English translation is being prepared.
Matched fieldsSummary
←Search results
Search “Hugging Face”
Baidu Apollo Go Opens Test Rides in London and Hong Kong
In October, Baidu Apollo Go invited media for the first time to test ride the RT6 in London, and together with FREENOW by Lyft plans to achieve fully driverless operation next year and launch commercial services in 2027; during the same period, it launched Hong Kong's first large-scale public test rides in Tung Chung, providing free rides to more than 1,000 residents from October 21 to November 3, after having already driven 280,000 kilometers safely in Hong Kong. Apollo Go also received the 2026 Self-Driving Industry Awards Consumer Service Award.
In October, Baidu Apollo Go invited media for the first time to test ride the RT6 in London, and together with FREENOW by Lyft plans to achieve fully driverless operation next year and launch commercial services in 2027; during the same period, it launched Hong Kong's first large-scale public test rides in Tung Chung, providing free rides to more than 1,000 residents from October 21 to November 3, after having already driven 280,000 kilometers safely in Hong Kong. Apollo Go also received the 202
Matched fieldsBody
Latent Space announces AIE CODE coding agent conference curation themes
Latent Space host swyx introduces the curation approach for the annual flagship conference AIE CODE, which is single-track, invitation-only, and aimed at about 1,000 agent engineers. This year's three themes are Scaling without Slop, System One models, and Increasing Ambition, among which Jarred Sumner will share Bun's rewrite in Rust and a Bun AOT preview. The article also rounds up that day's AI news, including Mistral's open-weight release and OpenAI's frontier model monitoring.
Matched fieldsBody
Edge0 Releases Audio8-TTS-Realistic: A 0.6B Open-Source Voice Cloning TTS Model
Edge0 has released the open-weight voice cloning TTS model Audio8-TTS-Realistic, with 601M parameters, supporting 23 languages, with weights and code released under Apache 2.0. The model requires only 1 reference audio clip and its corresponding text for zero-shot voice cloning, and can also directly generate speech when no reference voice is available. Self-reported results show an average error rate of 4.86% on the CV3 nine-language benchmark, lower than Qwen3-TTS-1.7B's 7.73%, but it lags behind 4B-class models on the MiniMax benchmark, with Thai WER reaching 48.57%.
Edge0 has released the open-weight voice cloning TTS model Audio8-TTS-Realistic, with 601M parameters, supporting 23 languages, with weights and code released under Apache 2.0. The model requires only 1 reference audio clip and its corresponding text for zero-shot voice cloning, and can also directly generate speech when no reference voice is available. Self-reported results show an average error rate of 4.86% on the CV3 nine-language benchmark, lower than Qwen3-TTS-1.7B's 7.73%, but it lags beh
Matched fieldsBody
Edge AI Daily Morning Brief: Google Releases Gemini 4 Carbon, Anthropic Files for $2 Trillion...
Google DeepMind released the cybersecurity frontier model Gemini 4 Carbon on October 28, topping the Artificial Analysis Intelligence Index v5.0 with a composite score of 59, priced at $2 per million input tokens and $10 per million output tokens. Anthropic filed its S-1 prospectus with the SEC on October 26, planning to list on Nasdaq under the ticker ANT, with a target valuation of $1.5 trillion to $2 trillion and fundraising of up to $100 billion. In the same period, Microsoft disclosed for the first time that Azure quarterly revenue reached $33.2 billion, up 48% year-over-year.
Google DeepMind released the cybersecurity frontier model Gemini 4 Carbon on October 28, topping the Artificial Analysis Intelligence Index v5.0 with a composite score of 59, priced at $2 per million input tokens and $10 per million output tokens. Anthropic filed its S-1 prospectus with the SEC on October 26, planning to list on Nasdaq under the ticker ANT, with a target valuation of $1.5 trillion to $2 trillion and fundraising of up to $100 billion. In the same period, Microsoft disclosed for t
Matched fieldsBody
OpenAI expands its misalignment monitoring system to cover training, evaluation, and external deployment
On October 29, 2026, OpenAI announced an expanded multi-layer misalignment behavior monitoring system covering training, evaluation, and internal use, and, after modifications, extended to external deployment. The system monitors every reinforcement learning run and evaluation for models at least as capable as GPT-6 Astra, with a four-layer architecture that escalates step by step from activation probes to safety and alignment investigation agents. In independent red team testing conducted by METR and Redwood Research from July 21 to August 11, two client-controllable request headers that could disable the new agent monitoring were found, along with a compression event gap from May to July 2026 in which approximately 10% of eligible internal Codex traffic was not monitored.
