NAll news

All news

Friday39 items

  1. heise onlineAnalysis
    AI score58

    Artificial Analysis releases Intelligence Index v5, with 60% of tasks now undisclosed

    Independent benchmarking organization Artificial Analysis has updated its main index to Intelligence Index v5. The index aggregates results from nine benchmarks and, for the first time, uses 60% undisclosed tasks to reduce the impact of models overfitting to public questions. The two newly added tests are Terminal-Bench-Science, which contains 70 agentic science tasks, and FrontierCode 2.0, recently launched by Cognition, which examines whether repository maintainers would adopt an agent's changes in real coding tasks; Humanity's Last Exam Diamond, AA-Omniscience v1.1 and CritPt v2 were upgraded, while GDPval-AA v2.1, SciCode and AA-LCR v1.1 were removed. Rankings changed after the test adjustments: Claude Opus 5.5 briefly led before being overtaken by the recently released Gemini 4 Carbon, Grok 4.7 and Kimi K3.1 Pro saw their scores decline, and the gap between the strongest open-source model and the top spot widened from 3 points to 12 points. The original article also notes that undisclosed benchmarks cannot be externally reviewed, making potential errors hard to detect.

    Same story · 8 reports

    #Evaluations / benchmarks#English text is being prepared.#Agents

  2. Google Developers BlogFirst-partyAnalysis
    AI score70

    Google A2UI Team and CopilotKit Explain How to Build Generative UI Applications

    The Google A2UI team and CopilotKit CEO Atai Barkai explain how to build generative UI applications using the A2UI open standard and CopilotKit. A2UI describes interfaces with declarative JSON, and the client only renders components from a trusted component catalog; the same response can be rendered to React, Angular, SwiftUI, Jetpack Compose, and Flutter. The article provides a code example for integrating CopilotKit with Google ADK, and explains the bidirectional flow, implemented via the AG-UI protocol, in which user actions are sent back and the agent regenerates the interface.

    #Tutorials / practice#Agents#English text is being prepared.

  3. Official Pydantic AI articleFirst-partyAnalysis
    AI score68

    Pydantic compares code mode and tool calling across 5,276 runs

    Michael Pfaffenberger of Pydantic used 5 small Pydantic AI applications and 8 models to conduct 5,276 scored runs, comparing code mode with direct tool calling. On long tasks, code mode reduced average cost by 58%, latency by 51%, and increased success rate by 16 percentage points; however, for short tasks with 5 items, 3 out of 5 tasks were instead 21% to 38% more expensive. The author argues that models are better at writing loops than at acting as the loop, and the gap widens as tasks get longer.

    Michael Pfaffenberger avatar

    #Evaluations / benchmarks#Agents#English text is being prepared.

  4. Latent SpaceAnalysis
    AI score54

    Elorian Releases DecoyCount Benchmark: Frontier Models All Score Below 50% on Visual Counting Accuracy

    Elorian and Mercor jointly released the DecoyCount benchmark, which tests whether models truly count objects or rely on visual shortcuts. Frontier models all scored below 50% accuracy, with the highest being GPT-6 Astra at 43.17%. Elorian says its visual reasoning model Glimpse scored 50.48% in a single run on 1,660 questions, making it the only model to exceed 50%, but this result comes from internal testing and did not use an agent or tool framework, whereas the six models in the paper each took the average of three runs. Elorian CEO Andrew Dai said the benchmark examines both vision and reasoning, and can distinguish whether a model failed to see the objects or miscounted them.

    #Evaluations / benchmarks#Multimodal#Reasoning

  5. Beijing NewsAnalysis
    AI score67

    Beijing News tests Qwen Go! and Lingguang 2.0: personal agent competition shifts toward autonomous task handling

    Beijing News Shell Finance reporters tested Alibaba's Qwen Go! and Ant's Lingguang 2.0. The former set up a daily 8 a.m. medication reminder with a single sentence and checked it item by item, while the latter, after authorization on Alipay, sorted out the past six months of bills. The article also reviews the layouts of Meta Muse, OpenAI dots, as well as ByteDance and Tencent, and points out that cross-ecosystem permissions are currently the hardest wall to break down, with ByteDance's SAEP published rejection list reaching 155 apps.

