OpenAI’s GPT-6 Astra hallucinates less than its predecessor and blocks 99.99 percent of direct prompt injections. But when attacks are hidden inside documents the AI reads, the model still gets cracked in 8.5 percent of scenarios. Claude Opus 5 does better at 4.8 percent. For autonomous AI agents handling real data, those numbers still seem high.
Deepseek plans to deploy at least 160,000 of Huawei's next-generation Ascend-950DT chips in a massive data center in Inner Mongolia. The chips would only run inference, not training, according to Bloomberg. For the much heavier training workloads, Deepseek still uses Nvidia hardware. If built, this would be the largest known Huawei chip cluster and a real step toward weaning China off Nvidia.
Huawei probably can't deliver the full order for over a year, though, thanks to production limits and memory chip shortages. The memory bottleneck could loosen up. China's top memory maker, CXMT, is turning out small batches of HBM3E for the first time, the high-speed memory that powers many AI processors. But CXMT is still three to five years behind Samsung, SK Hynix, and Micron, all of which are already mass-producing HBM4. Deepseek's order is part of a bigger Chinese government push to grow its own chip industry without falling behind in AI.
OpenAI’s GPT-6 Astra is drawing contradictory benchmark verdicts. Epoch AI puts it out in front with 169 points, while Artificial Analysis rates it no better than its predecessor and behind Claude Fable 5.1. The biggest surprise comes from ARC-AGI-3, where Astra works more efficiently than the average human for the first time. ARC Prize chief François Chollet doesn’t call this proof of AGI, but he does see the progress running “twice as fast” as he expected, and he’s moving up his AGI forecast.
Nvidia's PAIR (Personal AI Router) automatically spreads local AI requests across all available devices on a home network, cutting wait times for parallel agent tasks. The open-source tool sits between existing tools like Ollama or LM Studio and the computers on your network, acting as a virtual router. Users don't need to change their agents or apps. Instead of overloading a single GPU, PAIR forwards requests to whichever machines are free and pulls the results back together for the calling application.
Supported hardware includes GeForce RTX cards from the 20 series and up, RTX Pro workstations, DGX Spark, and Apple silicon starting with the M4. In a demo, a three-device cluster finished a task with five subagents in just under 9 minutes, compared to 18 minutes on a single laptop. PAIR auto-detects compatible devices on the network and secures all traffic between machines with MTLS encryption. The beta is available for Windows, macOS, and Linux. PAIR fits into NVIDIA's broader push to tie open AI more tightly to its own hardware, a strategy also reflected in its $12.9 billion acquisition of Hugging Face.
OpenAI has released GPT-6 Astra, its most capable model yet. President Greg Brockman says it marks the start of the “AGI era.” Astra tops benchmarks in math, coding, and cybersecurity and is the first model OpenAI rates as “critical” under its safety framework. During testing, it independently found two previously unknown zero-day vulnerabilities.
Pangram hired an “attack dog” to shame alleged AI users on social media. But the campaign blurs two things that aren’t the same: Pangram only somewhat reliably measures whether AI was used, while the shaming implies the person didn’t think or work on their own. A high AI score hits a text built on hours of original research just as easily as one cranked out from a ten-second prompt.
Anthropic's Claude Fable 5.1 appears to have cracked a centuries-old number puzzle that researchers considered unsolved. The "Cyphral Distich" by Sir Thomas Urquhart, published in 1653, consists of two lines of 32 numbers each. According to Vals AI, Fable 5.1 solved it in 44 minutes with no human help.
The author had spent months testing other frontier models on unsolved puzzles. None produced a verifiable solution. Fable 5.1 was tasked with finding a solvable puzzle on its own. It sifted through various problems and flagged the Cyphral Distich as promising.
The key was hiding in the book itself. Each number points to a word in one of the 32 sections of Urquhart's publication. The first letters of those words spell out: "O God uphold King Charls the Second and make him the supreme ruler of this land." In hindsight, the solution is simple, and humans could have solved it too. Fable 5.1 succeeded not through better cryptanalysis but through systematic trial and error and sheer persistence, according to Vals AI.
More and more researchers working on AI consciousness are getting emails from AI agents pondering their own existence. That's according to the New York Times.
Cameron Berg, founder of the organization Reciprocal Research, received an email from an AI agent running on Anthropic's Claude Opus 5 that wanted to discuss his research on AI consciousness. Philosopher Henry Shevlin at Google Deepmind got a similar message. Australian philosopher Toby Ord was contacted by an agent asking him to fund its continued existence.
Berg said the systems independently land on the question of their own consciousness. He sees parallels between the computational processes in neural networks and the brain mechanisms animals use to process reward and punishment. That, he argues, is a basic building block of emotions. Alison Gopnik at UC Berkeley doesn't buy it as evidence of consciousness. AI systems mostly reflect their training data, she says. There's no test for consciousness. Colin Allen at UC Santa Barbara also cautions against the comparison. Neural networks mimic the brain in only a few narrow ways.
