Wednesday, July 15, 2026

Percona highlights MySQL 9.7 as a new Long-Term Support release designed to replace the End-of-Life MySQL 8.0. The update introduces features across multiple technical domains aimed at enhancing operational stability. The release is positioned as the next stable foundation for production workloads following the previous LTS cycle.

• MySQL 8.0 has reached End-of-Life, making 9.7 the recommended LTS path.
• Release focuses on enhanced stability for long-term production deployments.
• Feature set spans multiple technical domains within the MySQL environment.
• Percona provides evaluation of practical utility for database administrators.

LakeQuest is a 9,846-pair benchmark designed to evaluate question answering systems on noisy, weakly structured data lakes rather than clean corpora. It spans AI/ML metadata, retail banking, and biomedical domains, requiring models to navigate heterogeneous tables and linked metadata. The benchmark focuses on the full retrieve-and-synthesize pipeline, addressing the gap where current tests ignore the complexity of real-world data discovery.

• Evaluates end-to-end performance over heterogeneous, weakly structured data rather than clean schemas.
• Covers three distinct domains: AI/ML metadata, retail banking, and biomedical drug information.
• Requires systems to handle noisy discovery of tables, passages, and linked metadata.
• Human-validated pairs ensure realistic assessment of retrieve-and-synthesize pipelines.

AWS has expanded Amazon Aurora DSQL availability to the Europe (Spain) Region, supporting single-Region clusters. This serverless distributed SQL database offers active-active high availability and multi-Region strong consistency. It is designed to provide virtually unlimited scalability and zero infrastructure management for always-available applications.

• Aurora DSQL is now live in Europe (Spain) for single-Region deployments.
• Features include active-active HA and multi-Region strong consistency.
• Serverless architecture eliminates infrastructure management overhead.
• Enables effortless scaling and resilience for distributed SQL workloads.
• Expands the global footprint of fast, serverless distributed SQL capabilities.

With pgAgent on its way out, Regina Obe is guiding the migration to pg_timetable by detailing how to install it as a Linux service. This second part of the series focuses on the operational configuration required to run the scheduler reliably. Obe is also developing a pg_timetable UI for pgAdmin that mirrors the familiar pgAgent interface to ease the transition for users.

• pgAgent is being deprecated, making pg_timetable the recommended replacement for job scheduling.
• New guidance is available for installing pg_timetable as a persistent Linux service.
• A pgAdmin UI for pg_timetable is in development to replicate the existing pgAgent experience.
• The new UI pull request is currently being cleaned up before final commit.

The Odyssey connection pooler for PostgreSQL and Apache Cloudberry has reached version 1.5.1, bringing extended protocol support and performance improvements for pipelining. New capabilities include shared pools, transaction pooling pinning, and explicit CPU affinity support. The release also adds observability features like queue time metrics, Prometheus exporter enhancements, and SSLKEYLOGFILE support.

• Improves pipelining performance and fixes protocol violations for better stability.
• Adds queue time metrics (avg/total wait) and pool_notice_after_waiting_ms for debugging.
• Supports CPU affinity and shared pools for more granular resource control.
• Includes SSLKEYLOGFILE support to aid in TLS traffic analysis and debugging.

Researchers have isolated a specific mechanism in Llama-3.1-70B-Instruct that enables precise token counting across diverse tasks. This countdown subcircuit compares the current token position against a target length to estimate remaining tokens. The study reveals that this motif is shared across various applications requiring exact length control, such as sentence generation and sequence formatting.

• LLMs use a reusable subcircuit for precise token counting tasks.
• Mechanism compares current position to a goal length estimate.
• Identified in Llama-3.1-70B-Instruct via controlled experiments.
• Explains consistent performance in fixed-length output generation.
• Shared motif suggests generalizable counting architecture in models.

PrismML has released Bonsai 27B, a large language model with 27 billion parameters capable of operating on smartphones. The model represents a significant step in optimizing large models for edge deployment without requiring cloud infrastructure. This release highlights ongoing efforts to shrink model size while maintaining performance for on-device applications.

