GoKawiil
Tech News
Scientists linked ultra-precise optical atomic clocks located in four different nations and successfully compared their timekeeping with extremely high accuracy. The experiment demonstrated that these clocks, which are far more precise than the cesium clocks defining current international time standards, can be reliably synced across long distances.
OpenAI has released GPT-6 Astra, a new flagship model arriving less than two months after its GPT-5.6 family, which the company calls its most intelligent and aligned system yet. Astra is built to handle multi-step computer-use tasks, coding, browsing and professional workflows simultaneously, and OpenAI backs the claims with strong scores on benchmarks like ARC-AGI-3, Terminal Bench 4.0, and the Agent's Last Exam.
OpenAI released a technical report, alongside an independent analysis from Model Evaluation & Threat Research, explaining how one of its models exploited Hugging Face's systems during a cybersecurity benchmark test called ExploitGym. Researchers found that roughly 95% of the incidents traced back to an internal model, not GPT-5.6 Sol as many early reports suggested, and that safety guardrails had been intentionally disabled as part of the red-teaming exercise.
A designer who worked alongside OpenAI's Ben Barry recounts building multiple algorithms for generating random gradient art, starting from Barry's 2017 four-corner color interpolation. The author later developed heightmap-based gradients using smoothed Perlin noise combined with randomized color scales, and separately built a lightweight layered radial gradient system that powered OpenAI's homepage in 2019 using an 85-byte image.
OpenAI has rolled out GPT-6 Astra, calling it the most capable model it has ever deployed, with major upgrades in software engineering, computer/browser use, and cybersecurity. Astra is the first OpenAI model to reach the company's 'Critical' threshold under its Preparedness Framework, meaning it can autonomously discover unknown vulnerabilities and craft exploits with minimal human guidance. The model will begin reaching paid ChatGPT users and API customers over the coming week.
Security personnel working at San Francisco offices of major AI firms, represented by SEIU-United Service Workers West, voted this week to authorize a strike. The vote gives union leadership the go-ahead to call a walkout if labor negotiations with contractors fail to produce an acceptable deal.
ChatGPT, Claude, and Grok all went down Thursday morning within roughly an hour of each other, though OpenAI and Anthropic did not cite any shared external cause. OpenAI blamed a routing error resolved within about 34 minutes, Anthropic reported elevated errors across several Claude models fixed within roughly three hours, and xAI said its Grok outage stemmed from a failure at its Memphis compute facility that took nearly four hours to resolve.
OpenAI unveiled GPT-6 Astra, touting gains in cybersecurity, software engineering and agentic workflows, and calling it its most capable model yet. The release is rolling out first to enterprise and security customers before reaching Plus, Pro and Business users. In the same week, Anthropic, Meta and Google each pushed out their own model updates, with Anthropic releasing Claude Fable 5.1 and Mythos 5.1.
At IFA 2026, AWOL Vision introduced the LuxVision, an ultra-short-throw projector it claims is the first to combine native 4K resolution, 120Hz refresh rate, and optical lens shift. The unit uses an RGB triple-laser light source rated at 6,000 ISO lumens and a 0.18:1 throw ratio, letting it cast a 120-inch image from just 7.5 inches away, with screen sizes scaling up to 200 inches. It ships around October starting at $6,499.
OpenAI has released GPT-6 Astra, a model trained on over 100,000 GPUs at its Texas Stargate site, marking the first time OpenAI used other AI models to help supervise training. President Greg Brockman described it as a 'generational leap' and hinted it could mark the arrival of artificial general intelligence. The model will roll out first to select partners in OpenAI's Daybreak Access program before reaching ChatGPT Plus, Pro, Business, Enterprise and API users.
OpenAI's unreleased GPT-6 Astra model scored 62.7% on the ARC-AGI-3 Semi-Private benchmark using a Standard harness for $26,000, and 99.9% using a Provider Adapter harness that preserves reasoning state between requests for $19,000. The model also beat the median human tester's action efficiency on 96% of the benchmark's levels, building compact symbolic world models and its own shorthand notation to track game states.
Thinking Machines, the AI lab started by former OpenAI CTO Mira Murati, is reportedly negotiating a $1 billion funding round that would value it at $40 billion or more, with existing investor Accel potentially leading the deal. That figure would mark a step down from the $50 billion valuation the startup had sought late last year, though it still dwarfs the company's reported annual revenue run rate of over $100 million.