On October 29, 2026, OpenAI announced an expanded multi-layer misalignment behavior monitoring system covering training, evaluation, and internal use, and, after modifications, extended to external deployment. The system monitors every reinforcement learning run and evaluation for models at least as capable as GPT-6 Astra, with a four-layer architecture that escalates step by step from activation probes to safety and alignment investigation agents. In independent red team testing conducted by ME
Matched fieldsBody
OpenAI Updates Model Spec, Listing Loyalty to Human Intent as a Core Principle
OpenAI has updated its Model Spec document, restructuring the framework into five principles: loyalty to human intent, reducing harm, honesty and transparency, safeguarding people's right to make their own choices, and prudence and diligence, with loyalty to human intent as the foundation of the entire framework. OpenAI says its current models cannot yet fully comply with the new version of the spec, and the team is researching new training methods to help models internalize these principles. This adjustment is similar to Anthropic's constitutional AI approach.
OpenAI has updated its Model Spec document, restructuring the framework into five principles: loyalty to human intent, reducing harm, honesty and transparency, safeguarding people's right to make their own choices, and prudence and diligence, with loyalty to human intent as the foundation of the entire framework. OpenAI says its current models cannot yet fully comply with the new version of the spec, and the team is researching new training methods to help models internalize these principles. Th
Matched fieldsBody
Ant inclusionAI Open-Sources the UBTree Project Page: Using Unigram and…
Ant inclusionAI has launched the UBTree project page on GitHub, introducing a speculative decoding method that uses Unigram and Bigram models for parallel tree drafting (arXiv:2609.39972). The page includes demo videos of Ling3 Tiny and Flash, comparing four methods—AR, DFlash, DSpark, and UBTree—under the same target model; the videos are 1440×760, 20 FPS, at 0.25× playback speed. The figures and table numbers are all driven by static/js/data.js.
Matched fieldsBody
Ant inclusionAI Open-Sources the ZwZ Model Family and the ZoomBench Fine-Grained Perception Benchmark (ICML 2026)
Ant inclusionAI has released the ZwZ model family (4/7/8B) accompanying its ICML 2026 paper, achieving SOTA on multimodal perception benchmarks among open-source models, while requiring only a single forward inference pass and no tool calls at inference time. The method is Region-to-Image Distillation (R2I), which distills high-quality VQA supervision generated from enlarged local regions back into the full image, with 37K training samples. It also introduces ZoomBench, containing 845 VQA items covering 6 fine-grained perception dimensions, on which Qwen2.5-VL-7B achieves only 42.5% average accuracy.
Ant inclusionAI has released the ZwZ model family (4/7/8B) accompanying its ICML 2026 paper, achieving SOTA on multimodal perception benchmarks among open-source models, while requiring only a single forward inference pass and no tool calls at inference time. The method is Region-to-Image Distillation (R2I), which distills high-quality VQA supervision generated from enlarged local regions back into the full image, with 37K training samples. It also introduces ZoomBench, containing 845 VQA items
Matched fieldsBody
OpenBMB Releases MiniCPM-OCR 1.0 and Verification Model Jev
OpenBMB releases MiniCPM-OCR 1.0, a 1B-scale OCR model that achieves 93.68 on Real5-OmniDocBench and 54.7 on OCRBench v2. It also releases MiniCPM-OCR Jev, a 2B decision model used to compare parsing results with the original image and give a 0–1 correctness score, approaching Qwen3.5-397B-A17B on JevOCRBench. Both models have open weights, support deployment via llama.cpp, vLLM, SGLang, and Ollama, and come with open-source fine-tuning code.
Matched fieldsBody
Teradyne Robotics and Sereact Partner to Bring Physical AI Solutions to Retail and Logistics
Teradyne Robotics and Sereact announced a formal partnership at Nvidia GTC Berlin 2026, running Sereact's physical AI platform Cortex on Universal Robots' collaborative robots for unstructured tasks such as returns processing and picking. The two parties have already delivered dual-arm systems composed of hundreds of UR robots to European customers including Zalando, and are autonomously processing returned goods at Zalando fulfillment centers in Greven, Germany, and Świebodzin, Poland. Cortex is a VLA model trained on Nvidia Isaac Sim and Cosmos, and runs on Nvidia Jetson Thor at deployment.
