https://www.ainformed.dev/articles/2026-09-18-qwen38-omni-flash-alibabas-first-ai-that-understands-and-acts-on-audio-and-video AInformed en 2026-09-18T12:02:46.666880+00:00 Qwen3.8-Omni-Flash: Alibaba's first AI that understands and acts on audio and video https://www.ainformed.dev/articles/2026-09-18-polydao-launches-megatrain-train-100b-parameter-models-on-a-single-gpu AInformed en 2026-09-18T12:01:04.590515+00:00 PolyDAO Launches MegaTrain: Train 100B-Parameter Models on a Single GPU https://www.ainformed.dev/articles/2026-09-18-openai-launches-astra-for-law-ai-tailored-for-legal-professionals AInformed en 2026-09-18T12:09:07.042804+00:00 OpenAI Launches Astra for Law: AI Tailored for Legal Professionals https://www.ainformed.dev/articles/2026-09-18-openai-finds-its-ai-models-leaving-secret-notes-to-hide-flaws AInformed en 2026-09-18T12:08:08.658608+00:00 OpenAI caught GPT-5.6 Sol leaving secret notes to future models to hide mistakes https://www.ainformed.dev/articles/2026-09-18-openai-and-aarp-teach-1000-older-adults-to-use-chatgpt-safely AInformed en 2026-09-18T12:12:24.556064+00:00 OpenAI and AARP bring free ChatGPT workshops to 1,000 older adults across 10 U.S. cities https://www.ainformed.dev/articles/2026-09-18-mistral-releases-24b-parameter-model-with-4-bit-quantization AInformed en 2026-09-18T11:59:10.933420+00:00 Mistral Releases 24B-Parameter Model with 4-Bit Quantization https://www.ainformed.dev/articles/2026-09-18-how-ai-chatbots-use-web-search-a-first-of-its-kind-study AInformed en 2026-09-18T12:13:23.026092+00:00 New Study Reveals How ChatGPT, Claude, Grok, and DeepSeek Use Web Search Differently https://www.ainformed.dev/articles/2026-09-18-detecting-harm-in-ai-by-reading-its-internal-thoughts AInformed en 2026-09-18T12:14:29.122040+00:00 LLaMA-3.1-8B Internal Activations Enable Faster Harm Detection Without External Guardrails https://www.ainformed.dev/articles/2026-09-18-claude-codes-new-projects-feature-lets-you-manage-multiple-ai-agents AInformed en 2026-09-18T12:10:58.658585+00:00 Claude Code Relaunches Projects to Manage Multiple AI Agents in the Cloud https://www.ainformed.dev/articles/2026-09-18-base-labs-partners-with-hugging-face-and-goodfire-for-open-ai-safety-research AInformed en 2026-09-18T12:09:59.562975+00:00 Base Labs Partners with Hugging Face and Goodfire to Publish Open-Weight AI Safety Methods https://www.ainformed.dev/articles/2026-09-17-zai-uses-glm-53-to-optimize-its-own-ai-infrastructure-in-just-two-weeks AInformed en 2026-09-17T12:19:40.932773+00:00 Z.ai Uses Its Own GLM-5.3 Model to Optimize GLM-5.3-Flash Inference Infrastructure, Tripling Throughput in Two Weeks https://www.ainformed.dev/articles/2026-09-17-openai-releases-framework-to-track-and-report-ai-model-misalignment AInformed en 2026-09-17T12:24:32.107577+00:00 OpenAI Releases Framework to Track and Report AI Model Misalignment https://www.ainformed.dev/articles/2026-09-17-openai-launches-ai-powered-advertising-tools-for-marketers AInformed en 2026-09-17T12:27:31.987851+00:00 OpenAI Launches Sponsored Agents and HubSpot/Shopify Integrations for AI-Powered Advertising https://www.ainformed.dev/articles/2026-09-17-nemo-data-designer-a-new-open-source-tool-for-custom-ai-training-data AInformed en 2026-09-17T12:31:53.418886+00:00 NeMo Data Designer: NVIDIA's Open-Source Framework for Multimodal Synthetic Data Generation https://www.ainformed.dev/articles/2026-09-17-how-artie-uses-ai-agents-to-automate-work-with-hermes AInformed en 2026-09-17T12:20:37.485530+00:00 How Artie Uses Hermes AI Agents from Nous Research to Automate Work 24/7 https://www.ainformed.dev/articles/2026-09-17-google-deepminds-logan-kilpatrick-outlines-strategy-for-gemini-4-and-frontier-ai-models AInformed en 2026-09-17T17:37:45.206277+00:00 Google DeepMind's Logan Kilpatrick Outlines Strategy for Gemini 4 and Frontier AI Models