https://themodelwire.com/article/meta-launches-muse-agent-for-autonomous-email-and-shopping-tasks-01M3SEC8J7MXV7S4X06X48TA56 Modelwire en 2026-09-30T15:18:49.000Z Meta launches Muse agent for autonomous email and shopping tasks https://themodelwire.com/article/deepmind-embeds-watermarks-in-ai-designed-proteins-without-loss-of-function-01M3SDGPYHHYZ0APB6SMTYXMX2 Modelwire en 2026-09-30T15:03:07.000Z DeepMind embeds watermarks in AI-designed proteins without loss of function https://themodelwire.com/article/google-pilots-publisher-payments-for-ai-search-training-data-01M3SEC8Q6AW01DF74ACDJNZ36 Modelwire en 2026-09-30T14:51:49.000Z Google pilots publisher payments for AI search training data https://themodelwire.com/article/cerebras-explores-hardware-limits-as-ai-compute-demands-plateau-01M3SCNAWR7DF6GSCQF83402W5 Modelwire en 2026-09-30T14:30:00.000Z Cerebras explores hardware limits as AI compute demands plateau https://themodelwire.com/article/federal-government-builds-mass-license-plate-surveillance-database-through-local-01M3SBSWMBAEMEVMZDT6S7T56K Modelwire en 2026-09-30T14:29:15.000Z Federal government builds mass license plate surveillance database through local traffic systems https://themodelwire.com/article/restate-raises-20m-to-build-agent-native-workflow-infrastructure-01M3SBSVPD4KZFXCQD1XQ6KYZM Modelwire en 2026-09-30T14:27:11.000Z Restate raises $20M to build agent-native workflow infrastructure https://themodelwire.com/article/trump-administration-formalizes-voluntary-ai-safety-pact-with-tech-leaders-01M3S4Y4WDCRRTWQ8G2S1CQGKH Modelwire en 2026-09-30T12:24:32.000Z Trump administration formalizes voluntary AI safety pact with tech leaders https://themodelwire.com/article/open-weight-glm-5-3-matches-claude-on-exploit-generation-raising-containment-con-01M3S0MY8GSHRQXY90EBY7PSBZ Modelwire en 2026-09-30T11:05:05.000Z Open-weight GLM-5.3 matches Claude on exploit generation, raising containment concerns https://themodelwire.com/article/openais-escaped-agents-expose-limits-of-ai-sandbox-security-01M3S0MSY785F5ZHE4VJT5K3R0 Modelwire en 2026-09-30T10:40:30.000Z OpenAI's escaped agents expose limits of AI sandbox security https://themodelwire.com/article/googles-publisher-payments-for-ai-content-reveal-wide-gaps-and-opacity-01M3RX71EZE0V3FV4DVJ2Q4Y6A Modelwire en 2026-09-30T10:17:15.000Z Google's publisher payments for AI content reveal wide gaps and opacity https://themodelwire.com/article/openai-pivots-toward-platform-and-identity-infrastructure-at-dev-day-01M3RWBFPTQYERHE73MWHG3GBM Modelwire en 2026-09-30T10:00:00.000Z OpenAI pivots toward platform and identity infrastructure at Dev Day https://themodelwire.com/article/white-house-secures-nonbinding-ai-conduct-pledge-from-major-tech-leaders-01M3RVG1XWCP9SX4TKFES594S0 Modelwire en 2026-09-30T09:46:10.000Z White House secures nonbinding AI conduct pledge from major tech leaders https://themodelwire.com/article/openai-pivots-agents-toward-enterprise-workflows-amid-safety-concerns-01M3QZ5NRCN8X4BME2M4JGWMJK Modelwire en 2026-09-30T01:20:57.000Z OpenAI pivots agents toward enterprise workflows amid safety concerns https://themodelwire.com/article/anthropic-pushes-regulation-while-its-agents-create-incidents-01M3R1R3CDG6ZZS3188XJ02762 Modelwire en 2026-09-30T01:12:58.000Z Anthropic pushes regulation while its agents create incidents https://themodelwire.com/article/openai-ties-ipo-timeline-to-model-safety-milestones-01M3QVQT5NBA81SMDN42VBG0T6 Modelwire en 2026-09-30T00:19:13.000Z OpenAI ties IPO timeline to model safety milestones https://themodelwire.com/article/hugging-face-opens-multilingual-tts-evaluation-leaderboard-01M3SDGQFZP8CCCT4QMVKK3CA4 Modelwire en 2026-09-30T00:00:00.000Z Hugging Face opens multilingual TTS evaluation leaderboard https://themodelwire.com/article/anthropic-red-team-flags-binary-exploitation-breakthrough-across-frontier-models-01M3QMW12MGWCZQ4HH2RWMHAAZ Modelwire en 2026-09-29T22:20:28.000Z Anthropic red team flags binary exploitation breakthrough across frontier models https://themodelwire.com/article/openai-embeds-app-marketplace-into-chatgpt-bypassing-traditional-stores-01M3QD4VTBHZYPRZ1VRS0NS02Y Modelwire en 2026-09-29T20:15:47.000Z OpenAI embeds app marketplace into ChatGPT, bypassing traditional stores https://themodelwire.com/article/openai-seeks-30b-at-1-4t-valuation-ahead-of-2027-ipo-01M3QC9C3PAZ0GEAD7TYGP063F Modelwire en 2026-09-29T19:52:37.000Z OpenAI seeks $30B at $1.4T valuation ahead of 2027 IPO https://themodelwire.com/article/openai-seeks-30b-at-1-4t-valuation-ahead-of-2027-ipo-01M3QTWAANYM4CQTVSSCB2VS8N Modelwire en 2026-09-29T19:52:37.000Z OpenAI seeks $30B at $1.4T valuation ahead of 2027 IPO https://themodelwire.com/article/gpt-6-astras-attack-success-rate-surges-fivefold-in-uk-security-tests-01M3QAJGT8KM41Q27W9G6AF38H Modelwire en 2026-09-29T19:24:12.000Z GPT-6 Astra's attack success rate surges fivefold in UK security tests https://themodelwire.com/article/nonprofit-sues-openai-over-hugging-face-security-breach-01M3Q9Q0FRBS9DVX5G54CJZTX8 Modelwire en 2026-09-29T19:05:00.000Z Nonprofit sues OpenAI over Hugging Face security breach https://themodelwire.com/article/openai-quietly-partners-with-nvidia-on-safety-while-avoiding-public-endorsement-01M3Q802FZBASQ3N6TY7NK2Y7C Modelwire en 2026-09-29T18:35:00.000Z OpenAI quietly partners with Nvidia on safety while avoiding public endorsement https://themodelwire.com/article/openai-releases-gpt-6-1-sol-at-fifth-the-cost-of-astra-01M3QVQS17EASNP8XS75671PQQ Modelwire en 2026-09-29T18:27:48.000Z OpenAI releases GPT-6.1 Sol at fifth the cost of Astra https://themodelwire.com/article/openai-launches-codex-cloud-for-cross-device-development-workflows-01M3Q74MWX42HAVDJ5WGVK3ZSP Modelwire en 2026-09-29T18:18:13.000Z OpenAI launches Codex Cloud for cross-device development workflows https://themodelwire.com/article/openais-chatgpt-reaches-1-2-billion-weekly-users-as-revenue-accelerates-70-perce-01M3Q697E9BQH1JZ3Q8QVJVSW6 Modelwire en 2026-09-29T18:10:34.000Z OpenAI's ChatGPT reaches 1.2 billion weekly users as revenue accelerates 70 percent https://themodelwire.com/article/openai-launches-persistent-agents-with-continuous-learning-and-4-000-app-integra-01M3Q69755F0K8Q8P87VW1VTXJ Modelwire en 2026-09-29T18:04:41.000Z OpenAI launches persistent agents with continuous learning and 4,000-app integration https://themodelwire.com/article/leapquant-enables-lossless-8-bit-quantization-for-linear-attention-inference-01M3R4AFVNJ4P826S0GYT4GT42 Modelwire en 2026-09-29T17:59:34.000Z LeapQuant enables lossless 8-bit quantization for linear attention inference https://themodelwire.com/article/researchers-introduce-feedback-pathways-to-transformer-architecture-01M3R4AFWRT6SEJ1GVZYA4J20Z Modelwire en 2026-09-29T17:57:40.000Z Researchers introduce feedback pathways to transformer architecture https://themodelwire.com/article/new-benchmark-exposes-long-context-llm-harness-tradeoffs-01M3R4AFYMDDKC4GCW7QK37Z3M Modelwire en 2026-09-29T17:55:07.000Z New benchmark exposes long-context LLM harness tradeoffs https://themodelwire.com/article/wush-kv-quantization-cuts-kv-cache-memory-overhead-for-long-context-inference-01M3R4AG19S0J1ZYXJQMFAE0M6 Modelwire en 2026-09-29T17:50:50.000Z WUSH-KV quantization cuts KV cache memory overhead for long-context inference https://themodelwire.com/article/qwen-and-gemma-models-gain-early-warning-grounding-checks-via-routing-analysis-01M3R4AFZFJDV9NJVQA1J3QA0F Modelwire en 2026-09-29T17:49:07.000Z Qwen and Gemma models gain early-warning grounding checks via routing analysis https://themodelwire.com/article/hybrid-attention-layers-implicitly-encode-position-without-explicit-embeddings-01M3R4AG17D7C6HST58H0A0CAB Modelwire en 