    #English text is being prepared.#Agents#OpenAI

  6. Baseten Engineering BlogFirst-partyAnalysis
    AI score56

    Baseten Optimizes Qwen3-ASR Batch Transcription, Boosting Single-GPU Throughput by 7x

    The Baseten engineering team, through native audio decoding, KV cache FP8 quantization, and reduced host-device synchronization, increased batch transcription throughput for Qwen3-ASR 1.7B on a single RTX PRO 6000 by 7x and reduced latency by 6.9x. After optimization, the transcription cost for one hour of audio is approximately $0.0013, i.e., 1/10 of a cent per audio hour, with no significant change in word error rate. The optimized version has been made the default deployment in the Baseten Model Library and can be called through an OpenAI-compatible transcription client.

    How Baseten gets 7x throughput on Qwen3-ASR batch transcription

    #Tutorials / practice#English text is being prepared.#Speech

  7. Sakana AI：BlogFirst-partyAnalysis
    AI score54

    Sakana AI uses AI to analyze the US midterm elections: about 710,000 X reactions to the Texas Senate race

    Sakana AI has released an AI analysis case study of the US midterm elections, focusing on the Texas Senate race and analyzing 2,042 posts and 705,976 reactions on X from two candidates and two state party organizations. Its narrative analysis system extracts narratives according to "who did what and how it was evaluated," and cross-checks them against the original text using multiple AI models; meanwhile, it uses the product Sakana Marlin to retrieve public information such as polls, campaign funding, prediction markets, and reports, and compares it with observations on X. The article states that most of the support Paxton received before the runoff came from intra-party mobilization, that after the runoff his support share fell from about 50% to just under 20%, and that the shift in evaluations of the two candidates preceded judgment adjustments by institutions such as the Cook Political Report by about two months.

    #English text is being prepared.#English text is being prepared.#Reasoning

  8. Smart City Industry AnalysisAnalysis
    AI score54

    A Beauty Company's FDE Implementation in Practice: An Eight-Step Pipeline for Cleaning Massive Multimodal Data into Trustworthy Data

    A beauty company's FDE implementation project processes text, image, audio, and video data collected from Xiaohongshu, Douyin, WeChat Channels, and Weibo through eight steps: defining data specifications, preserving original content, mapping data relationships, multimodal normalization, conflict and quality cleaning, generating structured data, performing acceptance checks, and ingesting according to rules and versions. The original article emphasizes that cleaning results must not overwrite the original text, and that OCR, ASR, and model labels are only added results; it also describes the practice of tracing back to the original evidence when ASR and OCR values conflict, and placing cases that cannot be confirmed into a pending-verification queue. This pipeline has also been organized by the author into a general solution reusable in scenarios such as consumer goods insights, automotive word-of-mouth, restaurant site selection, and brand public opinion.

    #Tutorials / practice#English text is being prepared.#Multimodal

  9. Hacker News TrendingAnalysis
    AI score76

    Fix and Troubleshooting for the Linux IPv4 TCP Handshake Use-After-Free Vulnerability CVE-2026-98432

    The Linux stable kernel fixed a use-after-free vulnerability, CVE-2026-98432, in the IPv4 TCP handshake path on October 11 and 15. The kernel CNA rated it CVSS 3.1 9.8 Critical, affecting versions 4.4 and later. As of October 29, kernels for all Ubuntu versions are still marked as affected, and the default kernels of RHEL 8, 9, and 10 and Amazon Linux 2 have no fix yet. Anthropic said its Claude model wrote, in a lab setting, remote root exploit code targeting servers with a multipath TCP listener, but did not disclose it.

    #Safety / alignment#Open-source ecosystem#English text is being prepared.

Wednesday20 items

  1. Tencent Cloud Developer CommunityAnalysis
    AI score57

    From ReAct to Supervisor-Worker: Core Concepts and Architecture Choices of AI Agents

    The article systematically reviews the core concepts and architecture choices of AI Agents, pointing out that the essential difference between an Agent and an ordinary large-model call lies in having a closed loop composed of goals, perception, decision-making, and action. It compares the problems each of the three single-Agent modes—ReAct, Plan-and-Execute, and Reflexion—solves and their costs, and explains the division-of-labor principles of Supervisor-Worker in multi-Agent scenarios. The author also lists engineering pain points during implementation, such as state persistence, termination circuit breaking, context engineering, and observability, and provides a decision sequence for selection starting from a single Agent.