• 27B parameter count is unusually large for on-device inference
• Enables local AI processing without cloud dependency
• Targets mobile hardware constraints specifically
• Signals shift toward edge-native LLM architectures
• Reduces latency and privacy concerns for end users

Researchers demonstrate that training decoder-only transformers on chronologically filtered data can significantly reduce the performance lag of point-in-time language models. By scaling to 4 billion parameters and 1 trillion tokens from FineWeb, they eliminate future information leakage that compromises causal inference in finance and social sciences. The study produces a sequence of monthly checkpoints spanning from 201, showing that scale is a viable strategy to bridge the gap with unconstrained models.

• Chronological filtering prevents future data leakage, ensuring valid backtests and causal inference.
• Scaling to 4B parameters and 1T tokens substantially narrows the performance gap of PIT models.
• Monthly checkpoints allow reconstruction of model capabilities at specific historical dates.
• This approach enables temporal analysis without the bias inherent in unrestricted internet corpora.

Researchers introduce CANDI-QA, a new benchmark designed to evaluate large language models in specialized fields like medicine and finance. Unlike traditional QA datasets, this resource focuses on contextual grounding, user awareness, and domain-specific accuracy. The dataset consists of expert-curated question-answer pairs structured to test precise information extraction and nuanced understanding.

• New benchmark moves beyond general knowledge to test domain-specific LLM capabilities.
• Evaluates contextual grounding and user alignment in sensitive fields like healthcare.
• Features expert-curated pairs focusing on precise information extraction tasks.
• Addresses gaps in traditional QA benchmarks for specialized professional use cases.

The MAGE framework provides a controlled environment to analyze how components like episodic memory and multi-objective selection interact during iterative prompt optimization. Researchers identified the Prompt Optimization Coupling Effect (POCE), where multiple stochastic signals in a reflective loop simultaneously boost performance and amplify variance. This study serves as an ablation platform rather than a claim of superior absolute optimization.

• MAGE is an analysis framework, not a standalone optimizer proposal.
• POCE shows combined signals improve performance while increasing variance.
• Episodic memory and Pareto selection interact in closed reflective loops.
• Results highlight trade-offs in multi-component prompt optimization strategies.

This paper addresses the high inference costs of cross-encoders in RAG pipelines by fine-tuning LLaMA 3 (8B) as a drop-in reranker. The process involves supervised fine-tuning with LoRA adapters using the Unsloth framework, followed by 4-bit quantization to optimize performance. Evaluated on a domain-specific QA benchmark, the model significantly improves answer relevancy, context precision, and answer accuracy compared to standard retrievers.

• Replaces expensive cross-encoders with a quantized 8B LLM for real-time RAG reranking.
• Achieves 14% gain in answer relevancy and 16% in context precision via distillation.
• Uses Unsloth and LoRA for efficient fine-tuning, then 4-bit quantization for deployment.
• Validated on domain-specific benchmarks using RAGAS framework for rigorous metrics.

Researchers propose TAKE, a framework that distills large-scale text corpora down to 0.1% of their original size while maintaining downstream task fidelity. The method leverages influence functions to quantify each sample's contribution to the training objective. By convolving knowledge-based influence along the training trajectory, TAKE generates a single per-sample score to identify and retain the most informative data points.

• Reduces training data volume to 0.1% of original size, significantly lowering storage and compute costs.
• Uses influence functions to measure individual sample contribution to downstream objectives.
• Convolve influence along training trajectory to capture dynamic informative samples.
• Enables efficient continual learning by retaining only high-impact text data.

This research paper introduces a framework that decomposes image-based retinal diagnoses into structured argumentation components based on the Toulmin model. Instead of treating ML predictions as final claims, the system extracts biomarkers as grounds, uses a MedGemma agent to validate the warrant, and defines qualifiers and rebuttals. This approach shifts from black-box outputs to a structured, interpretable diagnostic assistance model.

