Section · official announcements, collected automatically
AI News: official announcements from the AI giants
The announcements the big AI companies publish on their own blogs and newsrooms: the title, date and a short excerpt exactly as they published them, with a link to the original. This is not our reporting; we do not rewrite or summarise. The list updates every hour.
Latest weekly digest
Week of 11–17 Oct 2026
11–17 Oct 2026: 68 official announcements from 11 companies. Amazon 13, OpenAI 12, Apple 9, Google 9, Hugging Face 8, NVIDIA 8, Anthropic 3, Microsoft 2, xAI 2, Meta 1, Mistral AI 1.
Weekly archive
- 18–24 Oct 2026 11 items · in progress
- 11–17 Oct 2026 68 items
A week runs Sunday to Saturday. Every Sunday at 06:00 (Israel time) the week that ended on Saturday is closed and published automatically.
The live feed
Last update: 20 Oct 2026, 07:33 (Israel time)Announcements from Apple
10- AppleLocal and Distributed Large LLM Inference with MLX and RDMA on Mac Studio
Apple Machine Learning Research · machinelearning.apple.com
- AppleSelection-Based Structured Reasoning: Toward Efficient Multimodal Search Agents
Multimodal agents commonly generate freeform reasoning before each action. For small models, limited model capacity can result in lengthy reasoning that provides little useful guidance for action generation while incurring substantial inference…
Apple Machine Learning Research · machinelearning.apple.com
- ApplePlan or Retrieve First? A Scale Study of Tool Discovery Architectures
Tool-augmented LLM agents discover relevant APIs from large registries before planning execution, yet finite context windows prohibit injecting all tool definitions into a single prompt. The dominant approach, which we term Retrieve-then-Plan…
Apple Machine Learning Research · machinelearning.apple.com
- AppleCode-Switching Detection via Language-Selective Dimensions in Speech Foundation Models
Code-switching (CS), the alternation between languages within an utterance, is common in multilingual speech but underrepresented in pretraining corpora. Increasing CS coverage requires finding relevant utterances in large speech collections so that…
Apple Machine Learning Research · machinelearning.apple.com
- AppleThe Design Space of Tri-Modal Masked Diffusion Models
Discrete diffusion models have emerged as strong alternatives to autoregressive language models, with recent work initializing and finetuning a base unimodal model for bi-modal generation. Diverging from previous approaches, we introduce the…
Apple Machine Learning Research · machinelearning.apple.com
- AppleMix, Don't Tune: Bilingual Pre-Training Outperforms Hyperparameter Search in Data-Constrained Settings
For most languages of the world, language model pre-training operates in a data-constrained regime where models must repeat their training data many times, degrading generalization. Two remedies exist: aggressive hyperparameter tuning such…
Apple Machine Learning Research · machinelearning.apple.com
- AppleLarge Language Models for Machine Translation Quality Annotation: Humans and Models Are Both Challenged
Large Language Models (LLMs) are considered to be a more efficient and cost-effective alternative to human judgment for Machine Translation (MT) evaluation. With MT evaluation spanning a large number of language pairs,…
Apple Machine Learning Research · machinelearning.apple.com
- AppleMinions: A Dynamic and Distributed Virtual GPU Cloud Built on Idle Apple Silicon Devices
The increasing demand for GPU-accelerated machine learning inference has exposed limitations in centralized GPU infrastructures, including high cost, limited elasticity, and long wait times under resource constraints. At the same time, modern…
Apple Machine Learning Research · machinelearning.apple.com
- AppleSelf-Compacting Language Model Agents
Long agent traces composed of chains of thought and tool calls accumulate stale content that anchor subsequent generations, and eventually outgrow the context window. Existing scaffolds mitigate it with fixed-interval compaction triggered…
Apple Machine Learning Research · machinelearning.apple.com
- AppleNormalizing Trajectory Models
Diffusion-based models decompose sampling into many small Gaussian denoising steps, an assumption that breaks down when generation is compressed to a few coarse transitions. Existing few-step methods address this through distillation, consistency…
Apple Machine Learning Research · machinelearning.apple.com
Sources: OpenAI News · Google Blog · AI · Google DeepMind Blog · Google Research Blog · Anthropic Newsroom · Meta Newsroom · AI · Microsoft Source · AI · NVIDIA Blog · Generative AI · Apple Machine Learning Research · Mistral AI News · xAI News · Hugging Face Blog · AWS Machine Learning Blog. Titles and excerpts belong to the companies.