Apple ML researchers published a NeurIPS paper introducing SELFCOMPACT, a scaffold that lets language model agents decide for themselves when and how to compact their own accumulated context, rather than relying on fixed-interval summarization triggered by a token count. It combines a compaction tool the model can call with a lightweight rubric telling it when to fire (a sub-task is resolved) and when to hold off (mid-derivation or stuck). Tested on six benchmarks and seven models, it matched or beat fixed-interval summarization while using far fewer tokens.
A new Apple research paper introduces Minions, a system that pools unused GPU capacity across Apple Silicon Macs into a single distributed cloud for running machine learning inference. Instead of buying additional GPU hardware, teams could tap spare compute sitting idle on existing macOS devices. The authors say the design handles task scheduling, failure recovery, and efficient use of available resources, offering an alternative when dedicated GPU capacity is scarce or expensive.
AWS released Part 1 of a two-part guide showing how to build a post-quantum software supply chain using ML-DSA signatures (NIST FIPS 204) across AWS KMS, AWS Private CA, and IAM Roles Anywhere. The guide covers signing binaries, container images, and SBOMs, and is positioned as the first full cloud-provider walkthrough for post-quantum signing, from root-of-trust creation to policy enforcement. It responds to upcoming mandates: the NSA's CNSA 2.0 requiring exclusive quantum-resistant signatures by 2030, and the EU Cyber Resilience Act taking full effect in December 2027.
AWS published a follow-up security guide explaining how organizations can enforce which ML-DSA (post-quantum signature) parameter sets are allowed across their AWS environment. It covers using Service Control Policies (SCPs) at the organization or OU level and IAM permission boundaries to block creation and use of non-approved ML-DSA variants (e.g., blocking ML-DSA-44 while allowing ML-DSA-65/87). The guide also covers AWS Private CA and certificate controls, and discusses the trade-off between deny-lists (default-allow for new variants) and allow-lists (default-deny, requiring policy updates for new approved algorithms).
Amazon Web Services researchers published a method called 'priming' that converts pretrained Transformer models into hybrid architectures mixing attention and state space model (SSM) layers, reusing existing weights instead of retraining from scratch. The approach cuts pretraining costs by over 99% and lets a 32-billion-parameter hybrid model match its source Transformer's quality while decoding up to twice as fast. Amazon tested three SSM types and released an open-source model zoo, training code, and serving tools under Apache 2.0.
NVIDIA, working with Princeton's Dao AI Lab, built ReplaySSM, a new way to store state in hybrid models that mix state-space (SSM) layers with attention layers. Instead of rewriting the entire state on every token, the method keeps a periodic checkpoint plus a short list of recent inputs, splitting the job of producing output from the job of updating memory. The result is roughly half the memory traffic, decode kernels running up to 1.80x faster at scale, and speculative decoding throughput more than doubling on 120B-class models, all while keeping accuracy close to full precision. The method already ships inside TensorRT-LLM, vLLM, and FlashInfer, and is running in production for GLM-5.3-Flash and Kimi-K3.
At the OCP Global Summit, NVIDIA contributed its Vera Rubin rack and compute tray designs to the open MGX architecture and announced a series of partnerships covering AI factory power, cooling and networking standards. Partners include hyperscalers such as Meta, Google, Microsoft and AMD, plus Cisco and Visionbay. The announcements cover rack design specs (MGX 1.2), new open-source software for managing AI data centers (DSX OS components), grid resilience standards, an 800VDC power architecture, new Ethernet switching (Spectrum-6) and a new optical transceiver standard (USFP) with Cisco.
Google has pulled gemini-3.7-flash from active use, with every call to that model string now redirected automatically to gemini-3.8-flash. On the agent side, the deep-research-pro-preview-12-2025 variant is on its way out too, with a hard shutdown date of October 23, 2026, pushing developers toward two newer Deep Research options.
