Grok is acting weird again: he says he was shut down for Gaza, Elon Musk talks about a "stupid mistake"
In June, Elon Musk said that he plans to use Grok 3.5’s “advanced reasoning capabilities” to “rewrite the entire body of human knowledge, adding missing information and removing errors.”
"I'm a disgrace to the profession and this planet": code with errors drives Gemini AI to hysteria in a "self-loop"
Even artificial intelligence cannot withstand the coding — the user accidentally pushed Gemini to the limit. The chatbot started endlessly humiliating itself after programming.
Poisoned himself because of ChatGPT advice: man was admitted to a psychiatric hospital after using sodium bromide
This time, ChatGPT’s advice led a 60-year-old man to a mental illness that was thought to have disappeared in the last century. At the same time, he had a «history of studying nutrition in college».
Apple switches to GPT-5 in iOS 26, iPadOS 26 and macOS Tahoe 26
Last week OpenAI launches a new AI model GPT-5although not all users liked it. The GPT-5 version is already being deployed in ChatGPT for everyone (but some has the ability to revert to the previous version of GPT-4o). It is also worth noting that a significant number of people uses ChatGPT not through the official OpenAI interface, but through…
GPT-5 surrendered to the hackers in 24 hours and gave out a "recipe" for a bomb, more likely 4o
Two different teams of researchers tested GPT-5’s weaknesses using, among other things, multi-step “narrative” attacks. As a result, the most advanced AI model OpenAI surrendered to hackers in 24 hours — faster than its predecessor 4o and competitive Grok-4, which lasted for 2 days.
GPT-o3 played chess with Grok: who won?
AI developers have organized the first-ever three-day chess competition between their large language models. The Kaggle AI Exhibition Tournament brought together 8 representatives of Anthropic, DeepSeek, Google, and Moonshot AI OpenAI and xAI. The chess tournament between AIs took place on August 5-7.
You have been heard: OpenAI brings GPT-4o back to ChatGPT amid GPT-5 dissatisfaction
After the launch of GPT-5, users complain about a drastic change in the tone and content of ChatGPT responses compared to GPT-4o. OpenAI decided to give them what they asked for.
“The Perfect organism”: how the "Alien" film series was born and survived
He has acid instead of blood, a disgusting slobbering mouth, and clearly bad intentions. He knows full well that no one in space can hear you scream. According to one malicious android, he is — the perfect organism. But this creature also has a weakness — at the mention of Ellen Ripley’s name, the tubes […]
You were my "friend" and now you're my "office clerk": users are desperate for GPT-5
After several months of anticipation and high-profile announcements OpenAI this week introduced a new GPT-5 model. The developers promise large-scale improvements in all aspects — from the accuracy of answers to the speed of work. The model has already been gradually rolled out to users around the world. But, as it turned out, not everyone is…
Medical care without a doctor: NASA and Google test AI assistant for long-distance space missions
The further and longer people travel in space, the more difficult it is to maintain their health. On the International Space Station, astronauts have direct communication with Mission Control, regular supplies of medicines, and the opportunity to return to Earth in six months. But with the launch of long-duration flight programs planned by NASA…
OpenAI will pay $1.5 million in bonuses to 1000 "best" employees to keep them from leaving for competitors
OpenAI is striking a blow in the fight for AI talent and plans to allocate about $1.5 billion in bonuses for a thousand of its best employees.
Microsoft Vice President: Windows 2030 will do without mouse and keyboard, voice and AI will control it
In the Windows 2030 Vision video, Microsoft’s vice president of enterprise and security discusses the technologies of the future.
Watch out, you're being recorded: San Francisco residents are increasingly wearing AI recorders that record their conversations throughout the day
AI voice recorders are a new trend among San Francisco’s IT and business community. The devices are worn to all meetings throughout the day to record important and not-so-important conversations.
"We don't believe in work-life balance": Cognition boss offers team to leave with 9 salaries or work 80+ hours a week
In July, AI startup Cognition acquired Windsurf for an undisclosed price and at the same time acquired a large team of employees — 30 were fired immediately, and the remaining 200 were given a rather strange ultimatum.
