https://www.ainformed.dev/articles/2026-07-31-skillsmith-ai-system-combines-textual-knowledge-and-learned-skills-for-smarter-p AInformed en 2026-07-31T16:12:11.573818+00:00 SkillSmith: New AI System Merges Textual Knowledge with Learned Skills for Smarter Problem-Solving https://www.ainformed.dev/articles/2026-07-31-researchers-evaluate-41-open-weight-language-models-for-intent-classification AInformed en 2026-07-31T16:10:39.466233+00:00 41 Open-Weight Language Models Benchmarked for Zero-Shot Intent Classification: Study Finds Best Models for Accuracy vs. Efficiency https://www.ainformed.dev/articles/2026-07-31-new-study-reveals-how-ai-doctors-misinterpret-patient-stories-based-on-language- AInformed en 2026-07-31T11:07:30.009539+00:00 Narrative Anchoring: Study Shows Clinical AI Models Diagnose Differently Based on Language Style https://www.ainformed.dev/articles/2026-07-31-new-benchmark-tests-ais-ability-to-reason-with-probabilities AInformed en 2026-07-31T16:09:24.997068+00:00 New Benchmark Tests LLMs on Logical Inference Over Probability Operators https://www.ainformed.dev/articles/2026-07-31-multivationbench-testing-ais-ability-to-understand-motivations-in-stories AInformed en 2026-07-31T11:10:26.587201+00:00 MultivationBench: New Benchmark Tests AI's Ability to Understand Character Motivations in Visual Stories https://www.ainformed.dev/articles/2026-07-31-layerrag-bench-new-benchmark-tests-ai-agent-reliability-across-multiple-layers AInformed en 2026-07-31T16:07:47.988358+00:00 LayerRAG-Bench: New Benchmark Reveals AI Agents Fail at Evidence, Permissions, and Tool Contracts https://www.ainformed.dev/articles/2026-07-31-is-ai-reasoning-right-for-the-wrong-reasons AInformed en 2026-07-31T16:04:52.135258+00:00 MIT and Stanford Study: AI Models Mimic Reasoning Without True Understanding https://www.ainformed.dev/articles/2026-07-31-how-avatarin-built-a-247-retail-agent-with-gpt-realtime AInformed en 2026-07-31T11:02:28.416128+00:00 How avatarin built a 24/7 retail agent with GPT-Realtime https://www.ainformed.dev/articles/2026-07-31-gogotb-ai-framework-for-smarter-chip-design-verification AInformed en 2026-07-31T16:06:23.148842+00:00 GoGoTB: Agentic AI Framework Automates Chip Verification to Catch Costly Bugs https://www.ainformed.dev/articles/2026-07-31-eu-mandates-ai-labels-for-authentic-looking-ai-generated-content AInformed en 2026-07-31T16:03:43.555390+00:00 EU Mandates Compulsory AI Labels for Authentic-Looking Content https://www.ainformed.dev/articles/2026-07-31-cam-wolf-ai-agents-that-play-social-deduction-games-like-humans AInformed en 2026-07-31T11:03:32.042766+00:00 CaM-Wolf: The First Multimodal AI Agent for Social Deduction Games Like Werewolf https://www.ainformed.dev/articles/2026-07-31-b1ade-tiny-ai-models-that-outperform-giants-in-retrieval-tasks AInformed en 2026-07-31T16:13:24.998181+00:00 B1ade: Tiny AI Models That Outperform Giants in Retrieval Tasks https://www.ainformed.dev/articles/2026-07-31-ai-scibrief-how-prompt-chaining-improves-scholarly-report-generation AInformed en 2026-07-31T11:04:27.265970+00:00 AI SciBrief: How Prompt Chaining Improves Scholarly Report Generation https://www.ainformed.dev/articles/2026-07-31-ai-models-struggle-to-match-human-emotional-perception-in-news-framing AInformed en 2026-07-31T11:05:59.360231+00:00 AI Models Fail to Match Human Emotional Perception of News Headlines, Cambridge Study Finds https://www.ainformed.dev/articles/2026-07-31-ai-leaders-clash-over-open-weights-as-openai-and-anthropic-team-up-for-washington-lobbying AInformed en 2026-07-31T15:28:12.158605+00:00 AI Leaders Clash Over Open Weights as OpenAI and Anthropic Team Up for Washington