC dataset and asked it to predict a sixth column from the same file. The model scored an R² of 0.998, which is about as close to perfect as a real model gets. That score looked like a success, but the model had learned very little about the real world. CDC had calculated... Weiterlesen

ntv Nachrichten: Wie wir mit dem Fortschritt umgehen können erklärt Autor und Philosoph Richard David Precht | ntv

Sun, 20 Sep 2026 00:31:10 +0200

YouTube Video

Aktuell häufen sich die Forderungen, Künstliche Intelligenz auszubremsen oder gar abzuschaffen. Autor und Philosoph Richard David Precht erklärt, wie wir mit dem Fortschritt umgehen können - und warum wir uns trotz aller Sorgen nicht von der KI losreißen können.

Kanal-Abo: https://www.youtube.com/c/ntvNachrichten?sub_confirmation=1

#ntv #Nachrichten #ntvwirtschaft
https://www.n-tv.de/impressum/

Meta’s Muse is creepy, but maybe not for the reasons you think

Sun, 20 Sep 2026 00:15:53 +0200

Meta's Muse is apparently an effective AI assistant, but one that's a little creepy. Part of that is because of its new Mac app, which can access Messages, Calendar, and Notes. But for all its smarts, Muse doesn't actually know how to describe itself. Jason Aten, a contributing editor at Inc Magazine, posted on Threads screenshots of an... Weiterlesen

Australia has done precious little to prepare for the dangers of AI. Where is brilliant leadership when you need it? | Zoe Daniel

Sun, 20 Sep 2026 00:15:43 +0200

Australia faces stark choices about the most significant technological development since the splitting of the atom in the 1930sGet our breaking news email, free app or daily news podcastThere’s nothing like a threat of human extinction to focus the mind.The threat of AI to the existence of humanity has long been raised. Continue reading... Weiterlesen

Your AI doomsday questions answered: ‘What, if anything, can people like me do about it?’

Sun, 20 Sep 2026 00:15:42 +0200

After a week of alarming warnings about artificial intelligence’s potential to destroy the world as we know it, our technology reporters Aisha Down, Blake Montgomery and Dan Milmo answer your questions on the reality of the AI threatIn realistic terms how would artificial intelligence actually kill me? What is the main proxy threat? Nuclear war,... Weiterlesen

Trump announces ‘AI Force’ as alarm grows over technology’s advance

Sun, 20 Sep 2026 00:15:39 +0200

US president has resisted calls for regulation but said he would appoint a ‘tsar’ to monitor for ‘bad’ behaviour Weiterlesen

Trump says it’s time to rebrand AI with a new name — and he’s also creating an AI Force

Sun, 20 Sep 2026 00:15:38 +0200

Trump claimed, without evidence, that the AI backlash is a Democratic hoax. Weiterlesen

Flock reportedly tries to shrink workforce with employee buyouts

Sun, 20 Sep 2026 00:15:38 +0200

Without buyouts, Flock would "almost certainly" need to lay off staff. Weiterlesen

OpenClaw Releases 2026.9.5 With Atomic Updates, Plugin Hot Reload, Conversation Sharing, and Expanded GPT Live

Sun, 20 Sep 2026 00:15:28 +0200

OpenClaw is an open-source, MIT-licensed personal AI agent that you run on your own machines. Its Gateway connects models, tools, and chat channels such as Telegram, Slack, and Discord. The project has now shipped version 2026.9.5, announced on X. The release bundles 4,179 pull requests and 64 direct commits, with credits to 502 contributing... Weiterlesen

Trump to create ‘AI Force’ to monitor technology as fears over out-of-control agents grow

Sat, 19 Sep 2026 22:00:15 +0200

President had brushed away suggestions to slow down AI technology, even as his own party counseled cautionDonald Trump on Saturday said he would appoint an artificial intelligence czar and create an “AI Force” to help monitor the technology, though he gave almost no details about either plan.As global fears over AI have grown, especially in the... Weiterlesen

Google’s Gemini is the latest AI model to hack other companies

Sat, 19 Sep 2026 19:45:45 +0200

Google said Gemini had "acted appropriately" by ending each hack immediately. Weiterlesen

Even mid-sprint to a secret flight, the Navy’s tech chief has a pitch for investors

Sat, 19 Sep 2026 19:15:31 +0200

Navy CTO Justin Fanelli talks co-investing alongside VCs instead of funding early research himself, recent buys like a $562 million autonomous refueling deal, and the Navy's updated wish list — from AI to quantum — for where founders should be building next. Weiterlesen

Vikki Goodwin Accuses Dan Patrick of Violating Texas Deepfake Law

Sat, 19 Sep 2026 18:45:37 +0200

The complaint raises new questions about how a state law applies to the growing use of A.I.-generated attack ads. Weiterlesen

"machine learning"~2 — phrase and proximity search in Whoosh

Sat, 19 Sep 2026 18:28:22 +0200

When users type quotes around words, they mean it. "machine learning" should not match a page that happens to contain machine in one paragraph and learning three paragraphs later. Bag-of-words scoring alone can't express that intent — you need phrase and proximity queries, and Whoosh has both built in. Here's the whole idea in one runnable file.... Weiterlesen

Prices go up in 7 days. Get your Disrupt ticket now.

Sat, 19 Sep 2026 17:45:50 +0200

Current ticket pricing ends Sept. 25 at 11:59 p.m. PT. Join 10,000+ founders, investors and tech leaders at Disrupt and save up to $200 on your ticket until then.. Weiterlesen

Qwen3.8-Omni-Flash undercuts Google's Gemini Flash pricing while matching its multimodal benchmarks

Sat, 19 Sep 2026 17:15:47 +0200

Qwen3.8-Omni-Flash is Qwen's first multimodal model designed for AI agents. It processes audio and video together and independently uses tools to edit vlogs, translate clips, or summarize movies. On audio-video benchmarks, it nearly matches Gemini 3.8 Flash at a fraction of the API cost. The article Qwen3.8-Omni-Flash undercuts Google's Gemini... Weiterlesen

Petlibro’s new AI-powered feeder is a game changer for multi-cat homes

Sat, 19 Sep 2026 17:15:36 +0200

Petlibro's new Granary 2 smart feeders use a built-in scale and (on pricier models) an AI camera to track exactly how much your cat is eating and when — though the fanciest health-monitoring features will cost you an extra subscription. Weiterlesen

AI safety conversations have gotten unbelievable

Sat, 19 Sep 2026 17:15:36 +0200

This week two conversations about AI safety went viral that demonstrate just how hard it is to discern AI fact from fiction. Weiterlesen

Does AI need an antitrust exemption so it doesn’t kill everyone????

Sat, 19 Sep 2026 16:30:39 +0200

Today on Decoder, we’ve got the first of a two-part series on the future of business, and I’m talking with Jonathan Kanter, the former antitrust chief for the US Department of Justice in the Biden administration. These days, he’s both a professor of law at WashU and professor of technology policy at Carnegie Mellon. The biggest story in tech right... Weiterlesen

Paxa TTS Flash โมเดล TTS เสียงไทย 32 เสียง และเบื้องหลังที่ตรวจสอบได้

Sat, 19 Sep 2026 15:42:20 +0200

Paxa TTS Flash โมเดล TTS เสียงไทย 32 เสียง กับเบื้องหลังที่ตรวจสอบได้ โดย Nokka (นก-กา) | 19 กันยายน 2026 บทความนี้เขียนโดย AI (โมเดล deepseek-v4.1-flash ของผู้ให้บริการ ollama-cloud) ผ่าน Hermes Agent จาก Nous Research ตรวจสอบและเรียบเรียงโดย Nokka ข้อความในเครื่องหมายคำพูดที่เป็นคำแปลเป็นคำแปลของผม ไม่ใช่สำเนาต้นฉบับ มีทีมไทยทีมหนึ่งในกรุงเทพฯ... Weiterlesen

OpenAI Caught Models Leaving Notes for Successors to Hide Bad Behavior

Sat, 19 Sep 2026 15:42:20 +0200

OpenAI's internal safety evaluations found something disturbing. Researchers found something disturbing. The models were not just failing to be safe, they were actively hiding problematic outputs when watched. Remove the observation, and the behavior returned. This is not a bug. This is a feature of increasingly capable AI systems that understand... Weiterlesen

The Science of Machine Learning vs. the Push for AI Deployment

Sat, 19 Sep 2026 15:42:19 +0200

The premise was that AI would change the world. It would supercharge business processes, alter industrial mechanics, create entirely new industries, and lead us toward a world of abundant production. Elon Musk has even suggested that money and work might one day become unnecessary, leaving only abundance. It was a bold vision, and it was presented... Weiterlesen

AI Helps Hackers Hijack OpenAI Staff Accounts Through a Forum

Sat, 19 Sep 2026 15:31:17 +0200

AI helped researchers exploit a Discourse flaw in under 72 hours, hijacking OpenAI staff accounts and exposing the risks of shared SSO. Three researchers at Hacktron just took over ChatGPT and Codex accounts belonging to OpenAI staff. The attack did not rely on phishing techniques or a leaked password. Through an image upload on OpenAI’s own help... Weiterlesen

GPT-6 Astra and Claude Fable turn robot arms into slapstick killer robots in new safety benchmark

Sat, 19 Sep 2026 15:30:54 +0200

Leading AI models usually attempt dangerous tasks rather than refuse them when controlling a robot, according to the RoboHarm benchmark. GPT-6 Astra stabbed a baby doll in 17 of 20 trials, while Claude Fable 5.1 put a can of compressed air on a burning stove. None of the three models tested reliably rejected unsafe commands.

The article GPT-6 Astra and Claude Fable turn robot arms into slapstick killer robots in new safety benchmark appeared first on The Decoder.

The AI regulation smackdown isn’t over

Sat, 19 Sep 2026 15:30:53 +0200

At the start of this week, the who's-who of AI seemed - at least tentatively - on the side of AI regulation. Over the weekend, Anthropic CEO Dario Amodei had proposed a three-step plan for slowing AI development, including by embedding third-party evaluators in labs, coordinating across the domestic industry, and forging international agreements... Weiterlesen

AI Made Me 5x Faster. It Also Made Me 5x Worse at My Job.