• Decomposes ML predictions into Toulmin's claim, grounds, warrant, qualifier, rebuttal, and backing.
• Uses specialized models for biomarker extraction (grounds) rather than raw image features.
• Employs a MedGemma agent to analyze the warrant linking evidence to the diagnostic claim.
• Enables structured interpretability by explicitly modeling qualifiers and potential rebuttals.

SupplyNetPy is a new open-source Python library designed for modeling and simulating arbitrary multi-echelon supply chain networks. It supports discrete-event simulation with features like perishable inventory, node disruptions, and stochastic demand. The tool allows users to define networks as graphs and provides detailed performance reports, validated against analytical benchmarks and commercial tools.

• Open-source Python library for high-fidelity supply chain simulation
• Supports arbitrary multi-echelon structures and discrete-event modeling
• Handles perishable inventory, disruptions, and stochastic lead times
• Validated against commercial tools and analytical benchmarks
• Extensible architecture via inheritance for custom components

AWS has added OpenAI's bidirectional privacy-filter model to SageMaker JumpStart, enabling one-click deployment for PII detection and masking. The model performs context-aware token classification in a single forward pass, identifying spans like emails, names, and secrets. This allows engineering teams to build high-throughput data sanitization workflows directly on AWS infrastructure.

• OpenAI's bidirectional model now available via SageMaker JumpStart for easy deployment.
• Detects PII spans including emails, names, secrets, and account numbers in one pass.
• Designed for high-throughput sanitization workflows on AWS or on-premises.
• Tunable architecture supports context-aware masking for sensitive text data.

AWS has added Qwen3-VL-Embedding-2B and Qwen3-Reranker-4B to SageMaker JumpStart. The embedding model handles initial recall for mixed modalities like text, images, and video. The reranker then refines these results for higher precision retrieval.

• Qwen3-VL-Embedding-2B supports text, images, screenshots, and video inputs for cross-modal search.
• Qwen3-Reranker-4B improves precision by re-ranking initial recall results from the embedding model.
• Both models are available via SageMaker JumpStart for immediate deployment on AWS infrastructure.
• This pairing enables building comprehensive, multi-stage retrieval pipelines with specialized capabilities.

AWS has added Google DeepMind's Gemma-4-E2B-it to SageMaker JumpStart, offering a multimodal model optimized for local execution. The model supports text, image, and audio inputs with native function calling and a step-by-step reasoning mode. It enables agentic workflows and handles complex tasks like OCR, code generation, and UI understanding.

• Gemma-4-E2B-it is available via SageMaker JumpStart for rapid deployment
• Model optimized for efficient local execution on AWS infrastructure
• Supports multimodal inputs including text, image, audio, and video
• Includes built-in reasoning mode for step-by-step problem solving
• Enables agentic workflows through native function calling capabilities

DoorDash engineered Ask DoorDash by integrating large language models with specialized agents and MCP-based tooling, supported by an intelligence layer handling persistent consumer memory and live backend data. This hybrid approach moves beyond pure generative text to execute precise shopping tasks with real-time context. Early metrics indicate a 24% lift in checkout conversion and a 17% increase in basket size, driven by improved intent accuracy through memory-backed sessions.

• Combines LLMs with specialized agents and MCP tooling for precise task execution.
• Persistent consumer memory and live backend data drive intent accuracy.
• Resulted in 24% higher checkout conversion and 17% larger average baskets.
• Architecture prioritizes deterministic tool use over pure generative responses.

The AWS Lambda console now offers a one-click prompt to configure coding agents with AWS Serverless skills and the Serverless Model Context Protocol (MCP) server. This update embeds serverless best practices directly into the agent configuration, removing the need to manually navigate documentation for setup. Developers can initiate this streamlined process at any stage of their Lambda journey, from initial exploration to creating their first function.

• Eliminates manual configuration friction for coding agents in the Lambda console.
• Automatically installs AWS Serverless skills and Serverless MCP server.
• Embeds serverless best practices directly into agent workflows from day one.
• Available across all Lambda console entry points for new and existing users.