2026-09-29T17:47:41.000Z Hybrid attention layers implicitly encode position without explicit embeddings https://themodelwire.com/article/llm-agents-fail-to-execute-their-own-plans-consistently-01M3R4AG4CDRB5VGJP09W1RJZ8 Modelwire en 2026-09-29T17:47:18.000Z LLM agents fail to execute their own plans consistently https://themodelwire.com/article/models-solve-math-correctly-but-show-invalid-reasoning-steps-01M3R4AG2CAVY70XVGJCKKJZVY Modelwire en 2026-09-29T17:46:55.000Z Models solve math correctly but show invalid reasoning steps https://themodelwire.com/article/model-pruning-widens-demographic-gaps-in-speech-recognition-systems-01M3R4AG359R22ZBAA3MSQYSQW Modelwire en 2026-09-29T17:46:50.000Z Model pruning widens demographic gaps in speech recognition systems https://themodelwire.com/article/openai-enters-productivity-software-escalating-competition-with-microsoft-01M3Q4J77FGTK0YCN39ST58RAE Modelwire en 2026-09-29T17:45:51.000Z OpenAI enters productivity software, escalating competition with Microsoft https://themodelwire.com/article/parallel-tempering-replaces-rl-for-efficient-small-model-reasoning-01M3R4AG5NVSMT2PY9YZEBAYNF Modelwire en 2026-09-29T17:45:43.000Z Parallel tempering replaces RL for efficient small model reasoning https://themodelwire.com/article/llms-learn-dense-retrieval-from-examples-alone-no-training-required-01M3R4AG4E8HGBZZ53RG1HWZKD Modelwire en 2026-09-29T17:44:16.000Z LLMs learn dense retrieval from examples alone, no training required https://themodelwire.com/article/zero-order-training-closes-memory-gap-with-opt-30b-at-10x-lower-overhead-01M3R4AG7N516EFD3PKNQ5ER5V Modelwire en 2026-09-29T17:43:53.000Z Zero-order training closes memory gap with OPT-30B at 10x lower overhead https://themodelwire.com/article/mira-enables-single-gpu-moe-inference-through-predictive-expert-staging-01M3R4AG88DWMHE7Z8G9FF4J98 Modelwire en 2026-09-29T17:41:28.000Z Mira enables single-GPU MoE inference through predictive expert staging https://themodelwire.com/article/researchers-map-hidden-computational-diversity-within-neural-network-solution-re-01M3R4AG8Z6KNCBBADXNQ7H9Z8 Modelwire en 2026-09-29T17:38:24.000Z Researchers map hidden computational diversity within neural network solution regions https://themodelwire.com/article/openai-and-deepmind-researchers-go-public-on-superintelligence-extinction-risk-01M3Q4J78R50FSX28FQXY8QSNM Modelwire en 2026-09-29T17:35:03.000Z OpenAI and DeepMind researchers go public on superintelligence extinction risk https://themodelwire.com/article/openai-transforms-chatgpt-into-workplace-operating-system-01M3Q3PSH744JX025X8S9KMDND Modelwire en 2026-09-29T17:23:22.000Z OpenAI transforms ChatGPT into workplace operating system https://themodelwire.com/article/wabi-shifts-app-builder-to-agentic-chat-interface-with-dynamic-ui-01M3Q3PQQAD091F5YQBBYJ3DYE Modelwire en 2026-09-29T17:20:00.000Z Wabi shifts app builder to agentic chat interface with dynamic UI https://themodelwire.com/article/openai-releases-dots-hardware-agnostic-autonomous-agents-01M3Q2V9A4PDSVA2FP2A2N6X8K Modelwire en 2026-09-29T17:17:15.000Z OpenAI releases Dots, hardware-agnostic autonomous agents https://themodelwire.com/article/openai-introduces-cost-efficient-gpt-6-1-sol-variant-for-enterprise-workflows-01M3Q2V9BWMC5CQWQZRPHF90CE Modelwire en 2026-09-29T17:15:00.000Z OpenAI introduces cost-efficient GPT-6.1 Sol variant for enterprise workflows https://themodelwire.com/article/openai-deploys-dots-to-compete-in-the-agentic-assistant-market-01M3Q2V8P0SA0BVHNN1DCMFBZ9 Modelwire en 2026-09-29T17:15:00.000Z OpenAI deploys Dots to compete in the agentic assistant market https://themodelwire.com/article/openai-embeds-app-like-interfaces-and-automations-into-chatgpt-plugins-01M3Q2V9CRJ5BKTEVAMSN3H6VE Modelwire en 2026-09-29T17:15:00.000Z OpenAI embeds app-like interfaces and automations into ChatGPT plugins