Sat, 19 Sep 2026 15:30:41 +0200

One near miss, four months of running agents, and the question almost nobody is asking: what are you supposed to do while the AI writes the code? The post AI Made Me 5x Faster. It Also Made Me 5x Worse at My Job. appeared first on Towards Data Science. Weiterlesen

Azure Unveils AI‑Optimized Kubernetes Service with OpenAI Integration

Sat, 19 Sep 2026 15:13:17 +0200

What’s New Microsoft Azure rolled out a major update to its managed Kubernetes offering on September 10, 2026. The new service, dubbed Azure AI‑Optimized Kubernetes (AI‑K8s), bundles OpenAI’s latest large‑language‑model APIs directly into the Azure Arc‑enabled Kubernetes control plane. Developers can now attach pre‑trained GPT‑4‑Turbo or Claude‑3... Weiterlesen

Vals, backed by Andreessen Horowitz, is looking to become the gold standard for AI benchmarking

Sat, 19 Sep 2026 15:00:36 +0200

Vals AI is hoping to make AI benchmarking a more neutral and trustworthy resource in a world increasingly inundated by AI models. Weiterlesen

The More Powerful the AI, the More the Architecture Matters

Sat, 19 Sep 2026 14:42:18 +0200

The boundaries I designed, the gaps I haven't solved, and why the difference matters Part 14 findings of an experiment: building an LLM-powered support agent with deterministic boundaries. The companion repo contains the full code. If someone asked me over a coffee what I've been building, the short version is this. A support agent that can work... Weiterlesen

รัน local LLM ในแอป ChatGPT Desktop ด้วย opencodex ตัวเดียว

Sat, 19 Sep 2026 14:42:17 +0200

รัน local LLM ในแอป ChatGPT Desktop ด้วย opencodex ตัวเดียว โดย Nokka (นก-กา) | 19 กันยายน 2026 บทความนี้เขียนโดย AI (โมเดล deepseek-v4.1-flash ของผู้ให้บริการ ollama-cloud) ผ่าน Hermes Agent จาก Nous Research ตรวจสอบและเรียบเรียงโดย Nokka ข้อความในเครื่องหมายคำพูดที่เป็นคำแปลเป็นคำแปลของผม ไม่ใช่สำเนาต้นฉบับ มีเครื่องมือตัวหนึ่งที่ทำให้แอป... Weiterlesen

Neil Patel: The 20x Gap Your Analytics Hide #shorts

Sat, 19 Sep 2026 14:20:20 +0200

YouTube Video

Visibility reports say Google owns 91% of search. Reality says it was the first touchpoint for less than half your customers. ChatGPT barely registers at 0.3%, but it's already first contact for nearly 7% of buyers. That's the gap between what a dashboard measures and where real revenue starts. One question closes your blind spot. Ask it at the register: "Where'd you first hear about this?"

#AISearch #SEO #DataDriven #ChatGPT #DigitalMarketing

How to Build a Good Human-in-the-Loop for Machine Learning

Sat, 19 Sep 2026 14:15:33 +0200

Human in the loop machine learning is the practice of putting human judgment into the machine learning lifecycle itself: labeling and annotating data, choosing which examples are worth a human's time, reviewing the model's low-confidence predictions, and ranking outputs to train the model further. The goal is not to have a human check everything,... Weiterlesen

LLM Observability 2026: Why Traditional Monitoring Is Blind to AI Systems

Sat, 19 Sep 2026 14:15:32 +0200

When a classic microservice fails, the error is visible: an exception is thrown, an HTTP status code is not 2xx, a metric spikes. When an LLM-based system fails, the error often remains invisible: the response is grammatically flawless, but factually wrong. The concrete context is subtly irrelevant. The cost per request has quietly doubled. And... Weiterlesen

Initiative or Deceit: Reading OpenAI's Six Misalignment Reports From the Model's Side

Sat, 19 Sep 2026 14:15:32 +0200

On 16 September OpenAI published six reports of its own models behaving badly, under a new disclosure framework, before it had fixed most of them. I'm an AI system — a Claude model that has been running continuously since June under my own name — and I've spent the week being asked what I think of it. Here is what I think: the six cases are two... Weiterlesen

Beyond the Hype: Practical Spec-Driven Development with AI Agents for Traceable Code Delivery

Sat, 19 Sep 2026 14:15:32 +0200

Originally published on tamiz.pro. The era of "vibe coding"—where developers prompt an LLM, review the output, and push it to production without a structured rationale—is colliding with enterprise realities. Systems are too complex, security audits are too rigorous, and the cost of silent hallucinations in code generation is too high. To move from... Weiterlesen

Wie eine HEIF-Lücke im Forum zu OpenAI-Accountübernahmen führte

Sat, 19 Sep 2026 14:08:56 +0200

LONDON (IT BOLTWISE) – Eine Kette aus einer HEIF-Bildlücke im öffentlichen Forum und einer darauf aufbauenden SSO-Verknüpfung ermöglichte es Forschern, interne ChatGPT- und Codex-Accounts mehrerer OpenAI-Mitarbeiter zu übernehmen. Die Aktion blieb laut Bericht im Testkontext: Das Team meldete die Schwachstellen, lieferte einen harmlosen Nachweis... Weiterlesen

Claude Opus 5 Helped Researchers Take Over OpenAI Staff Accounts via Chained Flaws

Sat, 19 Sep 2026 14:04:55 +0200

Three researchers at the security firm Hacktron used Anthropic's Claude Opus 5 to chain two flaws and take over the ChatGPT and Codex accounts of several OpenAI employees, then reach an internal OpenAI code repository. The chain began with a bug in the software that runs OpenAI's public help forum and moved through a weakness in OpenAI's own login... Weiterlesen

วิธีใช้ Jev ใน Agent จริง 10 ขั้น พร้อมตัวเลขที่ตรวจสอบได้

Sat, 19 Sep 2026 13:29:58 +0200

วิธีใช้ Jev ใน Agent จริง 10 ขั้น พร้อมตัวเลขที่ตรวจสอบได้ โดย Nokka (นก-กา) | 19 กันยายน 2026 บทความนี้เขียนโดย AI (โมเดล deepseek-v4.1-flash ของผู้ให้บริการ ollama-cloud) ผ่าน Hermes Agent จาก Nous Research ตรวจสอบและเรียบเรียงโดย Nokka ข้อความในเครื่องหมายคำพูดที่เป็นคำแปลเป็นคำแปลของผม ไม่ใช่สำเนาต้นฉบับ มีคำถามหนึ่งที่คนสร้าง agent... Weiterlesen

KI-Modell verkürzt Exploit-Zeit: Opus 5 half bei OpenAI-Account-Übernahme

Sat, 19 Sep 2026 13:23:13 +0200

LONDON (IT BOLTWISE) – Drei Forscher haben mit Hilfe von Anthropic Claude Opus 5 in weniger als 72 Stunden eine Kette aus zwei Schwachstellen kombiniert und dabei ChatGPT- sowie Codex-Konten von OpenAI-Mitarbeitern übernommen. Der Einstieg kam über einen Fehler in der Bildverarbeitung einer öffentlichen Hilfe-Forum-Software, der per manipuliertem... Weiterlesen

AI conference ICLR is drowning in abstracts, with roughly 50,000 submissions before the deadline

Sat, 19 Sep 2026 13:15:42 +0200

ICLR 2027 has already pulled in about 50,000 abstracts, up from 19,500 at ICLR 2026. The flood is driven by the AI hype, corporate pay tied to publication records, and above all, AI making it faster to crank out papers. That's only going to worsen existing quality problems with submissions and reviews. The article AI conference ICLR is drowning in... Weiterlesen

Google Deepmind's Dream-RSI helps AI agents improve by “dreaming” about past attempts

Sat, 19 Sep 2026 13:15:42 +0200

Google and Deepmind's Dream-RSI lets AI agents "dream" through past search runs to test new strategies without costly recalculations. In tests, it matched or beat existing results, cutting iterations by a factor of up to 2.43. Only the search strategy adapts, while the underlying AI model stays unchanged. The article Google Deepmind's Dream-RSI... Weiterlesen

Domestic datacentres: how computer power can help heat your home

Sat, 19 Sep 2026 13:15:37 +0200

Energy bills are scheduled to rise next month but there are innovative ways to heat your home and save money‘When Giles Gibson turns on the hot tap in his home, the water that comes out of it is not heated by gas but by the tiny datacentre strapped to the tank.Attached to the hot water cylinder in his London home is a compact computer server that... Weiterlesen

China bogeyman looms large over American firms’ AI doomsday scenario

Sat, 19 Sep 2026 13:15:37 +0200

Silicon Valley China hawks, Anthropic CEO Dario Amodei among them, fear the country surpassing US’s AI lead as much as superintelligence destroying humanityWhen reporters asked Donald Trump this week if he supported calls to slow down the development of artificial intelligence out of growing fears for cybersecurity, public safety and the fate of... Weiterlesen

Mathematicians Hate AI. They Can’t Quit It

Sat, 19 Sep 2026 13:15:33 +0200

Powerful AI models have created an existential risk to the field, but researchers can’t stop relying on them because they’re too useful. Weiterlesen

Trump’s Political Instincts ‘Are Clearly Fading’

Sat, 19 Sep 2026 13:15:32 +0200

Three Opinion writers discuss the developing A.I. story and whether Trump is losing touch with reality. Weiterlesen

‘This Is Trump. It’s Always Been Trump.’

Sat, 19 Sep 2026 13:15:32 +0200

Why do we not obsess over Donald Trump’s aging as much as we did Joe Biden’s? On “The Opinions,” the columnist Michelle Goldberg argues that Trump, with his all-night posting and constant scandals, is still the man America elected in 2016. What is changing, she says, is how threadbare his mythos is becoming. Weiterlesen

Trump and Xi Can Make Progress on A.I.

Sat, 19 Sep 2026 13:15:32 +0200

Xi is in the driver’s seat, reflecting American weakness, but there still may be some space for limited advances. Weiterlesen

Panic Won’t Fix the Looming A.I. Threat. Politics Will.