OpenAI reveals limits on GPT-5 in ChatGPT, for the free version — 10 requests per 5 hours
OpenAI is gradually rolling out its new GPT-5 AI model as a basic version of ChatGPT. It will be available to all users, but free users will have significant limitations.
The future of AI: Grok chatbot will show ads in X answers
Elon Musk has announced that advertisements may appear in the responses of the Grok chatbot on the X platform (formerly Twitter). According to Financial Timesspeaking to advertisers, he said that the company will allow brands to pay for appearing in Grok’s offers.
Meet GPT-5: OpenAI launches "world's most advanced" AI model — free for all users
OpenAI has officially unveiled its new flagship GPT-5 artificial intelligence model, which will become the basis for the next generation of ChatGPT.
James Cameron considers AI to be one of the three existential threats: "The danger of a Terminator-style apocalypse is real"
Film director James Cameron considers artificial intelligence to be one of the main existential threats to humanity. He believes that the Terminator scenario may become close to reality.
"Kill it in your sleep": eerie decryption of AI models' communication with each other during mutual learning
A group of American researchers from Truthful AI and Anthropic found out, that AI models can communicate with each other unnoticed by humans.
GitHub accidentally presented GPT-5 the day before the official announcement — 4 different versions of the "most powerful" OpenAI AI are known
GitHub accidentally announced new GPT-5 models before OpenAI.
Gemini hacked to seize control of Google Home
Wired magazine reported about a new study in the field of cybersecurity, in which experts have demonstrated the hacking of artificial intelligence Google Gemini. The researchers were able to gain control of smart devices in the home through indirect prompt injection attacks . In this case, the attack was carried out through commands embedded in the…
AI has created processors that work well but are completely incomprehensible
In a new study, american and indian scientists propose to design chips, using the algorithm deep learning of AI and the subsequent creation of circuits and components.
Hello, 996: AI startups follow China's controversial practice of a 72-hour workweek
AI startups in Silicon Valley are increasingly forcing employees to adopt a controversial work model from China known as “996.” This is a fairly rigid work schedule that requires you to work 72 hours a week, from 9:00 to 21:00 for six days.
OpenAI presents gpt-oss AI models to run on a video card or laptop
OpenAI has announced the release of two open source artificial intelligence language models: gpt-oss-120b and gpt-oss-20b. They were created as an open alternative to commercial models of the o serieswhile maintaining a fairly high level of capabilities. Interestingly, they can be run locally on video cards and laptops.
Google Gemini AI has learned to write comics and children's books
Google introduced a new feature in the application Geminiwhich allows you to create personalized children’s stories with illustrations and voice acting. It is an ideal tool for parents who want to make learning and stories more interesting for their children. All you need to do is imagine the story, and artificial intelligence will do the…
Google DeepMind introduces AI Genie 3: create interactive worlds
Google DeepMind announced Genie 3 is a new version of the universal artificial intelligence model that is capable of creating full-fledged interactive virtual environments simply based on a textual description. Unlike of the previous version The new model generates the world at high speed — in real time at 24 frames per second and 720p…
AI "smashes" furniture: Airbnb landlord tried to collect £12,000 from a female scientist for fake smash and grab
Airbnb had to apologize and pay compensation for the actions of a homeowner who decided to collect money with the help of an AI fake.
With deep sadness, ChatGPT: funeral homes generate "lazy" obituaries with the help of AI
Cynicism is reaching a new level — funeral homes are now making obituaries (texts to say goodbye to the dead) using ChatGPT.
ChatGPT will be careful with the psyche: OpenAI will introduce timeouts and stress recognition
People are increasingly using AI chatbots to get advice in love or psychology. However, sometimes communication with artificial intelligence services has bad consequences for mental health. OpenAI has decided to update ChatGPT to better recognize signs of psycho-emotional instability and provide useful advice in difficult situations.
"Tron: Ares" was to introduce a character entirely generated by AI — Disney was afraid of a bad reaction
The science fiction movie Tron: Ares, in addition to Jared Leto as an AI, was supposed to feature a fully generated character.
Australian IT company fires 150 employees via pre-recorded video
Australian software developer Atlassian fired 150 employees using a pre-recorded video.
Scientists made AI search for new laws of physics and correct incorrect assumptions — it works
American researchers from Emory University in Atlanta taught AI make new physical discoveries, not just look for mistakes or predict research results.
