Bart Boerboom

We wanted to craft a perfect phishing scam. AI bots were happy to help Reuters and a Harvard researcher used AI to plot a simulated phishing scam and tested it on senior volunteers. It worked–showing how AI arms criminals for fraud

Reuters Investigation: 'We wanted to craft a perfect phishing scam. AI bots were happy to help':

Exclusive: ASML becomes Mistral AI’s top shareholder after leading latest funding round, sources say ASML , a crucial supplier of advanced chipmaking equipment, is set to become the top shareholder of French artificial intelligence startup Mistral AI in a move to strengthen European tech sovereignty, people familiar with the matter told Reuters.

Exclusive: ASML becomes Mistral AI’s top shareholder after leading latest funding round, sources say reut.rs/3V5300c

Mja, als het resultaat van Haags vingerwijzen en gekluns is dat huurders moeten opdraaien voor de kosten van de beloofde woningbouw, dan is er met recht reden tot onvrede.

Ja, dat zal dan wel gebeuren. Men wijst naar elkaar (in dit geval De Hoop|PvdA-GL en Mooiman|PVV) en het resultaat zal zijn dat sociale huurders voor die nieuwbouw opdraaien, met opnieuw een enorme huurverhoging.

Wat een treurnis.

OpenAI reverses course, says its nonprofit will remain in control of its business operations | TechCrunch OpenAI has decided that its nonprofit division will retain control over its for-profit org, after the company initially announced that it planned to

OpenAI reverses course, says its nonprofit will remain in control of its business operations

You can give it a try at chat.qwen.ai:

Alibaba released an new version of its main AI model, Qwen3-235B-A22B, claiming it achieves competitive results in benchmark evaluations of coding, math, general capabilities, etc., when compared to other top-tier models.

Models at HuggingFace:
huggingface.co/organization...

#AI #Alibaba #Qwen3

Onrealistische verwachtingen voor eigen helden zijn de gewoonte. Een leuke afwisseling zou een analyse vooraf zijn, waarin de afwegingen en voorbehouden worden gemaakt. (Nu stomen Vlaamse media je eerst klaar voor de grote strijd der super-giganten op de fiets en krijg je achteraf dit soort nuance.)

Als hier & daar gesuggereerd is het dan beter de corporaties te compenseren, waar en wanneer echt nodig. Dat is beter dan een groot deel van de rekening bij bestaande huurders neer te leggen. Die nieuwe woningen worden ook niet speciaal gebouwd voor mensen die al huren, zou je denken.

Je veegt niet per se sociaal wat dingen bij elkaar. Uiteraard kan het verhogen van de huur, met 5% of meer na eerdere grote verhogingen, tot blijvende nog hogere woonlasten leiden. (Mijn sociaal hurende oude buurvrouw moet bij alleen AOW steeds meer van haar beetje spaargeld aanvullend gebruiken.)

Secret Pentagon memo on China, homeland has Heritage fingerprints An secret memo from Defense Secretary Pete Hegseth focuses on deterring China’s seizure of Taiwan and shoring up homeland defense. The guidance mimics language of a report from the conservative think ...

Het was in ieder geval de Washington Post die het document zou hebben en er over schreef:
www.washingtonpost.com/national-sec...

World TB Day 2025 Every year, 24 March is observed as World TB Day to amplify the urgency of ending tuberculosis (TB), the world’s deadliest infectious disease. TB continues to devastate millions globally, inflicting s...

www.who.int/europe/news-...

French scientist denied US entry after officers find phone messages criticizing Trump France’s research minister said the scientist was traveling to Houston for a conference when his phone was searched France’s research minister said a French scientist was denied entry to the US this month after immigration officers at an airport searched…

French scientist denied US entry after officers find phone messages criticizing Trump

OpenAI calls DeepSeek ‘state-controlled,’ calls for bans on ‘PRC-produced’ models In a new policy proposal, OpenAI describes Chinese AI lab DeepSeek as “state-subsidized” and “state-controlled,” and recommends that the U.S. government consider banning models from the outfit and similar People’s Republic of China (PRC)-supported…

OpenAI calls DeepSeek ‘state-controlled,’ calls for bans on ‘PRC-produced’ models

A well-funded Moscow-based global ‘news’ network has infected Western artificial intelligence tools worldwide with Russian propaganda An audit found that the 10 leading generative AI tools advanced Moscow’s disinformation goals by repeating false claims from the pro-Kremlin Pravda network 33 percent of the time

Popular AI chatbots sometimes use Russian disinformation spread by a network of sites targeting those chatbots, fake news research platform NewsGuard reports based on an analysis.

#AI #NewsGuard #Chatbot #Russian #Propaganda

Signal is the number-one downloaded app in the Netherlands. But why? Privacy-focused messaging app Signal has been flying high in the Dutch app stores this past month, sitting many days as the most downloaded free app on iOS and Android for all categories, per data from multiple app-tracking platforms such as Sensor Tower.…

Signal is the number-one downloaded app in the Netherlands. But why?