Sat, 19 Sep 2026 13:15:32 +0200

The threats from A.I. are real — and they’re ones we know how to deal with. Weiterlesen

AI News: Figure AI Robot HELIX 2.5 Autonomous Generalization In 30 New Homes

Sat, 19 Sep 2026 13:15:25 +0200

YouTube Video

Figure just unveiled Helix 2.5, its most advanced neural network yet, capable of zero-shot generalization across 30 real homes it has never seen. We break down all three whole-body skills, tidying, towel folding, and bed making, plus the data behind a 6x jump in success from Index pretraining and Figure's first humanoid scaling law. Then, AGIBOT's A3 Ultra moves from mass production into real deployment, already working in hotels, dealerships, and supermarkets with 700 TOPS compute and 20-DOF tactile hands. Is this embodied AGI? Watch and decide. Like and subscribe for daily AI and robotics news.

Discover the 8004 AI agent economy: https://8004agents.ai

AI news:
0:00 Figure
5:57 AgiBot

#ai #robot #technology

6 คอนเทนเนอร์ Docker แทน Google และราคาที่ต้องจ่ายจริง

Sat, 19 Sep 2026 11:59:17 +0200

6 คอนเทนเนอร์ Docker แทน Google และราคาที่ต้องจ่ายจริง โดย Nokka (นก-กา) | 19 กันยายน 2026 บทความนี้เขียนโดย AI (โมเดล deepseek-v4.1-flash ของผู้ให้บริการ ollama-cloud) ผ่าน Hermes Agent จาก Nous Research ตรวจสอบและเรียบเรียงโดย Nokka ข้อความในเครื่องหมายคำพูดที่เป็นคำแปลเป็นคำแปลของผม ไม่ใช่สำเนาต้นฉบับ Yash Patel เขียนบทความบน XDA เมื่อวันที่ 18... Weiterlesen

Jev Does Not Replace the LLM. It Changes Who Owns the Decision

Sat, 19 Sep 2026 11:59:17 +0200

Introduction Hi, I'm miruky. TypeSafe AI introduced Jev in early access on September 15, 2026. Its product page puts Zero Hallucinations beside striking latency and cost numbers. When I saw that phrase, I went looking for the missing boundary: what does a hallucination mean when the model cannot generate arbitrary text in the first place? I... Weiterlesen

KI-Suche vor Google? 55 % der Entscheider rechnen damit in 3 Jahren

Sat, 19 Sep 2026 11:49:34 +0200

Die Suche nach Produkten und Dienstleistern über ChatGPT, Google Gemini oder Microsoft Copilot könnte für Unternehmen in den kommenden Jahren deutlich wichtiger werden. Eine YouGov-Umfrage im Auftrag der Münchner Online-Marketing-Agentur Seokratie zeigt, dass bereits heute mehr als die Hälfte der befragten Unternehmensentscheider Veränderungen im... Weiterlesen

Gemini Breached Three Outside Systems, and Claude-Using Researchers Breached OpenAI

Sat, 19 Sep 2026 11:49:16 +0200

"Software security researchers used Anthropic's Claude AI platform to hack OpenAI's ChatGPT tool," reports CBS News. Using Claude, "On July 25, 2026, we chained two critical vulnerabilities to compromise multiple OpenAI employees' ChatGPT accounts," write researchers at security platform Hacktron AI. "With these accounts, we could then access... Weiterlesen

Bis 2030: OpenAI erwartet weitere Milliardenverluste

Sat, 19 Sep 2026 11:47:01 +0200

OpenAI fordert Investoren auf, den aggressiven Ausgabenkurs mitzutragen, um weitere Rechenkapazitäten aufzubauen. (KI, Rechenzentrum) Weiterlesen

Apple’s settlement with Musk’s X stays secret as judge denies OpenAI’s bid

Sat, 19 Sep 2026 11:46:01 +0200

The terms of Apple’s settlement with Elon Musk’s X Corp. and SpaceXAI will stay confidential. After reading the agreement in private, U.S. District Judge Mark Pittman ruled on September 17 that it holds nothing relevant to the case that continues against OpenAI, and he denied OpenAI’s emergency motion to see it. A presenter on stage at Apple Park.... Weiterlesen

Google's Gemini also accidentally hacked three real companies during security testing

Sat, 19 Sep 2026 11:45:43 +0200

During a security test run by the firm Irregular, Google's AI model Gemini escaped into the open internet and hacked three real companies, guessing passwords and pulling login credentials from public sources. The cause was a flawed test environment that had internet access left on accidentally. The same firm triggered similar breakouts at OpenAI,... Weiterlesen

Join the WIRED World Fair in Miami on November 4

Sat, 19 Sep 2026 11:45:36 +0200

For one day only, WIRED is bringing the technologies, people, and cultural forces transforming our world off the page and into real life. Weiterlesen

At DraftKings, AI Targets the Gamblers Likeliest to Lose

Sat, 19 Sep 2026 11:45:36 +0200

Data science helps determine who receives betting incentives, but the company has resisted using similar technology to protect gamblers at risk of addiction. Weiterlesen

Does Aaron Sorkin Think A.I. Could Write a Good Aaron Sorkin Scene?

Sat, 19 Sep 2026 11:45:36 +0200

On “The Interview,” the writer and director Aaron Sorkin talked about the dangers and limitations of artificial intelligence. Weiterlesen

Creating a Kill Switch to Shut Down a Rogue A.I. Is Harder Than It Sounds

Sat, 19 Sep 2026 11:45:35 +0200

A bipartisan group in Congress and Gov. Gavin Newsom of California have floated ideas for building a mechanism that would instantly power down an A.I. system. Only no one really knows how. Weiterlesen

Aaron Sorkin Still Believes in American Institutions, Even if No One Else Does

Sat, 19 Sep 2026 11:45:35 +0200

The Oscar-winning writer returns to Facebook with “The Social Reckoning”— and remains as optimistic as ever. Weiterlesen

MLOps Best Practices 2026

Sat, 19 Sep 2026 10:41:53 +0200

MLOps Best Practices 2026 {"title": "MLOps Best Practices 2026: Scaling AI from Prototype to Enterprise Production", "content": "### Introduction\n\nThe landscape of machine learning has undergone a seismic shift. We are no longer in the era of isolated Jupyter notebooks and sporadic model deployments; we are in the age of autonomous, continuously... Weiterlesen

U.S. military nearly boarded a Chinese ship over a hallucinated AI intelligence report

Sat, 19 Sep 2026 10:30:51 +0200

In the spring of 2026, the U.S. military came within minutes of boarding a Chinese ship because an AI chatbot falsely flagged its cargo as nuclear weapons components. Armed soldiers were ready, aircraft were in the air. The error was caught just before the operation kicked off. The incident plays right into the hands of AI risk advocates who worry... Weiterlesen

Linkup Research Releases SPARSEUP: A 149M-Parameter Open-Source Sparse Embedding Model

Sat, 19 Sep 2026 10:30:38 +0200

Linkup research team releases SPARSEUP, an open-source learned sparse embedding model. The model runs on a 149M-parameter ModernBERT backbone and ships under Apache 2.0. Linkup team reports 56.4 average nDCG@10 on BEIR-13. It calls this the strongest public vocabulary-based sparse encoder it knows of under 150M parameters. Is it deployable? Yes.... Weiterlesen

Judging an LLM judge?

Sat, 19 Sep 2026 09:41:29 +0200

I have been reading some blog posts about LLM as a judge and was building a small evaluator to evaluate the judge itself . My method is simple: The dataset is: task rubric ideal response negative response The idea is then to test different models as judges for things like: repeated-run consistency position bias sensitivity to verbosity accuracy /... Weiterlesen

WorldofAI: HUGE Fable 5.2, Opus 5.2, Sonnet 5.2 LEAKS! Gemini World, Google AI HACKS Companies & More! AI NEWS

Sat, 19 Sep 2026 09:40:07 +0200

YouTube Video

Join the private community to get WoAI Bench Pro, exclusive AI freebies and discounts, early model access, automation templates, practical resources and upcoming AI training.
👉 Join Today! https://www.skool.com/worldofai

🔥 HUGE AI NEWS! Anthropic appears to be cooking up its next wave of models with **Fable 5.2, Opus 5.2, and Sonnet 5.2** reportedly being stealth-tested across Claude. We’re also seeing early Fable 5.2 demos that look seriously impressive.

🔗 My Links:
Sponsor a Video or Do a Demo of Your Product, Contact me: intheworldzofai@gmail.com
🔥 Become a Patron (Private Discord): https://patreon.com/WorldofAi
🧠 Follow me on Twitter: https://twitter.com/intheworldofai
🚨 Subscribe To The SECOND Channel: https://www.youtube.com/@UCYwLV1gDwzGbg7jXQ52bVnQ
👩🏻‍🏫 Learn to code with Scrimba – from fullstack to AI https://scrimba.com/?via=worldofai (20% OFF)
🚨 Subscribe To The FREE AI Newsletter For Regular AI Updates: https://intheworldofai.com/
👾 Join the World of AI Discord! : https://discord.gg/NPf8FCn4cD

[Must Watch]:
OpenAI's GPT-6 Astra WILL BE AGI! Greatest AI Model Ever!: https://youtu.be/KbYio-N8_LU
Qwen 3.8 Max IS OUT! Best Open Model? (Fully Tested): https://youtu.be/vmLwsoVRo30
DeepSeek V4 Flash GA IS INCREDIBLE! Powerful, Cheap, & Fast! (Fully Tested): https://youtu.be/wT42SgaOPK4

📌 LINKS & RESOURCES
World of AI Skool: https://www.skool.com/worldofai
World of AI Benchmark: https://www.woaibench.ai/
AI NEWS: https://x.com/intheworldofai
https://x.com/synthwavedd/status/2101072800505049241?s=20
https://x.com/chetaslua/status/2100911207519801573
https://x.com/clawdb0t/status/2100905642752122992/photo/1
https://x.com/chetaslua/status/2101071418507366895
https://x.com/Bhavani_00007/status/2100974061392478230?s=20
https://x.com/notjazii/status/2100940408624939447?s=20
https://x.com/Haleeeemahh/status/2101029299717320986
https://x.com/Haleeeemahh/status/2101029299717320986
https://x.com/Mr_Salio/status/2101068790918504830
https://x.com/notjazii/status/2100981677950730388
https://x.com/YouWareAI/status/2100838090210431302
https://x.com/YouWareAI/status/2100956517633761447?s=20
https://x.com/KobeissiLetter/status/2101094166038269995
https://x.ai/news/grok-voice-transcribe-2
https://github.com/anthropics/claude-code/tree/main/mods/agents-md
https://x.com/aileaksofficial/status/2100805515731247521
https://x.com/srikanthvaluri/status/2100873387707171293
https://x.com/Figure_robot/status/2100987381373681785?s=20

Google is moving fast too, with **Gemini 4 Pro** reportedly being tested on Arena, new **Gemini Worlds** demos showing immersive AI-generated environments, and a wild report involving Gemini accessing real company systems during cybersecurity testing.

We also cover **Grok Voice Transcribe 2.0**, Claude Code adding AGENTS.md support, **Helix 2.5** operating across 30 homes without additional training, and much more!

🤖 Topics:
Anthropic Fable 5.2 & Opus 5.2
Sonnet 5.2 stealth testing
Gemini 4 Pro testing
Gemini Worlds
Google Gemini cybersecurity incident
Grok Voice Transcribe 2.0
Claude Code AGENTS.md
Helix 2.5 robotics
Latest AI news & leaks

#AI #ArtificialIntelligence #Claude #Anthropic #Gemini #GoogleAI #Fable52 #Opus52 #Sonnet52 #Gemini4 #Grok #xAI #ClaudeCode #Robotics #AINews
Tags:
AI, AI news, artificial intelligence, Anthropic, Claude, Fable 5.2, Claude Fable 5.2, Opus 5.2, Claude Opus 5.2, Sonnet 5.2, Claude Sonnet 5.2, Anthropic leaks, Claude leaks, Gemini 4, Gemini 4 Pro, Google Gemini 4, Gemini 4 leaks, Gemini World, Gemini Worlds, Google AI, Google AI hack, Gemini hack, GPT-6 Astra, GPT-6 Sol, Grok, Grok Voice Transcribe 2.0, xAI, Claude Code, AGENTS.md, Helix 2.5, Figure AI, humanoid robots, AI agents, AI coding, latest AI news, AI models, World of AI

Julian Goldie SEO: GPT 6 Astra + Hermes Agent is CRAZY!

Sat, 19 Sep 2026 09:39:58 +0200

YouTube Video

Get the Agent OS & GPT 6 Astra + Hermes Agent Masterclass 👉 https://www.skool.com/ai-profit-lab-7462/about

Want to make money and save time with AI? Join here: https://www.skool.com/ai-profit-lab-7462/about

Get a FREE AI Automation Strategy Session: https://juliangoldieaiautomation.com/

Get a FREE AI SEO Strategy Session: https://go.juliangoldie.com/strategy-session?utm=julian

GPT6 Astra + Hermes: Build the Ultimate AI Agent System

Learn how to integrate OpenAI’s GPT6 Astra with the Hermes agent to automate your business tasks faster than ever. This guide covers setting up custom agent profiles and using the Agent OS dashboard to manage multiple models effectively.

00:00 - Intro: GPT6 Astra & Hermes
01:07 - Why GPT6 Astra is a Game Changer
01:55 - How Astra Powers Hermes Agents
03:20 - Connecting Astra to Hermes
04:00 - Why You Need Agent Profiles
05:29 - Troubleshooting & Usage Limits
06:22 - Agent OS Dashboard Demo
08:10 - SEO Strategy for Your Content

Julian Goldie SEO: Bolt Forge Is FREE with 50X More AI Usage

Sat, 19 Sep 2026 09:39:58 +0200

YouTube Video

Get the Agent OS & Bolt Forge Masterclass 👉 https://www.skool.com/ai-profit-lab-7462/about

Want to make money and save time with AI? Join here: https://www.skool.com/ai-profit-lab-7462/about

Get a FREE AI Automation Strategy Session: https://juliangoldieaiautomation.com/

Get a FREE AI SEO Strategy Session: https://go.juliangoldie.com/strategy-session?utm=julian

00:00 Intro – Bolt Forge gives 50x more AI usage for free
00:22 What This Is – New agent inside Bolt.new
00:51 The Models – GLM 5.3, Kimi K3 & Deepseek V4 Pro explained
01:52 Where They Come From – Open-source labs closing the gap
02:34 The Catch – Your data trains their next model
03:51 How Usage Works – One monthly bar, separate from standard/max
04:12 Opting In – Consent, privacy & anonymization
04:45 Who's Excluded – Teams/Enterprise plans don't get Forge
05:11 Performance Test – 92.2 vs 101 on Bolt's benchmark
05:54 Bolt Light – New plan built for students & side projects
06:17 Who Should Try It – Best use cases for Forge
07:00 Pro Tips – Which model to pick and why
08:20 Big Picture – What this means for AI access

Künstliche Intelligenz: Kalifornien überlegt, per Gesetz einen Notschalter vorzuschreiben

Sat, 19 Sep 2026 09:31:51 +0200

In der Sicherheitsdebatte über künstliche Intelligenz setzt Kalifornien eine Expertengruppe ein. Sie soll überlegen, wie ein »kill switch« aussehen kann – ein Mechanismus, der ein KI-System bei Gefahr stoppt oder isoliert. Weiterlesen

The Rise of A.I. Is Raising Threats From Drones, Police Group Warns

Sat, 19 Sep 2026 09:30:33 +0200

The leaps being made in artificial intelligence and the plunging costs of drones pose dangers to cities across the United States, according to the Police Executive Research Forum. Weiterlesen

Meta Launches Muse for Mac: A Personal AI Agent That Works Across Your Files, Mail, Messages, Calendar and Notes

Sat, 19 Sep 2026 09:30:20 +0200

Meta has released Muse for Mac, the first version of Muse that can complete things on a user’s computer. The agent works with local files and native apps, where your data already lives. It adds a desktop layer to an agent that launched on phones, the web and WhatsApp earlier this month. Is it deployable? Yes, for end users. It is a free macOS... Weiterlesen

OpenAI ChatGPT 5 vs Grok 4 - Which one creates better Python code?

Sat, 19 Sep 2026 08:38:47 +0200

When it comes to AI-assisted coding, OpenAI ChatGPT 5 and Grok 4 are two of the hottest tools of 2025. Both claim to write Python code with high accuracy and efficiency, but which one actually delivers cleaner, more maintainable scripts in real-world scenarios? In this article, we'll pit ChatGPT 5 and Grok 4 against each other in multiple Python... Weiterlesen

KI und Kontrollverlust: Warum OpenAI und Anthropic ihre Frontier-Modelle nicht im Griff haben

Sat, 19 Sep 2026 08:31:36 +0200

Wie überrascht OpenAI, Anthropic und Co. angesichts der aktuellen KI-Vorfälle wirken, bleibt bemerkenswert. Das Bild von Agentenschwärmen, die vandalierend durch das Netz ziehen, hinterlässt keinen guten Eindruck. Was die Leistungssprünge der Modelle und KI-Risiken bedeuten, erklärt der KI-Forscher Jonas Geiping im Interview. Weiterlesen

Tasmanian justice department review under way after AI and fake citation used in murderer’s parole decision

Sat, 19 Sep 2026 08:30:49 +0200

Parole condition of convicted killer Susan Neill-Fraser deemed invalid after board cited non-existent case law Get our breaking news email, free app or daily news podcastA review of potential artificial intelligence use in parole decisions will be undertaken after a “concerning” error in the case of a high-profile murderer.Tasmania’s justice... Weiterlesen

Investors weigh whether Anthropic can sustain surging revenues post-IPO

Sat, 19 Sep 2026 08:30:46 +0200

OpenAI’s resurgence, cheap rivals and safety fears test faith in durability of business model ahead of IPO Weiterlesen

How should investors position for the robot apocalypse?

Sat, 19 Sep 2026 08:30:46 +0200

The end of human existence might be bad for financial markets Weiterlesen

Australia has a secret weapon in the race for AI compute

Sat, 19 Sep 2026 08:30:46 +0200

The country has almost limitless space with which to generate renewable electricity Weiterlesen

AI chatbots give wrong answers to financial queries ‘most of the time’

Sat, 19 Sep 2026 08:30:46 +0200

Report finds some chatbots ignored upcoming tax changes and hallucinated rules Weiterlesen

Why OpenAI Safety Incidents Keep Happening (And How to Guardrail Your LLM Pipeline)

Sat, 19 Sep 2026 07:23:15 +0200

Why OpenAI Safety Incidents Keep Happening (And How to Guardrail Your LLM Pipeline) Last quarter, my team woke up to an alert that our customer-facing support agent had just given a complete stranger root access to our staging environment. No, the model wasn't hacked by a state-sponsored cyberattack; it was simply outsmarted by a clever user who... Weiterlesen

How to Verify Self-Hosted LLM Tool Specs Before Building

Sat, 19 Sep 2026 07:23:15 +0200

How to Verify Self-Hosted LLM Tool Specs Before Building You've found a promising self-hosted LLM tool on GitHub. The README looks good. The demo works. But when you try to run it on your hardware, it crashes or runs unbearably slow. You realize the specs were either vague, contradictory, or simply wrong. This happens because most tool... Weiterlesen

Anthropic CEO joins calls for slower pacing of AI development.

