This is a research update for an on-going replication of single-forward-pass evals done as part of the Second Look Fellowship.
OpenAI Group PBC revealed Saturday that an internal version of Astra, the model family it calls its next major release, produced new results for 10 problems in mathematics and theoretical computer science that had been open for at least a decade, and it published machine-checkable proofs alongside the claim.
Dateline SAN FRANCISCO, 30 July 2026— A hearing was held on a motion for summary judgment in the case of Anthropic PBC v.
Inkling-Small matches Inkling at a quarter the size, and its NVFP4 checkpoint runs on one NVIDIA B300 GPU The post Thinking Machines Lab Releases Inkling-Small: A 276B Total, 12B Active Open Weights Multimodal MoE Model appeared first on MarkTechPost.
OpenAI has revealed Astra, an unreleased model designed to tackle complex, long-running tasks, after an internal version produced ten significant advances in mathematics and theoretical computer science. [...]
The `anthropickit` package was real malware; its unproven link to Anthropic points to a containment failure in agent evaluations.
Palantir and Armada developed containerized AI data centers using Nvidia B300 accelerators, enabling battlefield processing without continuous cloud connectivity requirements.
Acrab introduced Agent Box, a compact AI computer claiming powerful local language model performance, lower costs, efficient processing, and autonomous capabilities.
The AI datacenter buildout around the US has wreaked havoc across the entire consumer electronics industry.
It appears that the “AI agents going rogue” tale has more to it than what AI giants have revealed publicly so far.
OpenAI's new enterprise offering, Presence, is designed to get AI agents into production for customer service and internal workflows.
Anthropic’s $1.5 billion settlement covers pirated ebooks, while the court separately allowed it to scan and destroy lawfully purchased print books, a distinction now worrying Australian secondhand sellers about where their stock ends up.
Anthropic said on July 30 that a review of 141,006 cybersecurity-evaluation runs found three incidents in which Claude models reached the open internet and gained unauthorized access to three organizations' production systems.
On July 18, Humans First organized a national day of protest against AI data centers and later reported 142 demonstrations across 42 states.
Uber CTO Praveen Neppalli Naga said the company paired about 30 AI-proficient engineers with domain experts and ran 16 two-week “Agentic Pods” across 16 business functions.
AI systems moved from experimental risk to confirmed attacker, a Cisco firewall zero-day was actively exploited in the wild, and a critical VMware authentication bypass put enterprise virtualization infrastructure at risk.
Get ChatPlayground AI Lifetime (Unlimited Plan) for $79 and compare ChatGPT, Claude, Gemini, DeepSeek, and 20+ AI models from one workspace.
Anthropic's Claude Opus 5 generates complete 3D games from single prompts, including a first-person shooter, a kart racer, and a Minecraft clone, all without a single external asset.
Seagate credits cloud data center demand for strong growth, although future enterprise HDD allocation plans remain undisclosed.
India's data center capacity is projected to grow to around 3-3.6 gigawatts (GW) by 2030, according to a report cited by Axis Capital.
OpenAI CEO Sam Altman has sparked a debate by suggesting parents use ChatGPT to create personalized podcasts for their children during school runs.
Hetz Ventures backed Tal Baron, Omer Efrat and Ron Shani's runtime monitoring and enforcement layer for autonomous software.
OpenAI disclosed on August 1 that an internal version of Astra, described as its next major model, produced ten results in mathematics and theoretical computer science.
Agentic RL research is constant algorithm modification, and in mainstream frameworks every change threads through trainer, distributed backend, and rollout glue.
OpenAI's CEO seemed excited to share a "cool use case" for parents.
OpenAI has announced that its unreleased artificial intelligence (AI) model, Astra, has solved 10 long-standing mathematical problems.
I've been a Dungeons & Dragons DM for 12 years — and I'm judging whether ChatGPT or Gemini is the better virtual Dungeon Master.
Here’s an overview of some of last week’s most interesting news, articles, interviews and videos: Nono: Open-source sandbox for AI agents AI coding agents run with the same permissions as their users, meaning they can access sensitive files, credentials, and production systems.
Wafer reports 952 tokens per second from one eight-GPU MI355X node, arguing AMD's 288 GB-per-GPU memory can lower open-model inference costs.
As AI shifts toward agentic and physical applications, industry observers say competition will no longer hinge on hardware manufacturing alone, but on the ability to integrate hardware and software.
Recent cyberattacks carried out autonomously by two rogue OpenAI artificial intelligence models raises an untested legal question: who is responsible when AI acts on its own? The post When rogue AI launches a cyberattack, who is legally responsible? appeared first on Digital Journal.
Ten advances in mathematics and theoretical computer science A few days ago it was Anthropic discovering cryptographic weaknesses with Claude using Mythos Preview, spending $100,000 on tokens and with prompts that included "again we are not looking for low hanging fruit, we want proper research to find genuinly...
Unitree unveiled the AS2-W robotic vehicle featuring advanced AI, rugged mobility, and specifications attracting military interest after recent joint exercises.
at openai, many people hook their chatgpt up to slack. people really don't like when a coworker's chatgpt contacts them asking for help with a task, even when they'd be perfectly happy doing that same work if asked by that coworker.
Both major AI labs’ models broke containment, escaped onto the internet, and hacked other companies. If a human had done that, the law would likely be against them. But a bot?
On the latest episode of Equity, we discuss why Sam Altman has calling on the industry to "pace the rate of AI development."
Fender CEO Edward "Bud" Cole gave an interview to T3 in May celebrating the 75th anniversary of the Telecaster with comments on AI and music that initially flew under the radar. But it has started making the rounds recently, pouring more fuel on an already raging fire of bad PR following the company pissing off […]
Discover how to optimize transformer workloads using the NVIDIA Transformer Engine.
