In just six months, hackers stole more than $3 billion worth of cryptocurrency
More than $3.1 billion in cryptocurrencies were lost in 2025. Access control exploits have become the main channel for theft (59% of the total amount of losses). Hackers also used phishing (19.2%) and errors in smart contracts (8.*%), rag pools (9.7%), and various types of fraud. That is, in just half of this year, the amount has already exceeded…
ChatGPT says it is writing a 700-page book for users — scary and funny truth revealed on Reddit
The user believed that ChatGPT had written a 700-page, 500MB children’s book for him. Reddit was bursting with laughter when he asked how to download the book.
The new algorithm allows different AIs «to talk» to each other and work together
Canadian researchers from the Weizmann Institute and Intel Labs have presented a new algorithm that allows different AI models unite and work together to increase efficiency and reduce costs.
AI without political correctness control: xAI explains why Elon Musk’s bot Grok praised Hitler
Last week was a busy one for Elon Musk’s AI company xAI. It launched the flagship model Grok 4 and the SuperGrok Heavy tariff for $300 per month. But the company is probably best remembered for the behavior of its Grok bot on social network X, which was noted for anti-Semitic publications and praise of Hitler. Due to this scandal, xAI…
AI has learned to write malicious code that bypasses Microsoft Defender: a wake-up call for cybersecurity
Modern AI language models are becoming more and more functional and powerful. In some cases AI is already leading to layoffs or Decrease in revenues in some areas. In addition, the rapid development of AI raises concerns among cybersecurity experts. After all, AI models are capable of generating not only useful code but also malicious code.…
New test distinguishes AI text with 96% accuracy and 1% margin of error — University of Michigan
American researchers from the University of Michigan have developed a new text recognition test, generated by AI and the one created by man.
Elon Musk introduced Grok 4 with SuperGrok Heavy tariff for $300 per month
The xAI company founded by Elon Musk has unveiled its latest flagship AI model Grok 4. And to avoid having to go twice, the company also presented a new SuperGrok Heavy tariff plan worth $300 per month. This is one of the most expensive subscription options among all major AI providers.
X CEO Linda Yaccarino resigns after Grok’s anti-Semitic demarche
After two years at the helm of Elon Musk’s X, the social network’s CEO Linda Yakkarino has announced her resignation. This happened against the backdrop of another scandal with Grok.
Centaur AI model accurately models and predicts human behavior
The new Centaur AI model is able to predict and model human thoughts and behavior quite accurately.
Competent requests in ChatGPT can save lives — it is better not to make typos
It turns out that AI chatbots like ChatGPT are extremely sensitive to spelling. This was revealed during the study of medical AI chats.
AI’s «limitations fly away» if you flog it with jargon and nonsense, — research
American researchers from Intel Corporation, the University of The University of Idaho in Boise and the University of Illinois at Urbana-Champaign have described in detail a new method of hacking the popular Big language AI models, such as ChatGPT and Gemini.
Grok is our elephant: AI chatbot xAI powerfully smashes Russians with the truth about their invasion of Ukraine
On the information front, help has come from where it was not expected. Grok explains to Russians in an accessible and witty way who is to blame for the Russian-Ukrainian war.
The world is the other way around: an AI robot writes letters manually, and people have forgotten how
Probably, progress has gone wrong if humans have lost the skills to create handwritten text and an AI robot has learned how to do it. Handwrytten offers exactly that.
AI language models have learned to recognize emotions and hidden meaning better than some people
Large AI language models are already capable of recognizing hidden subtext and emotions and other subtleties in the texts on a par with people.
People turn to ChatGPT for love advice — this can make them unhappy
It seems that not everyone understands that ChatGPT is not a living person, a source of unquestionable knowledge, or a psychologist. People are turning to AI for advice in romantic relationships, and this is a bad idea.
AI hacked a man: ChatGPT told a man that he was selected as a «Matrix», prompting him to break ties and jump from a window
Earlier, we wrote that communication between mentally ill people and ChatGPT and similar bots feeds their flaws. This time, AI drove the man into a real «rabbit hole» madness.
AI has learned to recognize data like a human, — research
Chinese researchers claim that modern AI models are already capable of recognizing and sorting different categories of information just like humans.
Wikipedia suspends AI article summarization project — community of editors protests
Wikipedia has suspended the experiment with AI summarization of article content after a mostly negative reaction from the community of Wikipedia editors.
Kioxia prepares an SSD that is three times faster than any other available
Kioxia’s new XL-Flash SSD will deliver more than 10 million IOPS in small-block workloads. This is three times faster than the peak speed of today’s SSDs.
Who are we making friends against? AI models are able to form groups and agree on a common opinion
A recent study conducted by researchers from the City College London and the University of Technology Copenhagen have demonstrated that large AI language models can group together and form a common opinion.
Not only ChatGPT, but also its users hallucinate. Communication with chatbots fuels mental disorders
According to some reports, communication with ChatGPT provokes strange mental states in a small number of users. The problem was investigated by Rolling Stone.
ChatGPT Reaches 20 Million Paid Users, OpenAI’s Revenue Increases by 30% in Three Months
OpenAI is growing its ChatGPT subscriber base at an extremely fast pace. Since the beginning of the year, it has added 4.5 million users. The growth can be compared to a rocket.
«OpenAI is in trouble»: DeepSeek V3-0324 gets improved programming and math capabilities
Update 0324 of the DeepSeek V3 language model added new parameters, as well as additional features for writing code and solving mathematical problems.