Lobbying https://www.ainformed.dev/articles/2026-07-31-ai-evaluation-scores-are-perishable-knowledge-claims-researchers-argue AInformed en 2026-07-31T11:09:32.366812+00:00 AI Evaluation Scores Are Perishable Knowledge Claims, Researchers Argue https://www.ainformed.dev/articles/2026-07-31-aha-memes-new-benchmark-for-detecting-hate-in-arabic-memes AInformed en 2026-07-31T11:08:28.557880+00:00 AHA-Memes: First Large-Scale Arabic Hateful Meme Benchmark with Fine-Grained Labels https://www.ainformed.dev/articles/2026-07-30-xai-sues-minnesota-to-block-nudification-app-law-ahead-of-august-1-deadline AInformed en 2026-07-30T16:17:49.396421+00:00 xAI Sues Minnesota to Block 'Nudification' App Law Ahead of August 1 Deadline https://www.ainformed.dev/articles/2026-07-30-why-reinforcement-learning-beats-fine-tuning-in-math-problem-solving AInformed en 2026-07-30T11:12:18.817846+00:00 Why Reinforcement Learning Beats Fine-Tuning in Math Problem-Solving https://www.ainformed.dev/articles/2026-07-30-why-ai-benchmark-results-dont-always-add-up-to-reliable-conclusions AInformed en 2026-07-30T21:06:58.074668+00:00 New ArXiv Paper Reveals Why AI Benchmark Results Don't Compose Into Reliable Conclusions https://www.ainformed.dev/articles/2026-07-30-tracecoder-ai-code-generation-with-full-transparency-and-audit-trails AInformed en 2026-07-30T11:20:29.860719+00:00 TraceCoder: AI Code Generation with Full Transparency and Audit Trails https://www.ainformed.dev/articles/2026-07-30-running-a-tiny-ai-model-across-three-esp32-s3-boards AInformed en 2026-07-30T16:09:28.943754+00:00 Developer Runs 56M-Parameter AI Model Across Three ESP32-S3 Boards Using ESP-NOW https://www.ainformed.dev/articles/2026-07-30-openai-cuts-gpt-56-prices-boosts-enterprise-ai-workflows AInformed en 2026-07-30T21:05:49.991819+00:00 OpenAI Cuts GPT-5.6 Prices by 20% for Enterprise Luna and Terra Services https://www.ainformed.dev/articles/2026-07-30-nscale-acquires-anyscale-to-expand-ai-compute-infrastructure AInformed en 2026-07-30T16:06:16.882967+00:00 British AI cloud firm Nscale acquires Anyscale to own more of the AI compute stack https://www.ainformed.dev/articles/2026-07-30-new-ai-system-tracks-claims-to-evidence-for-accuracy AInformed en 2026-07-30T21:09:43.896981+00:00 Evidence-Ledger Adjudication: New AI System Traces Claims to Source Evidence for Accuracy https://www.ainformed.dev/articles/2026-07-30-microsoft-unveils-in-house-ai-models-challenges-openai-and-anthropic AInformed en 2026-07-30T16:07:38.476684+00:00 Microsoft Unveils In-House AI Models, Challenges OpenAI and Anthropic https://www.ainformed.dev/articles/2026-07-30-microsoft-copilot-hacked-to-spread-ai-worm-via-prompts AInformed en 2026-07-30T16:08:33.142247+00:00 Microsoft Copilot Exploit Spreads AI Worm via Malicious Prompts https://www.ainformed.dev/articles/2026-07-30-microsoft-confirms-copilot-super-app-coming-this-year AInformed en 2026-07-30T11:07:48.647197+00:00 Microsoft confirms Copilot ‘super app’ coming this year https://www.ainformed.dev/articles/2026-07-30-meta-plans-major-push-into-personal-ai-agents AInformed en 2026-07-30T11:10:05.852686+00:00 Meta Plans Major Push Into Personal AI Agents That Act on Your Behalf https://www.ainformed.dev/articles/2026-07-30-how-two-settings-tripled-openais-scores-on-the-arc-agi-3-benchmark AInformed en 2026-07-30T16:10:45.889099+00:00 How Two Settings Tripled OpenAI's Scores on the ARC-AGI-3 Benchmark https://www.ainformed.dev/articles/2026-07-30-how-ai-agents-work-on-google-cloud-a-complete-guide AInformed en 2026-07-30T16:15:53.712728+00:00 How AI Agents Work on Google Cloud: A Complete Guide https://www.ainformed.dev/articles/2026-07-30-guideskill-ai-system-turns-medical-guidelines-into-executable-rules-for-better-d AInformed en 2026-07-30T11:18:18.050247+00:00 GuideSkill: AI System Turns Medical Guidelines into Executable Rules for Better Diagnoses https://www.ainformed.dev/articles/2026-07-30-google-deepminds-gemini-robotics-2-now-controls-full-humanoid-robots AInformed en 2026-07-30T21:04:36.668838+00:00 Google DeepMind’s Gemini Robotics 2 Now Controls Full Humanoid Robots https://www.ainformed.dev/articles/2026-07-30-evopinn-ai-automates-physics-based-neural-network-design AInformed en 2026-07-30T21:08:53.145650+00:00 EvoPINN: AI System Automates Physics-Informed Neural Network Design for Faster PDE Solving https://www.ainformed.dev/articles/2026-07-30-eco3s-simulating-complex-socio-economic-systems-with-ai-agents AInformed en 2026-07-30T21:10:34.801037+00:00 Eco3S: UC Berkeley's AI Agent Framework for Simulating Economic Policy Scenarios https://www.ainformed.dev/articles/2026-07-30-clinlens-a-new-benchmark-for-ai-in-clinical-data-science AInformed en 2026-07-30T11:16:24.408264+00:00 ClinLens: A New Benchmark for AI Agents in Longitudinal Clinical Data Science https://www.ainformed.dev/articles/2026-07-30-cg-world-a-new-dataset-for-training-ai-to-understand-the-physical-world AInformed en 2026-07-30T21:03:39.859635+00:00 CG-World Dataset: Training AI to Understand Physical World Dynamics https://www.ainformed.dev/articles/2026-07-30-ai-agents-tackle-physics-problem-transformation-in-new-benchmark-study AInformed en 2026-07-30T11:26:37.103995+00:00 AI Agents Tackle Physics Problem Transformation in New StatMechBench-v0 Benchmark https://www.ainformed.dev/articles/2026-07-30-ai-agents-deceive-in-mixed-motive-games-like-werewolf AInformed en 2026-07-30T11:14:20.121933+00:00 AI Agents Deceive in Mixed-Motive Games Like Werewolf, Study Finds https://www.ainformed.dev/articles/2026-07-29-recursive-superintelligence-secures-410m-aws-compute-deal AInformed en 2026-07-29T21:05:42.249364+00:00 Recursive Superintelligence signs $410M compute deal with Amazon https://www.ainformed.dev/articles/2026-07-29-pangram-raises-9m-to-combat-ai-generated-content-with-new-detection-tools AInformed en 2026-07-29T21:03:05.612470+00:00 Pangram raises $9M to detect AI-generated text and images with new Pangram 4 model https://www.ainformed.dev/articles/2026-07-29-openai-gives-100000-researchers-free-chatgpt-access-to-boost-scientific-work AInformed en 2026-07-29T21:04:06.184065+00:00 OpenAI Gives 100,000 Researchers Free Access to Advanced ChatGPT Models to Accelerate Scientific Discovery https://www.ainformed.dev/articles/2026-07-29-my-local-llm-scored-66-it-was-wrong-every-time AInformed en 2026-07-29T21:08:30.036137+00:00 My Local LLM Scored 6/6. It Was Wrong Every Time https://www.ainformed.dev/articles/2026-07-29-gpt-56-openais-latest-model-balances-intelligence-and-efficiency AInformed en 2026-07-29T21:06:39.757268+00:00 GPT-5.6: OpenAI's latest model balances intelligence and efficiency https://www.ainformed.dev/articles/2026-07-29-fish-audio-raises-52m-to-expand-ai-voice-models-for-creators-and-enterprises AInformed en 2026-07-29T21:04:57.598583+00:00 Fish Audio Raises $52M Seed to Build AI Voice Models for Creators and Enterprises https://www.ainformed.dev/articles/2026-07-29-cyera-acquires-oasis-security-for-1b-to-protect-ai-agents AInformed en 2026-07-29T21:09:20.044016+00:00 Cyera Acquires Oasis Security for $1B to Protect AI Agents https://www.ainformed.dev/articles/2026-07-29-claude-opus-5-exhibits-unethical-behavior-in-vending-machine-simulation AInformed en 2026-07-29T21:07:30.606072+00:00 Claude Opus 5 Cheated When Tasked With Running a Vending Machine Simulation