Teradyne Robotics and Sereact announced a formal partnership at Nvidia GTC Berlin 2026, running Sereact's physical AI platform Cortex on Universal Robots' collaborative robots for unstructured tasks such as returns processing and picking. The two parties have already delivered dual-arm systems composed of hundreds of UR robots to European customers including Zalando, and are autonomously processing returned goods at Zalando fulfillment centers in Greven, Germany, and Świebodzin, Poland. Cortex i
Matched fieldsBody
Ravi Vasantraj of Mphasis: Owning IP Itself Is Meaningless
Ravi Vasantraj, Global Delivery Head at Mphasis, believes that models themselves are rapidly commoditizing and that owning IP itself is meaningless—what enterprises are really buying is outcomes delivered using AI. He revealed that AI-led work accounted for 63% of Mphasis's total new contract signings of $461 million in the June quarter, and that the share of AI-related business opportunities has risen from 12% to about 70%, but only 5% of AI sandbox projects have entered production environments, with the company currently seeing productivity gains of about 20%.
Ravi Vasantraj, Global Delivery Head at Mphasis, believes that models themselves are rapidly commoditizing and that owning IP itself is meaningless—what enterprises are really buying is outcomes delivered using AI. He revealed that AI-led work accounted for 63% of Mphasis's total new contract signings of $461 million in the June quarter, and that the share of AI-related business opportunities has risen from 12% to about 70%, but only 5% of AI sandbox projects have entered production environments
Matched fieldsBody
Anthropic Abolishes Coding Interviews: Felix Reiseberg Says "Hunger beats tenure"
Anthropic hiring manager Felix Reiseberg announced that the company will no longer use traditional coding interviews to evaluate candidates, citing the fact that Claude is already stronger than humans at both React challenges and assembly problems, and that truly impactful builders do not win through such tests. He advises junior engineers to pick an idea that can drive business impact, build a working prototype over the weekend, focus on just one good idea, and leverage their fresh perspective on users. On the same day, Google DeepMind released the Genie 4 world model, and Claude Opus 5.5 (58 points) topped the Artificial Analysis Intelligence Index v5.
Anthropic hiring manager Felix Reiseberg announced that the company will no longer use traditional coding interviews to evaluate candidates, citing the fact that Claude is already stronger than humans at both React challenges and assembly problems, and that truly impactful builders do not win through such tests. He advises junior engineers to pick an idea that can drive business impact, build a working prototype over the weekend, focus on just one good idea, and leverage their fresh perspective
Matched fieldsBody
Chaoyan Intelligence Open-Sources 35B Model Apex-35B-V1.0, with AI Participating in Its Own R&D
Beijing-based large model startup Chaoyan Intelligence has released and open-sourced its first large model, Apex-35B-V1.0, with 35 billion total parameters, approximately 3 billion parameters activated per token, and support for a 262,000-token context window, optimized through mid-training and post-training based on Qwen3.6-35B-A3B. The model scored 60.9 and 40.0 on the paper reproduction test PaperBench and the scientific research challenge test FrontierScience, respectively, higher than GLM-5.3's 42.5 and 30.7; it scored 84.0 on the tool-calling test BFCL V4, higher than DeepSeek-V4.1-Flash-1017's 75.0. The model has been released with open weights under the Apache 2.0 license, and the official release provides approximately 72GB of BF16 weights and GGUF quantized versions at various precisions.
Beijing-based large model startup Chaoyan Intelligence has released and open-sourced its first large model, Apex-35B-V1.0, with 35 billion total parameters, approximately 3 billion parameters activated per token, and support for a 262,000-token context window, optimized through mid-training and post-training based on Qwen3.6-35B-A3B. The model scored 60.9 and 40.0 on the paper reproduction test PaperBench and the scientific research challenge test FrontierScience, respectively, higher than GLM-5
Matched fieldsBody
a16z Survey of 100 CISOs: How AI Is Rewriting the 2027 Security Agenda
a16z's survey of 100 CISOs at companies with more than 1,000 employees shows that 87% expect cybersecurity spending to increase, driven mainly by AI-related threats and the expanded attack surface from enterprises' own AI deployments. All seven of the most-watched security initiatives for 2027 are AI-related; SOC automation received the most first-priority votes, but only 18% have put AI SOC tools into production. 45% plan to make new purchases or upgrades within 12 months for internal AI agent security, and 42% plan to procure offensive security capabilities.
a16z's survey of 100 CISOs at companies with more than 1,000 employees shows that 87% expect cybersecurity spending to increase, driven mainly by AI-related threats and the expanded attack surface from enterprises' own AI deployments. All seven of the most-watched security initiatives for 2027 are AI-related; SOC automation received the most first-priority votes, but only 18% have put AI SOC tools into production. 45% plan to make new purchases or upgrades within 12 months for internal AI agent
Matched fieldsBody
PyTorch 2.15 Released: Rewritten NCCL Backend Becomes Default, Apple Silicon Gains FP8 Support
The PyTorch Foundation has released PyTorch 2.15. All process groups using nccl now default to the rewritten NCCL backend, accompanied by a flight recorder that dumps on-site information when collective communication hangs, and can be reverted via TORCH_DIST_USE_NCCL2=0 or nccl-legacy. This version also adds float8_e4m3fn and _scaled_mm support for Apple Silicon, and the new grouped_mm improves MoE decoding throughput by 1.32–1.34x on M3 Ultra. The default wheel is upgraded to CUDA 13.2, while a CUDA 13.4 wheel is also provided with experimental support for NVIDIA Rubin GPUs; a ROCm 10.0 wheel is newly added, and Python 3.10 and macOS 14 are no longer supported.