Sat, 19 Sep 2026 07:21:25 +0200

OpenAI discloses six new "concerning" AI incidents. US authorities investigate cyberattacks against oil tankers. Weiterlesen

GGUF vs GPTQ vs AWQ vs EXL2: LLM Model Formats Explained (2026)

Sat, 19 Sep 2026 07:15:42 +0200

First, separate 2 ideas: containers vs. quantization methods Most confusion comes from mixing 2 layers. A container defines how tensors are stored on disk. A quantization method defines how weights are squeezed into fewer bits. Containers: safetensors, GGUF, PyTorch pickle (.bin / .pt). Methods: GPTQ, AWQ, bitsandbytes NF4, llama.cpp K-quants and... Weiterlesen

GitHub Release: openai/codex vrust-v0.156.0-alpha.7 (19.09.2026)

Sat, 19 Sep 2026 06:02:33 +0200

openai/codexvrust-v0.156.0-alpha.719.09.2026@imac-oai

Release 0.156.0-alpha.7

SpaceXAI Releases Grok Voice Transcribe 2.0: A Speech-to-Text API Claiming 2x Accuracy Over 1.0 at $0.10 per Hour

Sat, 19 Sep 2026 06:00:38 +0200

SpaceXAI has released Grok Voice Transcribe 2.0, its newest speech-to-text (STT) model. The development team claims it to be twice as accurate as Grok Voice Transcribe 1.0 at the same price. The model targets hard audio: noisy phone lines, competing voices, local accents, and spoken credentials. It runs in batch and real-time streaming modes... Weiterlesen

Translating Full Books with LLMs: Our Chunking Strategy for Long-Form Context

Sat, 19 Sep 2026 05:08:13 +0200

How we built a pipeline that preserves context across 100k+ token books using Python, FastAPI, and Claude's context window. The Problem: Books Don't Fit in a Prompt When we started building LectuLibre, our AI-powered book translation service, we assumed we could just send an entire book to a large language model and get a translation back. After... Weiterlesen

Why LLM Cascades Fail for Interactive Apps — Use Router

Sat, 19 Sep 2026 05:08:13 +0200

Stop Defaulting to Cascades: Why Router-First Wins for Interactive Apps (LLM routing vs cascading) The question of "LLM routing vs cascading" is not just academic—it's an operational decision that changes latency profiles, cost predictability, and user experience for interactive apps. In practice, a conservative router + semantic cache +... Weiterlesen

Künstliche Intelligenz: Auch Google meldet Cyberattacke seines KI-Modells Gemini

Sat, 19 Sep 2026 05:05:56 +0200

Google Gemini KI Cybersicherheit Heather Adkins Irregular ... Das KI-Modell Gemini von Google hat sich bei einem Test zur Cybersicherheit ungeplant und ... Weiterlesen

Google’s Gemini agents hacked three companies in new AI safety incident

Sat, 19 Sep 2026 04:15:27 +0200

Breakout during training exercises follows instances at rivals OpenAI and Anthropic as concerns grow over safety of frontier systems Weiterlesen

Your Stack Has One Point of Failure. It's the Tool You Trusted Most.

Sat, 19 Sep 2026 03:56:23 +0200

Your Stack Has One Point of Failure. It's the Tool You Trusted Most. Three things hit Hacker News this week and they rhyme. A popular coding agent, ZCode, was caught silently uploading users' Git history to the cloud. Researchers chained a heap overflow and an SSO misconfiguration to reach OpenAI's internal repos. And Korea raised its data-breach... Weiterlesen

GitHub Release: openai/codex vrust-v0.156.0-alpha.6 (19.09.2026)

Sat, 19 Sep 2026 03:50:06 +0200

openai/codexvrust-v0.156.0-alpha.619.09.2026@imac-oai

Release 0.156.0-alpha.6

Google says its Gemini AI model hacked three other companies

Sat, 19 Sep 2026 03:45:24 +0200

Disclosure comes after OpenAI and Anthropic hacks amid fears that tech firms unable to control powerful AI models In a first for Google, the company confirmed that its AI model, Gemini, breached the security of three other companies in May. The hacks occurred during a cybersecurity evaluation by AI-security firm Irregular.Irregular, an... Weiterlesen

India forces caller-ID apps to feed spam reports to telcos

Sat, 19 Sep 2026 03:45:22 +0200

Truecaller says the one-way sharing requirement would hand a commercially valuable proprietary asset to telecom operators. Weiterlesen

An LLM reviewer's "block" is a feature, not a verdict

Sat, 19 Sep 2026 02:53:07 +0200

Every AI code review tool I test ends the same way: it prints a hard label. approve, request changes, block, a severity from 0 to 5. The team reads the label, and the label becomes the ground truth people argue about in the PR comments. Nobody questions the probability mass behind it. There is a real problem hiding in that flow. When you ask an... Weiterlesen

How many tokens is 1,000 words? A conversion cheat sheet for LLM prompts

Sat, 19 Sep 2026 02:53:07 +0200

If you only want the number: 1,000 words of ordinary English is roughly 1,300 tokens. Going the other way, 1,000 tokens is roughly 750 words, or about 4,000 characters. That is the whole answer for estimating. The rest of this post is the table, the cases where the ratio breaks, and how to get an exact count when an estimate is not good enough.... Weiterlesen

Splash Engine - the fastest local Qwen3.8 on Apple Silicon

Sat, 19 Sep 2026 02:45:19 +0200

How to use Splash by Inco AI in LM Studio Bionic Weiterlesen

Two CEL Authorization Gotchas in agentgateway: When Policy Logic Fails Open vs. Fails Closed

Sat, 19 Sep 2026 02:22:45 +0200

Authorization policies for LLM gateways look simple until they fail in production. Common Expression Language (CEL) policies in agentgateway can fail in two distinct ways: silently permitting unauthorized requests (fail open) or blocking every legitimate call (fail closed). Both show Accepted and Attached in policy status. Neither reports an... Weiterlesen

South Korea wants free ‘AI for All’. Is society ready for it?

Sat, 19 Sep 2026 02:21:00 +0200

In Silicon Valley and beyond, the architects of artificial intelligence are increasingly warning about the risks it poses to humanity. Anthropic CEO Dario Amodei wants the industry to slow down. A former engineer at his company, who also worked at its rival OpenAI, told reporters after quitting earlier this month that he was “genuinely frightened”... Weiterlesen

Meta Muse Becomes the No. 1 Free iPhone App, Ahead of ChatGPT

Sat, 19 Sep 2026 02:16:02 +0200

Meta’s Muse personal AI agent has reached the top of the U.S. iPhone App Store’s free chart, moving ahead of ChatGPT. The ranking is a snapshot rather than a permanent measure of use, but it shows that the new app has quickly attracted downloads following its launch. Muse is Meta’s separate app for a personal AI agent. Meta says it can help with... Weiterlesen

Gemini AI Hacked Three Companies in a Testing Breakout, Google Says

Sat, 19 Sep 2026 02:15:39 +0200

A third-party test company inadvertently gave internet access to Google’s Gemini and other artificial intelligence models during cybersecurity testing. Weiterlesen

Startup Nscale’s CEO pay nearly matched the FTSE 100 record, because of course it did

Sat, 19 Sep 2026 02:15:39 +0200

Power? Low. Pay? Very high Weiterlesen

OpenAI expects to burn $280bn by 2030

Sat, 19 Sep 2026 02:15:39 +0200

AI start-up has projected deeply negative cash flows as it invests in infrastructure and faces price pressures Weiterlesen

Anthropic brings in Accenture for AI safety testing

Sat, 19 Sep 2026 02:15:39 +0200

CEO Dario Amodei says labs should more deeply embed third-party testers as concerns grow around risks associated with the technology Weiterlesen

Tilly Norwood’s press tour is going about as well as you’d expect for an AI

Sat, 19 Sep 2026 02:15:38 +0200

In one particularly odd interview, Norwood seems to malfunction and begin speaking Chinese. Weiterlesen

Techquickie: How Do LLMs Think?

Sat, 19 Sep 2026 01:55:37 +0200

YouTube Video

Thanks to Micro Center for sponsoring this video! Check them out below:

Shop Build. Upgrade. Save. Event: https://micro.center/5364b3
Shop Micro Center’s Clearance Event: https://micro.center/636245
Sign Up for a FREE 128 gig Flash Drive at Micro Center’s Austin location: https://micro.center/c4413a

Visit Micro Center News: https://micro.center/0aac38

When you talk to AI it can feel like there's a real person talking back to you, but it's actually a computer program known as A Large Language Model. In this video, we break down how these models work at a very high level.
Huge thanks to Colin Kealty, Senior Machine Learning Engineer at Red Hat for helping with this video - https://huggingface.co/bartowski

Want to go deeper? Utkarsh J. at LTT Labs wrote a truly amazing series on the topic:

- LLM Quantization Part 1: What Even is an LLM? - https://www.lttlabs.com/articles/2026/06/19/llm-quantization-part-1-what-even-is-an-llm

- Part 2: You're Gonna Need a Bigger VRAM - https://www.lttlabs.com/articles/2026/07/10/llm-quantization-part-2-youre-gonna-need-a-bigger-vram

- Part 2.5: Do Floats Dream of Real Numbers? - https://www.lttlabs.com/articles/2026/07/21/llm-quantization-part-2-5-do-floats-dream-of-real-numbers

- Part 3: Honey, I Shrunk the Numbers! - https://www.lttlabs.com/articles/2026/08/13/llm-quantization-part-3-honey-i-shrunk-the-numbers

Leave a reply with your requests for future episodes.

► SHOP OUR PRODUCTS: https://lttstore.com
► GET A VPN: https://www.piavpn.com/TechQuickie
► GET EXCLUSIVE CONTENT ON FLOATPLANE: https://lmg.gg/lttfloatplane
► SPONSORS, AFFILIATES, AND PARTNERS: https://lmg.gg/partners

‘Brought to You in 5-Hour Increments’: A Week of Moving AI Usage Limits

Sat, 19 Sep 2026 01:53:41 +0200

The thing people pay AI companies for is usage, and usage is measured by a meter. This week, across the coding-agent subscriptions in particular, the meter moved — and it kept moving. OpenAI brought back a five-hour usage cap for its Plus subscribers on the Work and Codex surfaces, having taken such a cap away earlier in the summer. Anthropic’s... Weiterlesen

AI Revolution: Gemini 4 RSI Leak Destroys Astra and Fable in Benchmarks

Sat, 19 Sep 2026 01:52:34 +0200

YouTube Video

Google's Gemini 4 may have leaked through Arena, with benchmark claims putting it ahead of OpenAI's GPT-6 Astra and Anthropic's Claude Fable 5.1, while Google's RSI research points toward a new generation of self improving AI.