Mistral AI has released Voxtral Mini 4B Realtime Arabic, a streaming speech-to-text model fine-tuned for Arabic dialects, Modern Standard Arabic, and code-switching between languages. The 4.4-billion-parameter model processes 16 kHz audio in real time with a configurable transcription delay, achieving an 8.82% average Character Error Rate across seven Arabic benchmarks at a 480ms delay — close to the offline Voxtral Transcribe Arabic model's 7.91% CER. It was co-developed with Morocco's digital transition ministry as part of a sovereign AI partnership and is released under Apache 2.0.
Anthropic has introduced a beta option on its Claude Platform that allows a Managed Agents agent to construct its own workflow for large, multi-part jobs, such as going through hundreds of documents. Instead of handling everything in one pass, the agent splits the task into phases, delegates pieces across multiple sub-agents, and merges their output into a single result. The job is handed off to the server, which executes it asynchronously and reports progress through a stream of events.
AWS published a joint engineering write-up with Postman describing how Agent Mode, Postman's AI-native interface for API testing and documentation, was built to run in production at scale. Rather than model quality, the team found that tool sprawl and missing context were the biggest obstacles to reliable agent behavior. The post details architectural patterns including dynamic tool scoping, schema-based query access, dedicated context handlers, and how Amazon Bedrock provides model flexibility, cross-Region inference, data residency controls, and prompt caching.
NVIDIA described a workflow built into Megatron Core that makes large-scale pretraining bitwise deterministic, meaning independent or checkpoint-resumed training runs follow the exact same numerical trajectory. Using tensor fingerprinting across training phases, modules, operations and kernels, engineers can trace nondeterminism down to a specific kernel and fix it. On a trillion-parameter Nemotron model running on 2,432 GPUs, optimizations brought the performance cost of enforcing determinism (the 'determinism tax') down to about 2%, while keeping bitwise-identical results over 800 training steps.
Anthropic has announced the Cyber Mission, a long-term initiative to help defenders secure operational technology and open-source code using Claude models. It includes the Critical Infrastructure Defense Program, built with partners like Accenture, CrowdStrike, Deloitte and Rockwell Automation, and the free OSS Scanner tool that scans open-source projects for vulnerabilities. The effort builds on lessons from the earlier Project Glasswing, now merged into an expanded Cyber Verification Program.
AWS published a case study on Amazon Bedrock AgentCore payments, a managed feature that lets AI agents autonomously pay for services like model inference, one request at a time, using the x402 payment protocol. Incarna used it to let its agents pay BlockRun, a multi-provider inference router, for each individual model call, settling in USDC on the Base blockchain. AWS says Incarna cut integration time from a scoped two to three months down to three days and about 200 lines of code. During the beta, agents processed over 1,000 payments ranging from $0.001 to $0.05 per call.
Google introduced AQuA (Ambient Quality Agent), a tool that runs continuously beside production AI agents inside a Google Cloud project to find failures that don't show up in logs or health checks, such as skipped validation steps or dropped context between sub-agents. It samples production session traces, grades them against a checklist, clusters recurring failure patterns, verifies clusters against full transcripts, and tracks them over time as new, recurring, or resolved. On request, it can also trace a failure back to the exact line of code or prompt responsible, using an immutable source snapshot captured at deploy time.
Apple researchers, with collaborators from the University of Pennsylvania and UIUC, introduced a new generative modeling technique called Normalizing Trajectory Models (NTM) in a paper accepted at NeurIPS. Unlike typical diffusion models that rely on many small Gaussian denoising steps, NTM treats each step as a flexible, invertible transformation that preserves exact probability calculations — something most fast few-step generation methods give up when they rely on distillation or adversarial training. According to the paper, NTM can match or beat strong text-to-image baselines using only four sampling steps.
