20 articles — 🔴 2 critical , 🟡 12 important , 🟢 6 interesting
Google announced the general availability of the gemini-embedding-2 model, which supports text, images, video, audio, and PDF inputs mapped into a unified embedding space. The model was in preview since March 10, 2026, and is now available to everyone via the Gemini API.
Microsoft Research introduced AutoAdapt, a framework that automates the adaptation of general language models to specialized domains such as medicine, law, and incident response. The system autonomously chooses between RAG and fine-tuning, optimizes hyperparameters, and completes the job in approximately 30 minutes at an additional cost of around $4.
Apple's research team at ICLR 2026 has introduced MANZANO, a unified multimodal framework that addresses a long-standing trade-off between image understanding capabilities and image generation quality. The model uses a hybrid vision tokenizer that produces continuous embeddings for understanding and discrete tokens for generation, a shared encoder, and two specialized adapters — reducing the quality loss that typically occurs when a single model attempts to do both tasks.
Apple presented five research papers at ICLR 2026 in Rio de Janeiro, with the most notable being ParaRNN — a method enabling parallel training of nonlinear recurrent neural networks with a 665× speedup over sequential approaches, scaling RNNs to billions of parameters to compete with transformers.
Linux Foundation AI & Data published a practical guide for implementing the RGAF (Responsible Generative AI Framework) across nine dimensions of responsible AI, with a catalog of 35 concrete open-source tools and alignment with NIST AI RMF, EU AI Act, ISO/IEC 42001, and OECD principles.
Google DeepMind has signed a partnership with five of the largest consulting firms — Accenture, Bain, BCG, Deloitte, and McKinsey — to accelerate enterprise AI transformation, given that currently only 25 percent of organizations manage to deploy AI to production.
OpenAI introduced Workspace Agents, Codex-powered AI agents integrated directly into the ChatGPT interface. The agents run in the cloud, automate complex workflows, and help enterprise teams scale work through connected tools with an emphasis on cross-application security.
AWS has published an architecture that combines Amazon Bedrock, the Neptune graph database, and the Mem0 framework for persistent AI agent memory at the company-wide level, solving the problem of context loss between sessions and users.
Amazon announced a managed agent harness for Bedrock AgentCore that enables deploying a fully working agent in just three API calls, without writing any orchestration infrastructure. The harness is accompanied by the AgentCore CLI covering the full development cycle and pre-built skills for coding assistants, available in preview across four AWS regions.
SWE-chat has been published on ArXiv — a dataset of real so-called in-the-wild interactions between users and AI coding agents in production environments. Rather than another synthetic benchmark based on GitHub issues, this dataset captures how developers actually use autonomous systems during their everyday work — what they ask for, how they respond to the agent's suggestions, and where the agent fails — opening the door to more precise evaluation and targeted improvements in agent design.
New research shows that AI agents for computer control that successfully complete a task once may fail on an identical repeated attempt, with three key reasons being execution stochasticity, task specification ambiguity, and agent behavior variability.
NVIDIA and Google Cloud announced a joint collaboration to accelerate agentic AI and physical AI workloads, combining NVIDIA GPU infrastructure with the Google Cloud platform for robotics, autonomous systems, and agents.
NVIDIA and HuggingFace demonstrated Gemma 4 as a Vision Language Agent that autonomously decides when to use the camera and runs the entire pipeline, including speech-to-text and TTS, locally on an NVIDIA Jetson Orin Nano Super with 8 GB of memory, with no cloud dependency.
The AWS Machine Learning blog described how to use NVIDIA's open-source Parakeet-TDT-0.6B-v3 model for cheap multilingual audio transcription in the cloud. The model covers 25 European languages with automatic detection, and when combined with AWS Batch, processing one minute of audio costs just $0.00005 on Spot instances, or $0.00011 on on-demand g6.xlarge GPUs, with a scale-to-zero policy and the ability to process audio recordings longer than ten hours through buffered streaming.
Amazon SageMaker AI now automatically benchmarks generative AI models across different GPU configurations using the NVIDIA AIPerf tool, eliminating weeks of manual testing and providing recommendations ranked by cost, latency, or throughput.
GitHub has enabled Copilot Business and Enterprise users to bring their own API keys for major providers including Anthropic, Google, OpenAI, OpenRouter, and Azure in VS Code. BYOK models work within Copilot Chat and custom agents, with billing going directly to the chosen provider without consuming Copilot quota.
GitHub's latest updates to the Copilot cloud agent for Atlassian's Jira introduce a set of features that significantly deepen AI integration into project management. Teams using Jira as a task tracking system can now define their own custom agents, use Atlassian custom fields in rules, set customized branching rules per space, and receive code review request notifications directly in Jira — connecting the development flow between GitHub and the project management tool.
OpenAI has enabled free access to ChatGPT for verified physicians, nurses, and pharmacists in the US. The program focuses on clinical documentation, patient care workflows, and medical research, with verification conducted through partnerships with American medical entities.