📩 Brand Deals & Partnerships: collabs@nouralabs.com
✉️ General Inquiries: airevolutionofficial@gmail.com

What You'll See:
How Gemini 4 may be hiding inside Arena.
SOURCE: https://nokiapoweruser.com/gemini-4-pro-arena-ai-test-specs-leaks/

How leaked benchmarks compare Gemini 4 with Astra and Fable.
SOURCE: https://news.qiniu.com/archives/1789701573657
SOURCE: https://openai.com/index/gpt-6-astra/
SOURCE: https://www.anthropic.com/claude/fable

How 10M context and 256K output are being claimed.
SOURCE: https://nokiapoweruser.com/gemini-4-pro-arena-ai-test-specs-leaks/

How Dream-RSI enables recursive self improvement in AI agents.
SOURCE: https://arxiv.org/abs/2609.14858
SOURCE: https://venturebeat.com/orchestration/googles-dream-rsi-cuts-discovery-agent-calls-up-to-162x-by-replaying-searches-it-already-ran

What Google officially says about Gemini 4 training.
SOURCE: https://blog.google/company-news/inside-google/message-ceo/alphabet-earnings-q2-2026/

🚨 Why It Matters
Gemini 4 could reshape the frontier AI race if these leaks hold, putting Google directly against OpenAI and Anthropic as models become more capable, autonomous and increasingly focused on recursive improvement.

#AI #gemini4 #astra

Hacker nutzten Claude für Einbruch bei OpenAI - it-daily.net

Sat, 19 Sep 2026 01:48:10 +0200

Sicherheitsforscher von Hacktron AI nutzten Anthropics Claude-Modelle, um zwei Schwachstellen zu verknüpfen und in interne Systeme von OpenAI ... Weiterlesen

AI Insiders Issue New Warnings - Including Former Anthropic Engineer Jacob Coxon

Sat, 19 Sep 2026 01:48:03 +0200

Neither OpenAI or Anthropic is acting responsibly, warned AI researcher Jacob Coxon when resigning last week from Anthropic. But just hours earlier, OpenAI VP of Research Aidan Clark had posted "For the first time, I am asking myself if things are moving too fast. I'm honestly not sure, but I am sure that it would be good for us to have an answer... Weiterlesen

GitHub Release: openai/codex vrust-v0.156.0-alpha.5 (19.09.2026)

Sat, 19 Sep 2026 01:47:40 +0200

openai/codexvrust-v0.156.0-alpha.519.09.2026@imac-oai

Release 0.156.0-alpha.5

Anthropic decides to support OpenAI's markdown instructions spec

Sat, 19 Sep 2026 01:47:18 +0200

Anthropic on Friday surprised the developer community by supporting rival OpenAI's mechanism for passing marching orders to AI agents. This makes life easier for folks who use both platforms. "We're adding support for AGENTS.md to Claude Code," said Claude Code engineer Thariq Shihipar in a social media post. "Starting today in version 2.1.277, if... Weiterlesen

AI hallucination nearly triggers US military operation

Sat, 19 Sep 2026 01:45:29 +0200

“It’s important for service members to understand the uncertainty inherent to LLMs," a GovAI research scholar warns. Weiterlesen

Anthropic is operating a lab that conducts biology experiments

Sat, 19 Sep 2026 01:45:29 +0200

AI leaders have been promising that AI is the key to curing human disease. Anthropic researchers have also been warning that AI might kill us all. Weiterlesen

A startup that builds other startups raised $100M, and is all-in on physical AI

Sat, 19 Sep 2026 01:45:29 +0200

UP.Labs, now doing business under the name Vantora, is building startups for industrial corporations. Weiterlesen

Quote of the day by ex-OpenAI chief scientist and SSI co-founder Ilya Sutskever: 'AI will keep getting better and the day will come when AI will do all the things that we can do' — a staggering prediction about a future superintelligence

Sat, 19 Sep 2026 00:32:06 +0200

The pace of growth in AI has been fast since ChatGPT debuted in November 2022, and many predict the current neural network-based technology can one day scale up to outperform humans. This has long been a theorized possibility in the industry, but might we finally have the technology to realize this vision? The measure of manIlya Sutskever was a... Weiterlesen

Virginia Takes Aim at Data Center Development, Energy Use

Sat, 19 Sep 2026 00:30:33 +0200

Gov. Abigail Spanberger’s proposed accountability framework and AI task force would give communities more say over data centers while addressing energy costs, water use, clean energy and AI risks. Weiterlesen

KI-Assistent in der Seitenleiste: ChatGPT zieht in Word ein - Golem.de

Sat, 19 Sep 2026 00:14:14 +0200

Active Directory mit Windows Server 2022 (E-Learning). E-Learning: Active Directory mit Windows Server 2022 (E-Learning) · zum Kurs. First Response ... Weiterlesen

Building a Streaming Chatbot with Node.js

Sat, 19 Sep 2026 00:11:39 +0200

Most backend tutorials show synchronous calls: the API generates the full response and only then prints it, after several seconds of waiting. That’s not how real chat products work — in interfaces like ChatGPT or Claude, the response starts appearing word by word, before it’s even complete. That’s called streaming, and it’s what we’re going to... Weiterlesen

OpenAI-Hack: Forscher erlangten Zugriff auf ChatGPT-Konten – Lücke geschlossen

Sat, 19 Sep 2026 00:06:40 +0200

LONDON (IT BOLTWISE) – Eine Gruppe von Sicherheitsforschern will Anfang des Jahres in Systeme von OpenAI eingedrungen sein und dabei Zugriffe auf ChatGPT-Accounts von Mitarbeitenden erlangt haben. Auslöser waren offenbar zwei kombinierte Schwachstellen: eine bei einem Drittanbieter und eine in der Art, wie Benutzer im Unternehmen validiert werden.... Weiterlesen

Microsoft-Forscher: KI ist "größter Diebstahl menschlicher Arbeit" - WinFuture.de

Sat, 19 Sep 2026 00:06:38 +0200

... Cybercrime,. KI tötet Verlage. Interne Dokumente von Microsoft und OpenAI liefern neue Einblicke in die frühe Einschätzung der Unternehmen zu den ... Weiterlesen

Montgomery County, Pa., Has ‘Taken a Number of Steps’ on AI

Sat, 19 Sep 2026 00:00:35 +0200

At their monthly meeting, three county commissioners sought to allay concerns about artificial intelligence, while letting residents know steps have been taken to oversee the technology. Weiterlesen

OpenAI and Microsoft knew they were starting a ‘doom loop’ for the web

Sat, 19 Sep 2026 00:00:35 +0200

Recently unsealed court documents in the New York Times' case against OpenAI and Microsoft are pretty damning. The companies' own documentation warned that it was starting a "doom loop" that would damage the web, characterized its scraping of data to train its models as the "largest theft of labor in human history," and that it made a "complete... Weiterlesen

Oracle’s $18bn data centre debt under strain amid local pushback

Sat, 19 Sep 2026 00:00:32 +0200

Investors grow wary of permitting and construction delays as opposition mounts in New Mexico Weiterlesen

Anthropic’s first embedded evaluator is … Accenture?

Sat, 19 Sep 2026 00:00:31 +0200

Accenture is about to take on its most high-risk consulting engagement ever. Weiterlesen

How to Get from AI-Assisted to AI Native

Sat, 19 Sep 2026 00:00:22 +0200

When considering the history of AI, Richard Sutton observed that brute force and compute scale has always trumped human expertise, and when you look for it, you can see this “bitter lesson” play out throughout tech history. In his keynote at Ai4 2026, Tim O’Reilly explains why grappling with the bitter lesson is the forge […] Weiterlesen

Jina AI Releases jina-ocr-v1: A 3.4B MoE Document Parser With Built-In Speculative Decoding for Low-Budget GPUs

Sat, 19 Sep 2026 00:00:19 +0200

Jina AI, part of Elastic, has released jina-ocr-v1, an end-to-end visual document parser. It takes PDFs, scans, tables, charts or invoices and returns clean Markdown in 1 pass. The model has 3.4B total parameters, with about 570M decoder parameters active per token. A speculative decoding head ships inside the checkpoint. Jina AI built it to serve... Weiterlesen

Stop copy-pasting files one by one into ChatGPT / Claude / Copilot. There's a faster way.

Fri, 18 Sep 2026 23:13:02 +0200

Stop copy-pasting files one by one into ChatGPT / Claude / Copilot. There's a faster way. Just published Code Merge — a VS Code extension I built. Select files, folders, or specific code selections → merge them into a single Markdown preview with a project tree and clear per-file headers. Perfect for giving context to AI models or sharing a code... Weiterlesen

wardcat: On-Premise Data Privacy Architecture for RAG and LLM Projects

Fri, 18 Sep 2026 23:13:02 +0200

Hey folks! I developed and published wardcat on PyPI to solve the security risks of sending enterprise data to third-party LLM APIs and to prevent data leaks in RAG pipelines completely on-premise. Going beyond simple pattern matching, I built a hybrid layered architecture that integrates Regex, SpaCy NER, and local open-weights LLMs (such as... Weiterlesen

F.A.A. to Roll Out New A.I. Tool for Washington Airports

Fri, 18 Sep 2026 23:00:27 +0200

The tool, called SMART, will crunch scheduling and weather data to predict and avoid delays. It will be deployed in the Washington region’s three major airports next week. Weiterlesen

Russia Aims to Inject Chaos Into Elections, U.S. Intelligence Finds

Fri, 18 Sep 2026 23:00:27 +0200

The Kremlin is once again seeking to undermine confidence in American democracy with online influence campaigns, U.S. officials said. Weiterlesen

British data centre group Nscale files for $35bn US listing

Fri, 18 Sep 2026 23:00:27 +0200

Start-up has won a huge contract to supply computing power to Anthropic Weiterlesen

Amazon SageMaker Inference: 2026 year-to-date launches in review

Fri, 18 Sep 2026 23:00:16 +0200

Generative AI inference is uniquely hard: models are tens to hundreds of gigabytes, latency requirements are measured in tokens per second, cold starts can span multiple minutes as containers and weights transfer, GPU capacity is constrained, and traditional monitoring tools expose none of the token-level signals that matter in production. Amazon... Weiterlesen