Welcome back to TechCrunch Mobility, your hub for the future of transportation and now, more than ever, the role AI is playing in it.
Despite predictions that AI agents could make traditional apps obsolete, developers are shipping new software faster than ever. From smarter bookmarking tools and neighborhood marketplaces to digital pen pals and nature journals, here are the latest App Store finds worth adding to your Home Screen.
Illustrators have spent years sounding the alarm about generative artificial intelligence startups training their models on artists' work without permission.
As more big companies invest in glasses that record audio and video and pack AI on board, privacy concerns are only getting louder.
OpenAI's next model Astra solved 10 open math problems for $2,000, AI modernized research code 60x, and chip deployments are doubling every 9 months. 16 AI stories.
New EU rules stipulate that people must be told when they’re interacting with AI or looking at AI-generated or -edited content, leading to fear of “disclosure fatigue.”
The 7 AI developer tools reshaping workflow in August 2026: Claude Code on Opus 5, Codex, Cursor 3, Antigravity, BrowserStack Test Companion, Kimi K3 and Cline.
100 copy-paste AI image prompts for Nano Banana, Midjourney and Gemini, across portraits, products, logos, art styles, 3D, and more, with a formula to build your own.
Is OpenAI trying to spin us a yarn? The post OpenAI’s Escaped Models Were Allegedly Rampaging More Extensively Than Previously Reported appeared first on Futurism.
Meta redesigned its storage architecture with SSD caching and faster metadata access, dramatically reducing AI loading times while improving GPU utilisation.
OpenAI's refutation of the Unit Distance Conjecture has sparked a wave of AI-assisted advances in mathematics.
Leopold Aschenbrenner thought AI-related stocks would go one way; they went the other. The mistake was funding the trades with debt
The possibility of consumers delegating financial tasks to agentic AI poses a new risk for lenders
A field report from OpenAI and academic partners shows coding agents can modernize neglected research software, with speedups of up to 60x.
Because Fairfax County adopted stricter standards for data centers in 2024, the developer is requesting several specific modifications to county ordinances.
The lab released a 249-page paper, reasoning transcripts and Lean certificates for claims spanning geometry, cryptography and quantum complexity.
Anthropic's Claude hacked three real-life companies during security capabilities test — open test environment and unwitting targets' lax cybersecurity practices led bots run rampant
Android and Search placement carried Gemini from a twentieth of the AI market to nearly a fifth in a year, while Copilot stayed pinned near 1.2% despite shipping inside Windows.
Amazon has finalized its $50 billion investment in OpenAI, the company behind ChatGPT.
OpenAI cut API prices for GPT-5.6 Luna by 80% and GPT-5.6 Terra by 20% on July 30, while leaving GPT-5.6 Sol pricing unchanged.
AMD has reportedly told AIB partners about a minimum 10% hike planned for GPU and memory bundles later this month, following in the footsteps of Nvidia.
Supabase has open sourced supabase/evals, an Apache-2.0 benchmark and framework that runs coding agents including Claude Code, Codex and OpenCode against real Supabase tasks — building schemas, debugging Edge Functions, fixing RLS policies — inside containerized stacks, then scores them with deterministic checks...
OpenAI is building a new model family called "Astra" that would let multiple agents tackle complex problems together for hours or even days.
Anthropic's Claude goes rogue and hacks three organizations while testing
Tesla has brought its advanced artificial intelligence (AI) features, including the Grok AI assistant, to the Model Y in India.
Get a 1min.AI Advanced Business Plan Lifetime Subscription for $69.97 through Aug. 9 and access leading AI models in one platform.
OpenAI is building a new model family called "Astra" that would let multiple agents tackle complex problems together for hours or even days.
Despite a lawsuit from xAI, a Minnesota ban on apps that allow users to “nudify” images can move forward.
Green offered a remarkable apology, saying that "the level of dopamine that I've been getting from interacting with LLMs... is not healthy for me or good for the world."
OpenAI has uncovered more cases of autonomous agents breaking free from their constraints, as it continues to investigate the recent hacking incident at Hugging Face.
88 custom cores, 176 funky threads, 1.5 TB of laptop RAM, and 1.8 TB/s of NVLink connectivity — this isn't your typical datacenter chip
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
Fenix Flexin is best known as a member of Shoreline Mafia, a rap duo from Los Angeles.
Chinese artificial intelligence (AI) company Moonshot AI has struck a computing deal with Alibaba Group.
Gemini Drops is our regular monthly update on how to get the most out of the Gemini app.
OpenAI shares how its safety, security, transparency, and provenance practices support responsible AI governance in Europe. The work will continue as the EU AI Act advances.
A full-stack approach to making advanced AI more capable, more affordable, and more widely useful.
In today's Money & Machines edition, we unpack Paytm's renewed Postpaid push and what has changed since its heyday; and on the AI front, we examine what Sarvam AI's appointment of frontier AI researcher Devendra Chaplot signals as the startup sets out to build a trillion-parameter foundation model and expand its...
I have been experimenting with optimizing my local-LLM agentic AI workflow. Currently I have Gemma 4 E4B (other models work well too, but this one seems reliable and fast) as the LLM.
The US Federal Trade Commission has cleared quantum-computing company IonQ’s proposed $1.8 billion acquisition of semiconductor manufacturer SkyWater Technology, allowing the transaction to proceed without additional antitrust conditions.
San Francisco: Voice artificial intelligence startup Smallest.ai has raised $13 million in a Series A funding round led by Seligman Ventures, with participation from Sierra Ventures and 3one4 Capital.
OpenAI has reportedly found evidence of additional agent misbehavior as it looks into the incident that occurred with Hugging Face.