The PyTorch Foundation has released PyTorch 2.15. All process groups using nccl now default to the rewritten NCCL backend, accompanied by a flight recorder that dumps on-site information when collective communication hangs, and can be reverted via TORCH DIST USE NCCL2=0 or nccl-legacy. This version also adds float8 e4m3fn and scaled mm support for Apple Silicon, and the new grouped mm improves MoE decoding throughput by 1.32–1.34x on M3 Ultra. The default wheel is upgraded to CUDA 13.2, while a
Matched fieldsBody
Mistral says its trillion-parameter open-source model Le Chonk surpasses Chinese models in coding and cybersecurity
Mistral has released the official version of Le Chonk, a trillion-parameter open-weight model, calling it the world's best open-weight model in automated coding and cybersecurity, and making it freely available. Pierre Stock, Mistral's VP of Science, said the model was trained on increasingly difficult cybersecurity and coding problems, enabling it to match or surpass models from Chinese labs such as DeepSeek and Z.ai. Stock also stressed that the West needs its own advanced open-weight models, because closed models can be shut down remotely, whereas open weights cannot have access cut off once released.
Mistral has released the official version of Le Chonk, a trillion-parameter open-weight model, calling it the world's best open-weight model in automated coding and cybersecurity, and making it freely available. Pierre Stock, Mistral's VP of Science, said the model was trained on increasingly difficult cybersecurity and coding problems, enabling it to match or surpass models from Chinese labs such as DeepSeek and Z.ai. Stock also stressed that the West needs its own advanced open-weight models,
Matched fieldsBody
AI Drives a Wave of Formal Verification Startups as &Truth Raises $70 Million
WIRED reports that generative AI is pushing formal verification toward commercialization. &Truth has raised $70 million and launched publicly, while Harmonic has raised a cumulative $295 million, and its math AI agent Aristotle supports verification of C code and Verilog hardware designs. Stanford researcher Chuyue Sun used AI tools to complete an engineering project in less than half a day at a cost of under $500, whereas three experts working by hand took five months to complete the same work. The report also notes that the technology is still at the prototype stage, that customers' willingness to pay and adoption methods remain unclear, and that Antithesis's Will Wilson calls formal verification a "trap" that can only provide partial guarantees.
WIRED reports that generative AI is pushing formal verification toward commercialization. &Truth has raised $70 million and launched publicly, while Harmonic has raised a cumulative $295 million, and its math AI agent Aristotle supports verification of C code and Verilog hardware designs. Stanford researcher Chuyue Sun used AI tools to complete an engineering project in less than half a day at a cost of under $500, whereas three experts working by hand took five months to complete the same work.
Matched fieldsBody
Google DeepMind releases Gemini 4 Carbon, topping Artificial…
Google DeepMind has released Gemini 4 Carbon, which ranks first on the Artificial Analysis Intelligence Index v5 with a score of 59 and leads in 4 of 9 evaluations including HLE-Diamond and TB-Science. The model focuses on cybersecurity and long-horizon tasks, claiming 88.2% on SEC-bench Pro, 54.8% on ExploitGym, and 71.0% on CWE-bench, and is initially available only to trusted cybersecurity personnel through the Fairwind Program. In terms of specifications, it supports a 1M token context and 1M maximum output, with input covering text, images, video, and speech, and output in text; Artificial Analysis lists a standard price of $4/$20 per million input/output tokens, plus a 50% launch discount.
Google DeepMind has released Gemini 4 Carbon, which ranks first on the Artificial Analysis Intelligence Index v5 with a score of 59 and leads in 4 of 9 evaluations including HLE-Diamond and TB-Science. The model focuses on cybersecurity and long-horizon tasks, claiming 88.2% on SEC-bench Pro, 54.8% on ExploitGym, and 71.0% on CWE-bench, and is initially available only to trusted cybersecurity personnel through the Fairwind Program. In terms of specifications, it supports a 1M token context and 1
Matched fieldsBody
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.
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 Mo
Matched fieldsBody