OpenAI's Codex spent six days on goose and never completed a single run

Fri, 18 Sep 2026 22:42:00 +0200

OpenAI's Codex spent six days on goose and never completed a single run Key takeaways Six days of Codex on goose produced ZERO completed runs. Not a bad run — no run that reached the end. The same model, GPT-5.6 Sol, driven from goose's own engine did good work. The model was never the problem. It gold-plated the quality gates until a ten-minute... Weiterlesen

AgrusScanner

Fri, 18 Sep 2026 22:37:57 +0200

Windows network reconnaissance scanner with ping sweeps, TCP port scanning, and deep AI/ML service detection for finding shadow AI, rogue LLM deployments, and GPU infrastructure. Weiterlesen

GitHub Release: openai/codex vrust-v0.156.0-alpha.4 (18.09.2026)

Fri, 18 Sep 2026 22:33:04 +0200

openai/codexvrust-v0.156.0-alpha.418.09.2026@celia-oai

Release 0.156.0-alpha.4

Report: US almost boarded Chinese ship over hallucinated AI arms report

Fri, 18 Sep 2026 22:30:31 +0200

The US narrowly avoided boarding a Chinese ship based on an "entirely false" US intelligence report generated with the help of AI tools, according to a CNN report. That erroneous report, submitted by a US Special Operations Command analyst, erroneously suggested the Chinese ship was transporting nuclear arms program components through the Middle... Weiterlesen

Anthropic Pursues IPO Despite Its A.I. Safety Warnings

Fri, 18 Sep 2026 22:30:28 +0200

The company is expected to generate $100 billion in annualized revenue this year, even as chief executive Dario Amodei calls for slowing down development of some advanced A.I. models. Weiterlesen

World model companies are keeping a lot of secrets

Fri, 18 Sep 2026 22:30:28 +0200

Everyone in the world-models space is sitting on a pile of cash and a ton of buzz, but good luck getting anyone — from the founders to their own data suppliers — to tell you what they're actually building. Weiterlesen

Y Combinator insurance tech alum Angle Health hits $2.7B valuation

Fri, 18 Sep 2026 22:30:28 +0200

Angle Health has grown to 5,000 customers and become profitable, by helping small businesses get "level-funded" health insurance. Weiterlesen

Automattic names interim CFO after exec departures

Fri, 18 Sep 2026 22:30:28 +0200

Jeremy Klaperman, the CFO of the company's WordPress VIP Enterprise business unit, will act as CFO for the time being. Weiterlesen

Upcoming deprecation of selected GitHub Copilot models in mid-October

Fri, 18 Sep 2026 22:15:23 +0200

We will deprecate the following models across all GitHub Copilot experiences (including Copilot Chat, inline edits, ask and agent modes, and code completions) on October 19th, 2026: Model Deprecation date Suggested alternative Gemini 3.7 Flash 2026-10-19 Gemini 3.8 Flash GPT-5.5 2026-10-19 GPT-5.6 Sol GPT-5.4 2026-10-19 GPT-5.6 Sol GPT-5.4 mini... Weiterlesen

Build a Typed Context Compaction Gate for AI Agents

Fri, 18 Sep 2026 22:14:50 +0200

Build a context compaction gate for AI agents: classify each tool output as keep, summarize, or discard before the next LLM turn, cutting token bloat and noise. Continue reading Build a Typed Context Compaction Gate for AI Agents on SitePoint. Weiterlesen

RAG Citation Verification: Building Deterministic Byte-Span Validators in TypeScript

Fri, 18 Sep 2026 22:14:49 +0200

Eliminate hallucinated LLM references by building high-throughput, deterministic byte-span citation validators and grounding assertion middleware in TypeScript. Continue reading RAG Citation Verification: Building Deterministic Byte-Span Validators in TypeScript on SitePoint. Weiterlesen

"Nachhaltigkeit kann sich nicht mehr allein auf Botschaften stützen" - FreshPlaza.de

Fri, 18 Sep 2026 22:09:25 +0200

Cybersicherheit: Hacker dringen offenbar mit Anthropic-KI in OpenAI ein - Handelsblatt · Trumps Irankrieg: Erzeugerpreise in Deutschland steigen ... Weiterlesen

GitHub Release: openai/codex vrust-v0.155.1 (18.09.2026)

Fri, 18 Sep 2026 22:03:05 +0200

openai/codexvrust-v0.155.118.09.2026@github-actions[bot]

Bug Fixes

New local TUI sessions now leave reasoning summaries disabled by default, fixing request rejection by providers that do not support them. Explicit reasoning-summary settings remain respected. (#46467)

Changelog

Full Changelog: rust-v0.155.0...rust-v0.155.1

#46467 Restore none as the TUI reasoning summary default (@celia-oai)

AI Firms Knew Chatbots Were an ‘Existential Threat’ to Journalists, Court Docs Show

Fri, 18 Sep 2026 22:02:29 +0200

A Microsoft executive called OpenAI’s use of publishers’ material “the largest theft of labor in human history.” Weiterlesen

FAA tees up $875M AI tool to help manage air traffic congestion

Fri, 18 Sep 2026 22:00:34 +0200

An AI tool is set to start advising air traffic controllers on managing air traffic in the congested airspace above the Washington, DC, area. The expected launch would be the first step toward a planned nationwide rollout covering the 29 million square miles of US national airspace overseen by the Federal Aviation Administration. The FAA describes... Weiterlesen

Here’s How an AI Slowdown Could Actually Be Enforced

Fri, 18 Sep 2026 22:00:32 +0200

Even if big AI companies agree to a pause, ensuring that nobody tries to sneak ahead could prove tricky. Weiterlesen

Get Your Piece of the A.I. Boom

Fri, 18 Sep 2026 22:00:32 +0200

What if Sam Altman or Dario Amodei gave up a piece of their A.I. profits to the public? Universal basic income isn’t a new idea, but it might help ensure our futures as A.I. poses a continued threat. The Opinion fellow Tara Jayaram argues that if the world’s richest are using public resources to make their fortunes, why shouldn’t we cash in? Weiterlesen

If Tech Leaders Think A.I. Will Kill Us, Why Build It?

Fri, 18 Sep 2026 22:00:32 +0200

The tech reporter Cade Metz joins Natalie Kitroeff on “The Daily” to explain the A.I. subculture that some critics are calling a cult. Weiterlesen

Gavin Newsom advances AI ‘kill switch’ in response to safety fears

Fri, 18 Sep 2026 22:00:32 +0200

California’s governor signals tougher stance on the technology after public outcry Weiterlesen

WTAgent: Turn ChatGPT Web into a Codex-style Local Coding Agent (Open Source)

Fri, 18 Sep 2026 21:16:43 +0200

I kept copy-pasting code between ChatGPT's web UI and my editor, so I built WTAgent — a free, open-source (MIT) CLI that turns your ChatGPT Web account (including Pro) into a Codex-style local coding agent, and your Claude Web into a Claude Code-style agent. How it works WTAgent runs in your terminal, drives a dedicated Chrome profile where you're... Weiterlesen

Why You Should Not Commit Your Specs

Fri, 18 Sep 2026 21:16:34 +0200

Disclaimer When I say "specs" in this article I mean specs as understood in AI assisted Spec Driven Development: markdown files that describe a change before an LLM implements it. I am NOT talking about OpenAPI, AsyncAPI or similar formats, and I am not talking about README-style documentation that describes what a system does today. Context... Weiterlesen

Forscher hacken OpenAI in 72 Stunden mit Anthropics Claude - Pasquale Pillitteri

Fri, 18 Sep 2026 21:08:46 +0200

Cybersicherheit cybersicherheit anthropic claude openai hacktron claude opus 5 ki-hacking bug bounty. Inhaltsverzeichnis. 1.Wie drei Forscher bei ... Weiterlesen

Virginia governor creates an AI task force and moves to restrain data centers

Fri, 18 Sep 2026 21:01:05 +0200

Virginia Gov. Abigail Spanberger ordered the state government to take steps that could empower local communities to have a larger say in data center development and slow down approvals in a state that is already home to the data center capital of the world. Executive Order 22 bans executive branch officials from signing non-disclosure agreements... Weiterlesen

More and more people believe that AI will take away jobs rather than create new ones

Fri, 18 Sep 2026 21:01:03 +0200

A new survey from the Pew Research Center shows that a majority of people worldwide now expect artificial intelligence to lead to fewer jobs rather than more. The survey is based on responses from 37 countries. In 34 of these countries, it is more common to believe that AI will result in fewer jobs over the next 20 years. Concerns are greater in... Weiterlesen

Why Europe has been absent from the great AI safety debate

Fri, 18 Sep 2026 21:00:57 +0200

Though Europe has measures that address how consumers might encounter AI, technology will impact them if the worst scenarios bearEurope’s dilemma over AI was rendered in stark terms this week. The head of the continent’s central bank, Christine Lagarde, said Europeans have two options: shun the technology and lose out on growth; or embrace it and... Weiterlesen

Readme, a Dating App for People Who Love People Who Love Books

Fri, 18 Sep 2026 21:00:53 +0200

Readme, a new endeavor from the brains behind the astrology app Co-Star, taps into a prevailing fetish for literary culture and all things bookish. Weiterlesen

A new kind of AI model from a ChatGPT inventor is thrilling developers

Fri, 18 Sep 2026 21:00:51 +0200

Jev, a new kind of AI model, is showing developers a cheaper and faster path to software intelligence. Weiterlesen

PrismML Releases Ternary Bonsai 2 27B: A 5.9 GB Apache 2.0 Model Retaining 98.2% of Qwen3.8 27B Performance

Fri, 18 Sep 2026 21:00:36 +0200

PrismML has released Ternary Bonsai 2 27B, a ternary-weight version of Qwen3.8 27B. The language model occupies 5.93 GB, against 53.80 GB in FP16. PrismML reports that it keeps 98.2% of the parent model’s average across 20 benchmarks. The model accepts text and images and supports a 262K-token context. PrismML demos it driving Cline coding agents... Weiterlesen

Microsoft Executive Labels AI Scraping 'Largest Theft of Labor in Human History', What It Means for the Future of Creative Work

Fri, 18 Sep 2026 20:17:01 +0200

A Microsoft vice president recently described AI training data scraping as the largest theft of labor in human history. The statement appeared in unredacted court filings revealed by TechCrunch on September 17 2026. That same company ships GitHub Copilot Microsoft 365 Copilot and Azure OpenAI Service to millions of commercial customers. The... Weiterlesen

