10114
Discover the future of AI with ChatGPT! 🤖💬 Get AI insights, tips, tricks, and more. Questions: @AI_inquires_bot
AI just got called out
Polaroid has launched an anti-AI campaign aimed at the growing water use behind data centers.
Some updates point to more power. Others point to more limits, more regulation, and more control over who gets access to the best systems.
Читать полностью…
AI is moving too fast to ignore.
In just one week, the AI world saw new models, chip plans, coding agents, workplace tools, robots, and tighter access controls.
AI made pool floats weird
A new prompt is turning famous faces into human sized inflatable pool toys.
Elon Musk replied with a shorter version of the same idea, physics with math. The message is simple, code can change fast, but first principles stay useful for decades.
As AI writes more code, the advantage may move to people who understand the world beneath the software.
🇨🇳 The DR02 by DEEP Robotics is the world's first IP66-rated humanoid robot, meaning it's completely dustproof and water-resistant.
This humanoid from Hangzhou also does pushups like it’s nothing.
OpenAI built its own chip
OpenAI has introduced Jalapeño, its first AI chip, built with Broadcom. It was made for LLM inference, which means answering user requests inside tools like ChatGPT and Codex.
The AI model war got messy
Anthropic has accused Alibaba of running the largest distillation campaign it has seen against Claude, its AI model. According to Reuters, the campaign used nearly 25,000 fraudulent accounts and more than 28 million Claude interactions between April and June.
Trump backed off Anthropic.
After a week of tension, Donald Trump says he no longer considers Anthropic a national security threat. The dispute started after the U.S. ordered Anthropic to block foreign nationals from using Fable 5 and Mythos 5, two advanced Claude models tied to cybersecurity work.
AI data centers use a lot of water to keep servers cool. As demand for AI grows, cooling is becoming one of the biggest hidden problems behind the infrastructure race.
Nvidia says liquid cooling could reduce that pressure.
Each one turns ChatGPT into a more focused assistant for studying, strategy, or personal growth.
Better prompts do not make AI magic, but they can make the output far more useful.
Most users will never reach that level, but power users can turn a flat monthly fee into a massive compute discount.
That helps explain why AI companies are watching usage limits so closely.
ChatGPT lost its majority
ChatGPT’s global market share has dropped below 50 percent for the first time. OpenAI’s chatbot is still the most used AI assistant in the world, but the gap is getting smaller.
The AI race is no longer only about smarter models, it is about hardware, safety, talent, infrastructure, and real products reaching the market.
Читать полностью…
OpenAI, Google, Meta, Anthropic, NVIDIA, Slack, and robotics companies are all moving at the same time, but in very different directions.
Читать полностью…
The image is framed like a luxury product photo, but the joke is how strangely believable it looks. It turns celebrity recognition into a surreal object you could almost imagine seeing in a summer catalog.
It also shows how fast image models can remix culture into products that never existed.
AI security just got louder
OpenAI released GPT 5.5 Cyber, a model built to find and fix software vulnerabilities faster. The company says it topped the CyberGym benchmark, beating Anthropic models and older GPT versions.
The timing makes this more interesting. Anthropic’s Mythos model was recently caught under an export ban, while OpenAI is now pushing a model focused on cyber defense.
AI is moving deeper into security work, where speed can decide who finds a flaw first.
Math just got louder
Pavel Durov told students not to waste their time learning programming or AI. His advice was to study mathematics and physics instead, because those skills teach people how to think clearly.
The chip was designed from scratch and reached tape out in just nine months. OpenAI says its own models helped speed up parts of the design process.
Early tests show better performance per watt than leading AI chips, though final numbers are still being measured.
If this works at scale, OpenAI gets more control over the compute behind its products.
Distillation means using one model’s answers to train another model. Anthropic says operators linked to Alibaba’s Qwen lab targeted Claude’s software engineering and agentic reasoning skills. Alibaba has not commented on the claims.
This shows how the AI race is moving from model launches to model security.
Anthropic said it could not reliably separate access only for Americans, so it disabled both models for everyone. That decision sparked backlash from cybersecurity experts, who argued that defenders outside the U.S. could lose access while attackers still turn to other models.
This shows how AI safety is moving beyond product launches and into export controls, identity checks, and national security fights.
Its system can run with coolant at about 113 degrees Fahrenheit, warm enough to avoid traditional chillers in some setups.
If this works at scale, AI data centers could use far less water while still running powerful chips.
The idea is simple, when you give AI a clear role and a clear goal, the answer gets much better.
This carousel shares 7 prompt structures for deeper learning and sharper thinking.
A SemiAnalysis estimate suggests a $200 per month ChatGPT Pro plan could represent up to $14,000 in API usage if pushed to its full potential.
The analysis compared high usage across OpenAI and Anthropic plans, arguing that the real value of these subscriptions depends on how heavily people use them.
Google is gaining fast because most users do not pick tools in a vacuum. They often use whatever is already closest to their browser, phone, search habits, and daily apps.
The AI assistant market may not be won by the best model alone, but by the company that controls where people already spend their time online.
For the next two weeks, Plus and Pro users can invite up to three friends to try Codex.
When a friend sends their first Codex message, you’ll both get another banked reset.