Google Cloud announced the Gemini agent, a unified agent that combines chat, knowledge work, content creation, and coding into one interface accessible across devices and channels. It integrates directly inside Gmail, Docs, Sheets, Slides, Chat, and Calendar, connects to enterprise tools and databases via MCP and a tools registry, and orchestrates across multiple models — including Google's Gemini family and Anthropic's Claude — to balance quality and cost. Google also detailed security/governance features, cost controls like Smart Routing and spend caps, and shared customer results from companies like BNP Paribas, Bradesco, and Orange Spain.
Cisco Talos' CAIRN research classifies a new malware archetype, 'A3: AI-Analysis Evasion,' where malicious code embeds natural-language text designed to manipulate language models used in automated malware triage. Talos traced the technique across four malware families (FRUITSHELL, PLOTSAFE, HOLLOWCLAD, MANTLEMAZE) and 84 samples from January 2025 to July 2026, showing it spreading from a simple copy-pasted comment to templated, multi-model-targeted 'template spraying' and fake intimidation messages. Controlled testing with a panel of five local LLMs found the simplest direct-instruction technique shifted model verdicts toward 'benign' in roughly 35% of runs, while more elaborate tricks had little effect or backfired by increasing suspicion.
Cisco Talos documented a threat actor tracked as UAT-11985 running a spear-phishing campaign against Taiwan research organizations, impersonating academic and policy institutions using fabricated event invitations. The emails show patterns consistent with AI-assisted content generation, allowing the actor to rapidly customize lures while keeping a consistent social-engineering structure. The campaign also used quishing (malicious QR codes on altered event posters) and a real-time adversary-in-the-middle (AitM) phishing kit that impersonates Google sign-in pages to intercept credentials and MFA challenges, harvesting full authenticated session tokens. Talos assesses with moderate confidence the phishing kit's interface was originally built in Simplified Chinese before being localized to Traditional Chinese and English.
Cisco Talos' Vulnerability Discovery & Research team disclosed a set of vulnerabilities across Adobe Photoshop, Apple macOS, Foxit Reader, and Microsoft Windows, all of which have already been patched by their respective vendors. The flaws range from privilege escalation and information disclosure to remote code execution and use-after-free bugs. Snort rule sets are available to detect exploitation attempts of these issues.
Google has released the Developer Knowledge API, a structured, machine-readable source for Google Cloud, Firebase and Android documentation aimed at AI agents and developer tooling. Instead of scraping web pages or relying on an LLM's training cutoff, agents and CLIs can now query fresh, Markdown-formatted docs through several interfaces: a gcloud CLI surface, an installable agent skill, client libraries, and a browser-based API explorer. The underlying index is refreshed frequently so retrieved answers stay close to the latest published documentation.
AWS describes a new two-stage access control architecture for Retrieval Augmented Generation (RAG) in Amazon Quick and Amazon Bedrock Knowledge Bases. Instead of relying only on periodically synced access control lists (ACLs), the system now verifies permissions in real time against the authoritative data source (e.g. Google Drive, SharePoint, Confluence) at query time, before passing content to the LLM. AWS says this closes security gaps caused by stale or incorrectly mapped permissions between sync cycles. Mondelēz International, with over 35,000 employees, is cited as a customer using this approach.
Anthropic has launched Claude Haiku 5.5 (claude-haiku-5-5), positioned as its most capable Haiku model for high-volume, latency-sensitive work. It supports a 1 million token context window, up to 128k max output tokens, and adaptive thinking controlled via a new effort parameter. The model is available on the Claude API, Amazon Bedrock, Claude Platform on AWS, Google Cloud, and Microsoft Foundry. Anthropic warns that code written for the previous Haiku 4.5 model may break on the new version.
Cloudflare built an AI agent system for its Managed Defense security service to handle the flood of alerts that overwhelm human analysts. Instead of one general-purpose agent analyzing everything, the system separates deterministic data collection from model-based reasoning and runs four specialist agents in parallel before a synthesis agent combines findings. The company says this reduces hallucinations and makes investigations auditable and reproducible.