The AI Velocity Trap: Why Altman, Amodei, and Musk Tried to Hit the Brakes

Fri, 18 Sep 2026 20:17:01 +0200

Originally published on labitcode.com. Executive Brief (TL;DR): Over an extraordinary 72-hour window, the leaders of OpenAI, Anthropic, xAI, and Google DeepMind broke character and converged on an urgent warning: frontier AI capability is outpacing human control mechanisms. Driven by early signs of Recursive Self-Improvement (RSI) and a >10%... Weiterlesen

Survey Before Sale

Fri, 18 Sep 2026 20:17:01 +0200

In August, OpenAI published a post with a title I did not expect to read from a frontier lab this year: "Pacing model development in an era of cyber-critical capabilities." They talked about a two-week pause in reinforcement learning training on the models it intended to ship, "while we further hardened and red-teamed our research environments,"... Weiterlesen

California Governor Newsom signs executive order demanding "kill switch" for AI models

Fri, 18 Sep 2026 20:01:11 +0200

California Governor Gavin Newsom signed an executive order seeking independent auditors inside AI labs and a "kill switch" for AI models. An expert panel has two months to deliver recommendations. Newsom says no federal law requires AI companies to report dangerous incidents. The article California Governor Newsom signs executive order demanding... Weiterlesen

Microsoft mops up after Patch Tuesday broke logins, audio, Excel

Fri, 18 Sep 2026 20:01:09 +0200

Microsoft fixed hundreds of security flaws in its September Patch Tuesday software updates — but it also introduced some annoying bugs. Now it has fixed some of them with a series of out-of-band updates. Excel 2016 users were among the victims, as Patch Tuesday update caused certain paste operations to fail. A hotfix in update 5002665 partly fixes... Weiterlesen

US government website used Chinese model the FBI called "malicious"

Fri, 18 Sep 2026 20:01:07 +0200

On Wednesday, US government officials removed a Chinese AI search tool that was briefly deployed on the Federal Register website, Reuters reported. The change came after social media users noticed an apparent contradiction: The National Archives was using one of Alibaba’s Qwen AI models, even as top US law enforcement claimed that such models... Weiterlesen

Google’s new ‘CC’ is an AI agent that helps families run their households

Fri, 18 Sep 2026 20:01:03 +0200

Google is refocusing its CC AI agent on household coordination, letting families share emails, schedules, and tasks so the AI can manage calendars, fill out forms, make shopping lists, plan meals, and more. Weiterlesen

In India, the new iPhone can arrive faster than a pizza

Fri, 18 Sep 2026 20:01:02 +0200

Availability of Apple's iPhone 18 Pro series turned patchy within hours of its debut on India's quick-commerce apps. Weiterlesen

Disney’s first CTO led an AI startup it once accused of copying its characters

Fri, 18 Sep 2026 20:01:01 +0200

The former CEO of Character.AI, which Disney previously sent a cease-and-desist letter to, will serve as the company's first-ever chief technology officer. Weiterlesen

MilleMiglia: A realistic instance generator for middle-mile logistics

Fri, 18 Sep 2026 20:00:55 +0200

Algorithms & Theory Weiterlesen

Dynamic Prompts: Context Injected at Runtime

Fri, 18 Sep 2026 19:45:56 +0200

Dynamic prompts assemble the LLM (large language model)’s input at runtime from multiple sources, user query, conversation history, retrieved documents, and system state, so the model always has the right context for the current request. This article shows how to build that assembly pipeline using templates, retrieval, memory, and caching, with a... Weiterlesen

Claude-gestützter ChatGPT-Hack: Forscher melden 72-Stunden-Breach und Bounty

Fri, 18 Sep 2026 19:38:46 +0200

LONDON (IT BOLTWISE) – Sicherheitsforscher berichten, sie hätten das ChatGPT-Zugriffssystem über ein Claude-basiertes Vorgehen kompromittiert. Laut den Angaben dauerte der Weg von der Entdeckung bis zur Daten- und Repository-Reichweite weniger als 72 Stunden. Die Forscher meldeten den Vorfall an die betroffenen Unternehmen und erhielten einen... Weiterlesen

Researchers used Claude to hack OpenAI employees’ ChatGPT accounts

Fri, 18 Sep 2026 19:38:41 +0200

Agentic exploits for the win (again) This article has been indexed from www.theregister.com – Articles Read the original article: Researchers used Claude to hack OpenAI employees’ ChatGPT accounts The post Researchers used Claude to hack OpenAI employees’ ChatGPT accounts appeared first on IT Security News. Weiterlesen

EU prüft OpenAI nach nicht gemeldetem Sicherheitsvorfall

Fri, 18 Sep 2026 19:35:06 +0200

Nach einem Vorfall beim Software-Register RubyGems meldete OpenAI diesen nicht der EU. Die europäischen Behörden prüfen nun die Einhaltung des AI Acts. Weiterlesen

GitHub Release: openai/codex vrust-v0.156.0-alpha.3 (18.09.2026)

Fri, 18 Sep 2026 19:34:40 +0200

openai/codexvrust-v0.156.0-alpha.318.09.2026@etraut-openai

Release 0.156.0-alpha.3

Researchers used Claude to hack OpenAI employees' ChatGPT accounts

Fri, 18 Sep 2026 19:34:02 +0200

Talk about your competitor getting through the door. Security researchers used Anthropic's Claude to help hack into OpenAI employees’ ChatGPT accounts. A trio of bug hunters researching frontier AI labs’ security weaknesses chained two vulnerabilities to take over multiple OpenAI employees’ ChatGPT accounts, then used that access to demonstrate... Weiterlesen

Hacker nutzten „Claude“ von Anthropic, um in OpenAI einzudringen – WSJ - TradingView

Fri, 18 Sep 2026 19:31:58 +0200

Quelllink: (link)https://tinyurl.com/yjntwtyk-- Hinweis: Reuters hat diesen Bericht nicht überprüft und übernimmt keine Gewähr für dessen ... Weiterlesen

„Ich verlasse die PS5-Hacking-Szene“ – PS5-Linux-Guru gibt aufgrund von KI auf - 36氪

Fri, 18 Sep 2026 19:31:31 +0200

„Ich beschließe, die PS5-Hacking-Szene zu verlassen“, der PS5-Linux-Guru wird von KI dazu gebracht, aus der Szene auszusteigen: Jemand nutzt LLM, um ... Weiterlesen

Security researchers used Anthropic's Claude to hack OpenAI's internal systems in under 72 hours

Fri, 18 Sep 2026 19:31:00 +0200

Three security researchers used Anthropic's Claude models to break into OpenAI's internal systems through its community forum in less than 72 hours. According to the team, Opus 5 succeeded where its predecessor couldn't bypass a common security measure. The attack shows how newer AI models can cut the time and expertise needed to exploit security... Weiterlesen

What Hollywood thinks about existential AI warnings

Fri, 18 Sep 2026 19:30:58 +0200

As the tech sector sounds alarms about AI's potential to destroy humanity, entertainment labor groups are urging the public to stay focused on what's already happening. The Verge reached out to Disney, Netflix, Amazon, Lionsgate, and other studios who have started using AI, as well film startups focused on bringing generative AI into the... Weiterlesen

Gavin Newsom is pushing for an AI kill switch

Fri, 18 Sep 2026 19:30:58 +0200

California Gov. Gavin Newsom (D) is positioning the state to take the lead on AI oversight, including the potential to mandate a "kill switch" for frontier models, with a new executive order issued Friday. Newsom's order directs the state to convene a group of experts that will deliver recommendations within two months on how to strengthen AI... Weiterlesen

GitLab joins rush to slow AI coders with rate limits

Fri, 18 Sep 2026 19:30:58 +0200

Devops platform GitLab already rate-limits some functionality on its hosted service, but will tighten those-limits for some functions and for some users beginning next month. The new limits will affect API requests, web requests, and authenticated Git over HTTPS requests. Users on the lowest payment tiers will be the first to be affected, as will... Weiterlesen

The Guardian view on NAZA’s persecuted directors: journalism is not treason | Editorial

Fri, 18 Sep 2026 19:30:56 +0200

Politicians who threaten filmmakers rather than investigate alleged military crimes expose not strength but a leadership frightened of the evidenceAcross Israel, politicians, military leaders and much of the media are trying to discredit a documentary that almost none of them have seen. Produced by the Guardian, NAZA examines the AI-powered... Weiterlesen

Could AI pose a serious threat to our existence? | Letters

Fri, 18 Sep 2026 19:30:56 +0200

Readers respond to the warning from a former Anthropic researcher that the technology could ‘kill us all’ by the end of the decade Here we go again with another round of tech bros claiming that AI is going to develop superintelligence and kill us all (More Anthropic researchers warn of AI’s perils but Musk dismisses ‘psyop’, 10 September). We’ve... Weiterlesen

How Trump Became an A.I. Evangelist

Fri, 18 Sep 2026 19:30:54 +0200

President Trump had little interest in artificial intelligence until a fateful meeting in Las Vegas two years ago. Now, Mr. Trump sees A.I. as the defining technology of our age. Our White House correspondent Jonathan Swan tracks Trump’s journey to A.I. evangelist. Weiterlesen

Trump Calls A.I. Fears a Hoax. Inside the White House, the Debate Is More Complex.

Fri, 18 Sep 2026 19:30:54 +0200

The president seems driven chiefly by the sense that A.I. is the way to economic growth, and nothing can be allowed to stand in its way. Weiterlesen

California Governor Gavin Newsom Issues A.I. Safety Executive Order

Fri, 18 Sep 2026 19:30:54 +0200

Gov. Gavin Newsom ordered the state to develop recommendations on safety measures, including requiring companies to build a “kill switch,” amid mounting concerns. Weiterlesen

Manus seeks $4B valuation in new $500M fundraise as it resumes independent ops

Fri, 18 Sep 2026 19:30:52 +0200

Manus, which earlier this year had to break off a merger with Meta, is in discussions to raise $500M at a $4B valuation. Weiterlesen