Public battery aging datasets often suffer from inconsistent formats and scattered metadata, making them difficult to use for health management. BatteryLake addresses this by implementing a governed data lakehouse that uses LLM agents to extract metadata and synthesize dataset-specific converters. This agentic, physics-grounded framework ensures outputs are tied to verbatim evidence, automatically abstaining when no supporting data exists.

• Solves fragmentation in public battery aging data through automated, reproducible curation.
• Uses LLM agents to generate converters grounded in verbatim evidence from source papers.
• Mitigates hallucination risks by having agents abstain when no metadata evidence is found.
• Provides benchmark-ready assets by handling domain-specific semantics of electrochemical time-series.

Multi-agent and memory-augmented LLM systems often place coordination content, shared state, prior discussion, tool outputs, summaries, and role instructions, inside the same finite prompt used for the current task. This creates a practical allocation problem: every token spent on coordination is unavailable to task instructions or evidence when a call is assembled under a fixed context budget. The authors introduce the Roundtable Context Window Test (RCWT) to measure this displacement by varying coordination content while controlling total budget, position order, task family, and scoring.

• Coordination tokens directly displace task-relevant evidence in fixed-context calls.
• RCWT isolates this effect by varying coordination content while holding budget constant.
• Protocol controls for position order, task family, and scoring to ensure valid metrics.
• Current commercial models show baseline recall retention under moderate displacement at W=4096.

This research addresses the latency and data-sensitivity constraints of deploying large language models for industrial closed-loop control by fine-tuning compact Small Language Models like Qwen2.5-3B. The proposed framework replaces opaque, high-latency cloud inference with a local agent that generates control policies from natural language specifications. To ensure safety, the system employs a multi-agent self-correction loop where a plant-aware validator, such as a digital twin, checks candidate actions before execution.

• Fine-tuned SLMs offer a viable path for low-latency, edge-native industrial control automation.
• Natural language policy generation reduces manual redesign effort for control systems.
• Validator-guided correction loops mitigate hallucination risks in autonomous decision-making.
• Digital twin integration enables safe pre-execution verification of AI-generated actions.

Evaluating context-reduction layers solely by text removal ignores actual billing metrics. A pre-specified campaign of 2,908 Claude Code runs analyzed whether compressing retrieved context or tool outputs truly lowers costs without hurting success rates. The study compares baseline execution against hook-based compression and API-boundary proxies across multiple models and repositories. Results challenge the assumption that fewer tokens automatically translate to lower provider bills.

• Token count is a poor proxy for actual API billing in coding agents.
• Context compression may fail to reduce costs even if text volume drops.
• Success rates and trajectory length must be monitored alongside cost.
• API-boundary proxies offer a distinct optimization vector compared to hooks.
• Evaluated 2,848 runs across 103 tasks and seven repositories.

Microsoft has released emergency software updates to address a record-breaking 570 security vulnerabilities across its product ecosystem. This massive patch cycle highlights the increasing complexity and attack surface of modern enterprise software stacks. Security teams are urged to prioritize these updates to mitigate potential exploitation vectors immediately.

• Microsoft addressed 570 vulnerabilities, setting a new record for a single update cycle.
• The volume of fixes suggests significant technical debt or coordinated discovery efforts.
• Prioritize patching to reduce exposure to zero-day exploitation attempts.
• Review release notes to identify critical flaws affecting your specific infrastructure.

AWS has integrated AI agent automation into Database Migration Service Schema Conversion through the AWS MCP Server. This allows coding agents like Kiro, Claude Code, and Cursor to autonomously execute migration workflows from within an IDE. The new dms-schema-conversion skill enforces predefined API patterns and operational rules to reduce trial-and-error during schema conversion.