    #Tutorials / practice#Agents#Reasoning

  2. Shao Meng
    AI score51

    Replicating Top-Tier iOS App Design with a Combination of Three Agent Skills

    The author introduces an approach to replicating top-tier iOS app design using three Agent Skills: Appllama Skills provides design research and Expo/RN native-feel build standards, Expo Skills offers official best practices such as project structure, Router, animations, and EAS publishing, while Emil Kowalski Skills turns motion and interface taste into executable rules. The three respectively answer whom to learn from, how to do the engineering, and how to keep motion details from feeling cheap; the medium is skills.sh and Agent Skills, rather than the traditional reverse-engineering IPA toolchain. Appllama's MCP is billed under Pro, and the design skill can be installed separately; Expo Skills can be used in Claude Code and Codex via the official plugin, while other Agents use the skills CLI.

    Quoted postjaimin · @jaimintf

    reverse engineer any ios app. app design → https://t.co/rhCq0WZIve best practices → https://t.co/BaQpMSbiAt animations → https://t.co/YZVV8SagDO https://t.co/OzB1WoaIu7

    #English text is being prepared.#Agents#English text is being prepared.

  3. Alibaba Cloud Developer Community AI
    AI score55

    Encapsulate repetitive API testing rules into SKILL.md so AI generates test cases to standard

    Using the example of team testers who repeat API testing rules to AI every day, the author wrote roughly 300 words of rules—covering logging in to get a token, three-layer assertions, and an exception scenario checklist—into SKILL.md. After encapsulation, one only needs to say "use api-test-standard to test this API," and the assertion layers and exception coverage stay consistent, while colleagues can also reuse the same set of rules. The article presents a four-step encapsulation method: record the most frequently repeated words verbatim; split into preconditions and execution steps, assertions, and exceptions; turn vague statements into executable rules; and add a self-check checklist at the end. It also lists five pitfalls, including writing it as a colloquial chat log, SKILL.md exceeding 500 lines, writing too empty a description, not verifying after writing, and only encapsulating without iterating.

    #Tutorials / practice#Agents#English text is being prepared.

  4. Alibaba Cloud Developer Community AIAnalysis
    AI score60

    How SmartCall Implements a Dual-Channel RAG Knowledge Base Within a Single Phone Call

    The SmartCall open source project breaks down the implementation of an agent knowledge base in voice customer service scenarios, using two parallel channels: interceptor gating and Function Calling tools. The article points out that the server-side VAD of end-to-end real-time voice models will answer preemptively, so the core contradiction shifts from retrieval accuracy to whether retrieval can beat automatic model responses; the solution relies on fail-open and audio hold as its pivots. The article also provides a four-step extension process for integrating new sources such as RAGFlow and a ten-item pitfalls checklist.

    #Tutorials / practice#RAG#Speech

  6. Ethan MollickAnalysis
    AI score55

    Ethan Mollick tests Genie 4 generating a Victorian afternoon tea scene in real time

    Ethan Mollick used Genie 4 to generate a Victorian afternoon tea scene in real time, entering commands such as "a mouse in formal attire," "a porcupine walks by," "the mouse grows wings," and "the mouse flies toward the teacup," and said he could walk freely around the scene. He stated that the visuals still have flaws, but that the progress compared with two years ago is enormous.

    Same story · 2 reports

    #English text is being prepared.#Multimodal#Video

  7. Artificial Analysis full articleFirst-partyAnalysis
    AI score66

    Artificial Analysis releases Intelligence Index v5, with 60% of evaluations switched to private datasets

    Artificial Analysis releases Intelligence Index v5, adding two new benchmarks, FrontierCode 2.0 and Terminal-Bench-Science, upgrading HLE-Diamond, AA-Omniscience v1.1, and CritPt v2, with 60% of evaluations based on private test sets or held-out answers. Claude Opus 5.5 (max) ranks first with 58 points, followed closely by GPT-6 Astra (max) at 57 points and GPT-6.1 Sol (max) at 55 points, with Gemini 4 Argon (high) at 50 points. The gap between open-weight models and the leaders has widened from 3 points to 12 points; the released-weight DeepSeek V4.1 Flash 1017 (max) and GLM-5.3 (max) both score 32 points. The per-task cost of leading models differs by more than 13 times: GPT-6.1 Terra (max) is $0.96, while Claude Fable 5.1 (max) is $13.04. A Progress Index v1 was also released, placing models released since November 2022 on the same capability scale for comparison.

