Nvidia open-sources Flora to make robot video labeling 125x cheaper than GPT-6 Astra
Nvidia's new data engine Flora labels one hour of robot video for 72 cents instead of $88 with GPT-6 Astra, while scoring slightly higher on quality. The open system, built on three fine-tuned Qwen3.5 models, catches errors in robot actions far more reliably than any competitor but sometimes falls short on descriptive accuracy. Nvidia is positioning Flora as a key piece of its strategy to control the entire robotics value chain. The article Nvidia open-sources Flora to make robot video
Get smart on it
Large AI models that train robots need videos annotated with three types of labels: descriptions of actions, step-by-step performance ratings, and task completion judgments. An open-source labeling system called Flora accomplishes all three at a projected cost of 72 cents per video hour, compared to roughly $88.50 with a commercial alternative, though these are estimates rather than actual production costs. Flora outperforms the commercial alternative at detecting where actions begin and end and is better at catching partially failed tasks, though the commercial model sometimes produces more precise descriptions and scores better on certain benchmarks. Nvidia is releasing Flora openly to help researchers scale robotics datasets and to keep developers using its hardware.
Relevant courses
Want to go deeper than the news? Explore live, cohort-based AI courses taught by practitioners.
Browse AI courses on MavenRelated news
OpenAI's latest AI models are coming to Google Cloud and Antigravity
Google would certainly prefer that everyone used its Gemini models with Google Cloud services, but soon, you'll have more options. OpenAI and Google have come to an arrangement that will see the former's frontier models arrive in Google Cloud services. Perhaps more importantly, GPT models will soon appear in Antigravity, Google's agent-powered coding tool. For enterprise customers and AI enthusiasts, OpenAI's latest model will appear in Google's "Model Garden" in the coming weeks, adding to its
Also reported by The Decoder and SiliconANGLE
Apple releases iOS and macOS 27.2 with new Siri AI features, new Music logo
Today, Apple rolled out the 27.2 updates for its operating systems, including iOS, macOS, watchOS, tvOS, visionOS, and so on. Let's start with iOS 27.2. This update is heavily focused on Siri AI. It adds a memory feature to Siri, so it can "recall your past conversations" and reference them when chatting with you. It also expands what data sources Siri can access when giving responses to include voicemails, Messages attachments, and more.Read full article Comments
Anthropic says one AI agent beats many subagents now that models can handle complexity alone
Anthropic says companies should evaluate whether they can consolidate sub-agents into a single agent that loads context on demand. In one architecture comparison, the single-agent setup ran about twice as fast. Anthropic has published a migration kit on GitHub to help teams make the transition. The article Anthropic says one AI agent beats many subagents now that models can handle complexity alone appeared first on The Decoder.
Claude Duet app is Anthropic’s bet on agents
Claude Duet is available through waitlist for Pro and Max plan users
Also reported by The Verge and SiliconANGLE
Anthropic Launches OSS Scanner, a Free AI Vulnerability Scanner
Anthropic has launched OSS Scanner, a free, opt-in vulnerability scanner for critical open-source projects. The Anthropic OSS Scanner runs the company’s strongest models, including Claude Mythos, against enrolled codebases. Reports go straight to maintainers without human review. That trade delivers faster findings, but shifts triage work onto project teams. 11 days after the October 8 launch, Anthropic reported 77 critical and 803 high model-rated findings. TL;DR Scale: 114 projec
OpenMobile-2 proves open-source smartphone agents can rival proprietary models
Smartphone agents don't get better by processing more data. They improve by training on as many different apps as possible. The researchers behind OpenMobile-2 built an open training environment with more than 110 applications, including 35 replicas of popular apps. On real Android devices, their model scores highest among open agents, the authors say. The article OpenMobile-2 proves open-source smartphone agents can rival proprietary models appeared first on The Decoder.