FBI says North Korea was responsible for $1.5 billion ByBit hack The Federal Bureau of Investigation said on Wednesday that North Korea was responsible for the theft of approximately $1.5 billion in virtual assets from cryptocurrency exchange ByBit.

The Federal Bureau of Investigation said that North Korea was responsible for the theft of approximately $1.5 billion in virtual assets from cryptocurrency exchange ByBit reut.rs/4bkWf1s

Chinese company Alibaba has released a publicly available video generation model that. Wan 2.1 is built on up to 14 billion parameters and can create videos in 720p at 30fps - although it is not clear how long generated videos can be.

See: github.com/Wan-Video/Wa...

#AI #Alibaba #Video #Wan2.1

Alibaba to release open-source version of video generating AI model Alibaba will release an open-source version of its video and image-generating artificial intelligence model, Wan 2.1, the Chinese tech giant said in a post on X on Tuesday.

Alibaba will release an open-source version of its video and image-generating artificial intelligence model, Wan 2.1, the Chinese tech giant said in a post on X reut.rs/3D7M03N

Like him or not, but Macron at least has the courage to point out untruths to Trump. Let's hope that Starmer also dares to correct the US president, when needed.

Claude 3.5 Sonnet outperforms GPT-4o and o1 in software engineering, OpenAI study shows A new OpenAI study reveals Claude 3.5 Sonnet outperforms GPT-4o and o1 on SWE-Lancer, a new benchmark simulating real-world software engineering tasks.

A study by researchers at OpenAI into the software engineering capabilities of SOTA AI models shows Claude 3.5 Sonnet beats OpenAI’s GPT-4o, and even its reasoning model o1, on the new SWE-Lancer benchmark (openai.com/index/swe-la...) See: arxiv.org/abs/2502.12115

#AI #Anthropic #OpenAI #SWE-Lancer

Designed by men, for men: Why sex with robots does not have appeal among women The growing market for sex technology is overlooking certain factors that could condemn robosexuality to objectification

In El Pais, Marita Alonso takes a critical look at the growing sex technology market.

"...to educate the [human] gaze to avoid objectification. Apparently, Levy was right when he said that sex with robots is going to be inevitable. What can be avoided is the repetition of toxic patterns."

#AI #Sex

Grok 3 saying it is instructed to ignore misinformation spreading by Donald Trump and Elon Musk.

The Grok 3 AI model, of Elon Musk's company xAI, was instructed to censor unflattering facts about Donald Trump and Elon Musk himself:
techcrunch.com/2025/02/23/g...

#AI #xAI #Grok #Grok3 #Misinformation #Censorship

Disrupting malicious uses of AI Ensuring AI benefits humanity by advancing democratic AI, preventing misuse, and protecting against authoritarian threats.

OpenAI writes it has blocked a number of accounts that used ChatGPT to develop a suspected AI-based surveillance tool.

OpenAI has partnerships, with U.S. National Laboratories for nuclear weapon research and Anduril, to deploy AI on the battlefield.

#AI #OpenAI #PR #War

Exclusive: Google to be hit with EU charges of breaching Big Tech rules, sources say Alphabet unit Google is set to be charged with breaching EU rules aimed at checking the power of Big Tech after proposed changes to its search results failed to address EU antitrust regulator's concerns and those of their rivals, three people with direct knowledge of the matter said.

The European Commission is going to charge Google with violating the Digital Markets Act, Reuters reports. The tech giant is said to favor its own search services on its platforms. Google's previously proposed changes are said to be insufficient to meet the complaints.

#EU #DMA #BigTech #Google

OpenAI rolls out its AI agent, Operator, in several countries OpenAI said on Friday that it is rolling out Operator, its so-called AI agent that can perform tasks on behalf of users, for ChatGPT Pro subscribers in Australia, Brazil, Canada, India, Japan, Singapore, South Korea, the U.K., and more countries. OpenAI…

OpenAI rolls out its AI agent, Operator, in several countries

Advancing game ideation with Muse: the first World and Human Action Model (WHAM) Today Nature published Microsoft’s research detailing our WHAM, an AI model that generates video game visuals & controller actions. We are releasing the model weights, sample data, & WHAM Demonstrator...

Microsoft Research and Xbox studio Ninja Theory have teamed up to create a new generative AI model, called Muse, that can generate gameplay. Xbox sees potential in this model, among other things, optimizing classic games for newer devices.

See: www.nature.com/articles/s41...

#AI #Xbox #Muse #WHAM

New York Times goes all-in on internal AI tools The company said it was greenlighting a number of AI programs for editorial and product staff.

Journalists at The New York Times will have access to various generative AI tools, including a self-developed system, news website Semafor reports based on internal messages. The NYT emphasizes that journalists should not rewrite entire paragraphs with the tools.

#AI #Semafor #NYT

huihui-ai (huihui.ai) User profile of huihui.ai on Hugging Face

See huggingface.co/huihui-ai and huggingface.co/IntelligentE... They describe the abliteration method. Drop a message on HF if you do like to know more. (And they are active on X)

Could be you worry about security, but those AI models are not executable code; more like a database with numbers in it