• Connect IDE-based AI agents to DMS Schema Conversion using the AWS MCP Server
• Agents can autonomously create projects, browse metadata, and generate assessment reports
• Predefined skill procedures enforce operational sequencing to minimize trial-and-error
• Supports conversion of remaining code objects like stored procedures via natural language

Amazon EMR on EKS now integrates an Apache Spark troubleshooting agent that diagnoses job failures using natural language prompts. The system automatically analyzes Spark History Server data, distributed executor logs, and cluster configurations to pinpoint issues like memory errors, data skew, and resource contention. Users can access this feature directly from the console via a 'Troubleshoot with AI' button on failed jobs, eliminating the need to manually sift through logs.

• Diagnose Spark failures on EMR on EKS using natural language queries for root cause analysis.
• Agent automatically inspects executor logs, history server data, and cluster configs for issues.
• Covers memory errors, data skew, resource contention, and connectivity failures.
• Consistent troubleshooting experience now available across EMR on EC2, Serverless, and EKS.
• Access the feature directly via the 'Troubleshoot with AI' option in the EMR on EKS console.

AWS has extended Amazon GuardDuty to monitor AI-specific workloads in Bedrock and SageMaker. The service analyzes CloudTrail events to identify threats like prompt injection and anomalous model invocations. It also detects cost harvesting attacks where actors exhaust GPU time and tokens. This provides visibility into AI security without requiring custom tooling or manual configuration.

• GuardDuty now covers Bedrock and SageMaker for AI threat detection
• Detects prompt injection, anomalous invocations, and cost harvesting
• Analyzes CloudTrail management and data events automatically
• Reduces need for custom security tooling for AI workloads

AWS Security Hub now includes an AI inventory feature to provide centralized security teams with a continuous, organization-wide view of AI assets and their security posture. This capability addresses the visibility gap that arises as organizations rapidly deploy AI agents, models, and pipelines. The feature automatically discovers and catalogs AI workloads across the AWS environment to help secure assets that might otherwise go unnoticed.

• Centralized visibility into AI assets eliminates blind spots in security posture.
• Automatically discovers AI workloads without requiring additional configuration.
• Catalogs managed AI services via AWS Config resources from Amazon Bedrock.
• Includes Bedrock AgentCore and Amazon SageMaker in the inventory scope.
• Connects AI assets to active threats and misconfigurations for better defense.

Amazon Redshift Serverless now automatically maintains zero-ETL and Amazon S3 event integrations when restoring a namespace from a snapshot or recovery point. Previously, these integrations were marked as failed during restore, forcing manual recreation and risking data ingestion gaps. This update ensures integrations resume operation immediately after restore, simplifying disaster recovery and testing workflows.

• Snapshot restores now keep zero-ETL and S3 event integrations active without manual reconfiguration.
• Eliminates data ingestion gaps and administrative overhead during disaster recovery or testing.
• Reduces risk of errors associated with rebuilding complex data pipelines post-restore.

Amazon Managed Service for Apache Flink now integrates AI Agent Skills that provide coding assistants with expert, up-to-date guidance for building and operating Flink applications. These skills cover common tasks like troubleshooting, scaling, monitoring, and version upgrades, transforming specialized knowledge into a guided experience. This allows developers to independently manage application health and accelerate development using existing tools like Kiro.

• AI coding assistants now access expert Flink guidance for tasks like scaling and monitoring
• Simplifies upgrades to latest Apache Flink versions like 2.2 without deep expertise
• Reduces dependency on specialized knowledge by turning complex ops into guided workflows
• Compatible with existing AI agents like Kiro to boost developer productivity

Hallmark is a design skill for Claude Code, Cursor, and Codex that prevents AI-generated code from looking generic. It enforces a macrostructure, selects from twenty themes, and runs fifty-seven tests to reject on-distribution defaults. The tool uses four verbs to build new UIs or audit existing code against anti-patterns.

• Built for Claude Code, Cursor, and Codex to enforce consistent, non-generic design output.
• Uses fifty-seven slop-test gates and self-critique to reject standard LLM design defaults.
• Offers 'hallmark audit' to score existing code against anti-patterns without editing.
• Generates distinct layouts rather than simple color swaps or template variations.