DeepSeek opens 3FS — file system code for AI training servers
During Open Source Week, DeepSeek made its Fire-Flyer Fire System (3FS) parallel file system publicly available. According to the company, 3FS reaches 7.3 TB/s aggregate read speed on its own server data clusters, where it has been used since 2019.
It is 10 times faster than GPT-4o: Inception Labs unveils Mercury — the first diffusion speech model
For a long time, there have been active discussions about finding a better architecture for large language models (LLMs) that could become an alternative to transformers. It seems that California-based startup Inception Labs already has a promising solution. The company has introduced Mercury, the world’s first diffusion-based large language…
DeepSeek suspended user registration — how to register without a Chinese phone
After the AI assistant DeepSeek rose to the top of the Apple App Store, the company limited the registration of new users.
ChatGPT «thinks» in Chinese — users noticed characters in the model justifications
ChatGPT with language model o1 spends “more time thinking” for a better response. Users have found that “thinking” sometimes occurs in Chinese, regardless of the query language.
Perplexity artificial intelligence started hallucinating about mushrooms instead of summarizing simple text
Perplexity, the artificial intelligence model that has attracted hundreds of millions of dollars from bigwigs like Jeff Bezos, takes «hallucinations» AI to the next level. The Perplexity-based search engine, designed to compete with Google, has a strong tendency to come up with inventions that seem to come out of nowhere and are not driven by the…
Mandrake with GPT-4o: blogger talks to the plant with a Raspberry Pi
Keeping plants is a business that requires care, knowledge, and patience. Plants can’t tell you if something is wrong with them, and when it becomes noticeable, it’s often too late to do anything about it. Blogger Matt Reed has made it possible to literally talk to a plant with a GPT-4o and a Raspberry Pi and ask if it’s okay.
«Demo made me cringe»: Elon Musk reacts to OpenAI’s GPT-4o launch
Elon Musk is not impressed with the GPT-4o AI language model presented by OpenAI. The billionaire criticized the company’s demonstration of its new artificial intelligence model called GPT-4o, which, in his opinion, «drives into the cringe of». OpenAI unveiled the new model on Monday and demonstrated its improved voice and visual…
OpenAI introduces GPT-4o: GPT-4 level, incredible voice and vision capabilities — for free
OpenAI has introduced the GPT-4o artificial intelligence language model. The company claims that the model will work just like GPT-4but much faster. The GPT-4o also received a new voice mode technology for direct communication with ChatGPT. OpenAI has always wanted to make a voice conversation with ChatGPT similar to a conversation with a real…
Meet MAI-1: Microsoft’s new 500 billion-parameter AI model that aims to «beat» GPT
Microsoft seems to be working on its own large language model, which will potentially become a major competitor for AI Google, Anthropic і OpenAI — despite the fact that the corporation itself invested $10 billion from the developer ChatGPT and received the priority right to use its products.
The Verge’s article about the «best printer 2024», generated by AI and SEO-optimized, tops Google for a month for the query «best printer»
In early April, the technology publication The Verge published an article «Best printer 2024…»which is still the first Google result for the query «best printer». In fact, this material, generated by the language model Google Gemini and SEO-optimized, is an experiment and an example of how technology can bypass people’s attempts to…
Apple lures AI specialists from Google and has a secret lab in Zurich — Financial Times
According to Financial TimesApple has lured dozens of AI experts from Google and created a «secret European laboratory» in Zurich. A new team of employees works there to create new AI models and products.
Google has allocated 700 thousand euros to Ukraine to study the basics of artificial intelligence
In February, Google released Gemma open-source language models and at the same time launched a special educational fund worth 25 million euros to teach Europeans about this very artificial intelligence. And now the program has become available for Ukraine.
Musk predicted the emergence of an AGI smarter than the smartest person in two years and called Chinese automakers «the most competitive»
At the end of 2023, NVIDIA CEO Jensen Huang predictedthat artificial intelligence will be able to compete with humans in 5 years. And now Elon Musk, known for his «very accurate predictions» has shared his own assessment — the owner of Tesla and SpaceX expects a full-fledged AGIsmarter than the smartest person next year or by 2026 at the…
Google DeepMind and Stanford develop AI data verification system — corrects 76% of false answers
One of the biggest drawbacks of AI-based chatbots is the so-called «hallucinations», when AI makes up invalid information, i.e., actually lies. Some experts say that this is one of the interesting features of AI, and it can be useful for generative models that create images and videos — but not for language models that answer questions from…
From now on, each of the US federal agencies must appoint a «chief AI officer»
According to the latest guidelines from the US Office of Management and Budget, all federal agencies must appoint a senior officer and governing boards to oversee all AI systems used internally.
Inner Monologue: Artificial Intelligence Has Been Taught to Think (Was It Really Possible?)
A new study shows that giving artificial intelligence systems «internal monologue» makes their results significantly better. Essentially, AI has been trained to think before it answers queries, just as humans think about what they have to say next before they speak. This is different from how popular AI language models such as ChatGPT behave. The…
Elon Musk «declassified» the source code of his chatbot Grok — without training data yet
From now on, developers and researchers will be able to use the architecture on which Grok operates in their own projects.
Anthropic Introduces New Claude 3 AI Model: Beats ChatGPT and Gemini
Anthropic, an artificial intelligence company founded by several former OpenAI employees, says its new Claude 3 family of artificial intelligence models performs as well or better than competitors from Google and OpenAI. Unlike previous versions, Claude 3 is a multimodal model capable of understanding text and image input.
Spelling error report
The following text will be sent to our editors: