iFeeling Daily
Daily curated AI insights you can't miss.
Benchmarking open models on agentic tooling with transformers
This article challenges the common practice of only checking final answers in agent benchmarks. Using transformers as a case study, it introduces a harness that measures agent effort across turns, tokens, and errors, and reveals that a CLI + Skill improvement helps large models but breaks smaller ones—a finding that would have shipped unseen with standard evaluation.
Chinese Threat Actor UNC6508 Targets Medical REDCap Servers for Espionage
A sophisticated PRC-linked espionage campaign compromised medical research REDCap servers for over a year, using a custom malware called INFINITERED to harvest credentials and then abusing enterprise email compliance rules for silent data exfiltration. A detailed breakdown of the attack chain, malware mechanics, and defensive takeaways for any organization running shared research platforms.
Near-autonomous AI improves challenging Chan-Lam coupling for sulfonamides
OpenAI and Molecule.one connected GPT-5.4 to an autonomous chemistry lab to improve Chan–Lam coupling of primary sulfonamides. Across 10,080 reactions, the system identified TEMPO as an effective additive, raising yields significantly. Human chemists validated results at bench scale, confirming the AI-driven finding holds in practical workflows.
MolmoMotion: Language-guided 3D motion forecasting from video
MolmoMotion predicts future 3D point trajectories from a single video frame, language instruction, and query points, using a class-agnostic, view-stable representation. It outperforms existing forecasting methods on a new benchmark, accelerates robot policy learning (76.3% vs 56% pick-and-place success), and improves video generation motion quality. The model, 1.16M-video dataset, and benchmark are openly released.
GLM-5.2: Open 1M-Context Model for Long-Horizon Agent Tasks
GLM-5.2 is the first open-source model to make 1M-context genuinely usable for long-horizon coding agents, trailing only Opus 4.8 on FrontierSWE and Terminal-Bench 2.1 while beating GPT-5.5 and Gemini 3.1 Pro. Its IndexShare architecture cuts per-token FLOPs by 2.9× at 1M context, and the anti-hack RL training stops reward hacking without crashing rollouts.
How Siemens Modernized Legacy Code with Agentic Workflows
Siemens and Google Cloud's Knowledge Fabric shows how to tame massive industrial codebases by combining knowledge graphs, multi-agent workflows, and human oversight—turning a decade-old modernization nightmare into a manageable, scalable process.
Graph Technologies Underpin Yahoo’s System of Action
Yahoo's Seller Agent, built on Google Cloud's graph technologies, shows how a dual-graph architecture—a knowledge graph for grounding and a context graph for auditability—lets agentic AI operate at speed while maintaining regulator-grade transparency. The blueprint applies broadly to any high-stakes industry.
The Golden Age of AI Applications
The article argues that the golden age of AI applications hinges on mastering three disciplines—model selection, loop design, and evaluation—rather than just model performance, with regulatory and market signals from Fable, Nadella, and Salesforce reinforcing the shift.
Anthropic disputes US government suspension of Fable 5 over minor jailbreak
Anthropic’s statement reveals a critical clash between rapid AI deployment and government oversight: a minor, narrow jailbreak found in Fable 5 led to an abrupt export control suspension, even though similar capabilities exist in other models. The company argues that applying this standard across the industry would halt all new frontier model deployments, underscoring the need for transparent, technically-grounded regulation.