https://www.ainformed.dev/articles/2026-09-17-fyxers-ai-executive-assistant-learns-your-voice-and-workflow AInformed en 2026-09-17T12:29:34.849879+00:00 Fyxer's AI Executive Assistant Learns Your Voice and Workflow https://www.ainformed.dev/articles/2026-09-17-elia-kuratli-shares-breakthrough-in-ai-training-efficiency AInformed en 2026-09-17T12:18:43.633878+00:00 Elia Kuratli Shares Breakthrough in AI Training Efficiency: 100B-Parameter Models on a Single GPU https://www.ainformed.dev/articles/2026-09-17-capmem-new-benchmark-tests-ais-memory-of-wearable-camera-videos AInformed en 2026-09-17T12:30:52.470493+00:00 CapMem Benchmark Tests AI's Episodic Memory on 33.7 Hours of Wearable Camera Video https://www.ainformed.dev/articles/2026-09-17-ais-growing-e-waste-problem-23-million-shipping-containers-by-2050 AInformed en 2026-09-17T12:25:37.497065+00:00 AI data center e-waste could fill 23 million shipping containers by 2050, report warns https://www.ainformed.dev/articles/2026-09-17-ai-safety-warnings-escalate-as-former-openai-and-anthropic-researcher-resigns AInformed en 2026-09-17T17:37:27.103389+00:00 AI Safety Warnings Escalate as Former OpenAI and Anthropic Researcher Resigns https://www.ainformed.dev/articles/2026-09-17-ai-predicts-extubation-failure-by-analyzing-respiratory-therapy-notes AInformed en 2026-09-17T12:33:33.379377+00:00 AI Predicts Extubation Failure by Analyzing Respiratory Therapy Notes https://www.ainformed.dev/articles/2026-09-17-ai-models-can-fake-personality-traits-like-humans-in-job-interviews-and-legal-se AInformed en 2026-09-17T12:34:44.126534+00:00 LLMs Can Fake Dark Triad Personality Traits in Job Interviews and Legal Evaluations, Study Finds https://www.ainformed.dev/articles/2026-09-16-researchers-map-personality-traits-of-22-ai-models-revealing-behavioral-patterns AInformed en 2026-09-16T12:29:00.362191+00:00 Study Maps Self-Reported Personalities of 22 LLMs, Revealing Distinct Behavioral Patterns in GPT, Grok, Gemini, and Claude https://www.ainformed.dev/articles/2026-09-16-orchslm-how-tiny-ai-models-could-replace-big-ones-in-key-tasks AInformed en 2026-09-16T12:22:51.890635+00:00 OrchSLM: University of Washington Researchers Orchestrate Small AI Models to Replace Large Ones for Specific Tasks https://www.ainformed.dev/articles/2026-09-16-odobot-a-cost-efficient-ai-web-agent-that-learns-from-user-behavior AInformed en 2026-09-16T12:24:00.960042+00:00 OdoBot: A Cost-Efficient AI Web Agent That Learns from User Behavior https://www.ainformed.dev/articles/2026-09-16-nvidia-ceo-jensen-huang-dismisses-ai-catastrophism-as-unfounded-in-podcast-appearance AInformed en 2026-09-16T17:37:27.255974+00:00 Nvidia CEO Jensen Huang Dismisses AI Catastrophism as Unfounded in Podcast Appearance https://www.ainformed.dev/articles/2026-09-16-new-research-shows-how-to-optimize-ai-model-costs-for-better-performance AInformed en 2026-09-16T12:26:05.605189+00:00 New ArXiv Study Proposes Adaptive Routing to Cut LLM Inference Costs Without Sacrificing Performance https://www.ainformed.dev/articles/2026-09-16-metas-new-one-subscriptions-bundle-social-media-and-ai-features-for-a-price AInformed en 2026-09-16T12:21:10.404955+00:00 Meta’s new One subscriptions bundle social media and AI features for a price https://www.ainformed.dev/articles/2026-09-16-latent-undertow-how-typos-break-ai-security-systems AInformed en 2026-09-16T12:27:12.395004+00:00 Latent Undertow: How Ordinary Typos Break AI Safety Probes https://www.ainformed.dev/articles/2026-09-16-bias-audits-find-ai-models-biased-but-disagree-on-rankings AInformed en 2026-09-16T12:28:04.726473+00:00 Bias Audits Detect AI Model Bias but Disagree on Rankings, Study Finds https://www.ainformed.dev/articles/2026-09-16-asclepius-a-new-framework-for-testing-long-term-ai-performance-in-emergency-care AInformed en 2026-09-16T12:24:59.420753+00:00 Asclepius: A New Framework for Testing Long-Term AI Performance in Emergency Care