    Same story · 8 reports

    #Evaluations / benchmarks#Agents#English text is being prepared.

  8. Artificial AnalysisAnalysis
    AI score65

    Artificial Analysis releases Intelligence Index v5, with 60% of evaluations based on private datasets

    Artificial Analysis releases Intelligence Index v5, adding two new benchmarks, FrontierCode 2.0 and Terminal-Bench-Science, upgrading HLE-Diamond, AA-Omniscience v1.1 and CritPt v2, removing GDPval-AA v2.1, SciCode and AA-LCR v1.1, and changing 60% of the index to evaluations based on private test sets or held-out answers. It also launches Progress Index v1, which uses a unified scale to track changes in model capabilities over time. The results show that Claude Opus 5.5 (max) ranks first with 58 points, followed by GPT-6 Astra (max, 57) and GPT-6.1 Sol (max, 55), while Gemini 4 Argon (high) scores 50; the gap between open-weight models and the leaders has widened from 3 points to 12 points, and among released weights, DeepSeek V4.1 Flash 1017 (max) and GLM-5.3 (max) lead with 32 points. The per-task cost of the top ten models differs by more than 13 times, with GPT-6.1 Terra (max) lowest at $0.96 and Claude Fable 5.1 (max) highest at $13.04.

    Same story · 8 reports

    #Evaluations / benchmarks#Agents#English text is being prepared.

  9. A Yi AI Notes
    AI score55

    The author tested an email Agent on 5,039 emails and includes 5 scenario prompts

    The author ran their own QQ Mail inbox of 5,039 emails through the email Agent on mobile; of 69 emails from the past month, about 20 needing action were filtered out in 43 seconds. The article provides prompts for 5 common scenarios, plus a memory template of "user profile + 10 behavioral guidelines," usable by copying without writing any code. Core scenarios include finding hidden AI auto-charges and SaaS renewals in 17 seconds, and merging multiple banks' Chinese and English monthly statements and calculating foreign currencies separately in 30 seconds.

    Quoted postAYi · @AYi_AInotes

    https://t.co/BSzz78Spop

    #Tutorials / practice#Agents#English text is being prepared.

  10. Shao Meng
    AI score64

    Blake Anderson open-sources ai-app-playbook, using a Coding Agent to build consumer-grade apps on a fixed stack…

    Blake Anderson has open-sourced "ai-app-playbook," an operating playbook for coding agents to execute, covering product selection, building, launch, and first-month profitability. It has two layers: a human-facing playbook/00–13 chapters and an Agent-facing AGENTS.md (Phase 0–14, with each step marked [AGENT] or [HUMAN]); before starting, one must pass the Gate A scoring. The tech stack is fixed on Cloudflare Workers, D1, Better Auth, OpenRouter routing to DeepSeek, Superwall plus Stripe, etc., and explicitly lists a long set of non-choices such as Vercel, Supabase, and Next.js.

    Quoted postBlake Anderson · @blakeandersonw

    Here's my current tech stack for scaling consumer apps in 2026. It's by far the best & cheapest stack for agent-native development ⬇️ https://t.co/0yqdwDeCp5

    #English text is being prepared.#Agents#English text is being prepared.

  11. Vaibhav SrivastavAnalysis
    AI score53

    GPT-6.1 Terra beats Astra on vision benchmarks, roughly 47% faster and about 92% cheaper

    Vaibhav Srivastav, citing SkalskiP's vision benchmark results, said that GPT-6.1 Terra beats Astra on text, object detection, and counting tasks, while also being roughly 47% faster and about 92% cheaper. The original post includes leaderboards and example links.

    Quoted postSkalskiP · @skalskip92

    GPT-6.1 Terra beats Astra in our vision benchmarks wtf OpenAI? had to double-check the results - text - object detection - counting also ~47% faster and ~92% cheaper ↓ leaderboards and examples https://t.co/4AvboaD16g

    #Evaluations / benchmarks#Multimodal#OpenAI