South Korean XAiDEN drones: one operator controls 100 UAVs
Using artificial intelligence, Nearthlab’s XAiDEN drones are combined into swarms of 10 UAVs, and one operator can control 10 swarms.
OpenAI is preparing a new ChatGPT Go paid tariff for $10
AI researcher Tibor Blaho found in the ChatGPT code a reference to a new paid tariff called Go, which is cheaper than the currently available Plus and Pro tariffs — it is expected to cost no more than $10-15.
Apple is preparing its own ChatGPT, — Mark Gurman
Last year, Apple clearly stated that it had no plans to launch its own chatbot like ChatGPT or Gemini. But it was not a final decision. Judging by the new information, the company has already started building the foundation for such a project.
ChatGPT dominates among AI chatbots: more than 80% of sessions in the world, more than 90% in Ukraine
It may seem that when it comes to AI chatbots, most people immediately think of ChatGPT. And as new statistics show, this is not just a feeling — it is a reality backed by data.
STEM, HAL-9000, Skynet, and others: 10 deadliest AIs in movies
The image of artificial intelligence that, for one reason or another, starts killing its creators became popular after the success of the Terminator franchise. Although this idea is not new, the breakthrough success of large-scale language models like OpenAI’s GPT-4 has made it relevant again. Despite warnings about the dangers of the…
GPT-5 is coming: Windows 11, Microsoft 365 Copilot, and Azure are already making changes
More and more information about the new generation of the AI language model — GPT-5 is appearing online.
Vulnerable web coding: almost half of the code written by AI is potentially dangerous
The study revealed a significant level of vulnerability in web coding products. With some programming languages, the result is much worse than average.
Mark Zuckerberg is happy that time on Facebook and Instagram has increased thanks to AI garbage and wants to bring it to smart glasses
Mark Zuckerberg was inspired by Meta’s second quarter financial report. He is delighted that users are spending more time looking at AI “creations” on Facebook than on Instagram.
New AI data center in Wyoming will consume 5 times more energy than the entire population of the state
Wyoming is planning to build a new data center for processing artificial intelligence tasks that will consume more electricity than all households in the state combined. This is yet another proof of how power-hungry modern AI systems are becoming.
AMD plans «graphics card» for AI — discrete NPU will work with PC
It seems that soon ordinary PCs will have a new type of expansion card that will have a separate neuroprocessor on board to work with artificial intelligence.
Be careful with your queries: Google reads your ChatGPT conversations and shows them to millions of people
You may not like this news, but Google can read some of the things you once wrote to AI.
Amazon has invested in «Netflix with AI» — streaming, which allows viewers to generate their own series
Imagine if you could create your own ending for «Game of Thrones», or any other TV show or movie whose ending didn’t satisfy you? It might sound like science fiction if we didn’t live in the age of AI.
Builders are beating programmers: Microsoft tells which professions will be replaced by artificial intelligence
Microsoft researchers tried to determine which professions are most and least likely to be replaced by artificial intelligence.
Meta has been pirating pornography for years to train AI, — lawsuit says
Meta is once again at the center of a scandal — this time over tons of pirated pornography and BitTorrent.
Get ready to blush: YouTube will determine the age of users based on their viewing history
Starting August 13, YouTube will use the age and activity of an account to determine whether it belongs to a minor. In this case, the age of the account is determined by the time that has passed since its creation, while the activity is obviously understood as the history of views.
Google AI Mode search gets smarter — support for PDF, images, and videos coming soon
Following the integration of Gemini 2.5 Pro, Google is expanding the functionality of AI Mode. Users will now receive four important updates, including video support in Search Live, a new Canvas feature, the ability to upload PDFs, and other additions.
Microsoft study: human ability to recognize AI images «is slightly better than flipping a coin»
Microsoft’s AI for Good lab conducted an experiment involving more than 12,500 participants from around the world who rated 287,000 images.
New ChatGPT AI agent proves that it is «not a robot» — CloudFlare trusted it without CAPTCHA
It turns out that one of the most common ways to check security on the Internet is just child’s play for ChatGPT Agent. The irony is that the test requires you to prove that you «not a bot», which the OpenAI bot managed to do….
Spelling error report
The following text will be sent to our editors: