Friday, July 10, 2026

Masahiko Sawada has committed a patch for PostgreSQL 20 that introduces min() and max() aggregate functions for the uuid data type. Since uuid already possesses a full set of comparison operators and a btree operator class, it is fully ordered. This update resolves the final gap by implementing the uuid_larger() and uuid_smaller() functions to support these aggregates.

• UUIDs are now fully supported in min/max aggregations without custom casting.
• Completes the aggregate coverage for uuid, matching other primitive types.
• Enables direct ordering and selection of earliest/latest UUIDs in queries.
• Requires PostgreSQL 20 to utilize the new uuid_larger/smaller functions.

Tencent has published Hy3, a new large language model available as open weights. The release includes technical details and resources for researchers and developers to evaluate and deploy the model. This entry adds another capable open-source option to the current landscape of foundation models.

• Tencent contributes Hy3 to the open-weights ecosystem
• Provides technical resources for model evaluation
• Adds competitive alternative to existing open LLMs

This paper identifies a failure mode in critic-free RL where uniform credit assignment reinforces erroneous low-probability tokens alongside plausible ones. The authors propose Tail-Aware Credit calibratiOn (TACO) to differentiate and suppress credit for these tail tokens. This calibration prevents the indiscriminate reinforcement of flawed reasoning patterns during model training.

• Uniform credit assignment in critic-free RL reinforces bad tail tokens equally to good ones
• TACO calibrates credit to suppress undesired low-probability token reinforcement
• Mitigates positive-credit contamination that degrades reasoning capabilities
• Addresses a critical gap in current LLM reinforcement learning methods

SQuaD-SQL enables small language models to rival large language models on Text-to-SQL tasks by using LLM-guided knowledge distillation and synthetic data generation. This approach addresses the high computational costs of deploying LLMs in resource-constrained environments. The method combines structured knowledge extraction with synthetic data to boost small model efficiency without sacrificing accuracy.

• SLMs can approach LLM performance on Text-to-SQL via distillation
• Reduces computational footprint for resource-constrained deployments
• Uses LLM-generated synthetic data to train smaller models
• Enables efficient natural language interaction with databases

This paper models in-context search as approximate inference where a base LLM provides a prior and self-reflection updates the posterior. The authors analyze the sampling complexity, defining the number of sequential attempts needed for high success probability. They prove that when reflections reliably pinpoint early mistakes, the process achieves exponential improvements over the base model's zero-shot pass rate.

• In-context search acts as posterior updating over reasoning traces using base model priors.
• Exponential performance gains occur when reflections accurately localize early reasoning errors.
• Polynomial sequential attempts can solve problems with exponentially small zero-shot pass rates.
• Theoretical framework quantifies the sampling complexity required for high success probability.

Pyligent introduces a training and inference framework that treats reasoning as a validated search over partial solution chains rather than a simple linear progression. The system uses a task validator to label continuations and failures, converting these search trees into supervised targets for three specific actions: continue, finish, and backtrack. This approach allows models to recover from delayed failures by backtracking to the most recent viable prefix, with optional traces summarizing abandoned branches.

• Shifts reasoning from linear chains to validated search trees with explicit backtracking.
• Uses supervised targets for continue, finish, and backtrack actions based on validator feedback.
• Evaluates performance on hidden directed graph tasks to isolate delayed-failure recovery.
• Generates optional traces to summarize abandoned reasoning branches for transparency.

Amazon SageMaker HyperPod now enables deep health checks for Slurm-orchestrated clusters using continuous provisioning. This feature allows proactive verification of GPU accelerator health on running instances before jobs commit resources. By targeting entire instance groups or specific nodes, users can run hardware stress and connectivity tests to prevent wasted compute time from single unhealthy nodes.

• Deep health checks validate GPU health and connectivity on running instances before job allocation.
• Continuous provisioning now pairs with hardware validation to prevent all-or-nothing scaling failures.
• Proactive testing avoids wasting hours of compute time on single unhealthy nodes.
• Users can target specific instances or entire groups for comprehensive stress testing.

AWS Neuron 2.31.0 introduces NKI 0.5.0 with MX FP8 scale dtype support, tensor indirection for optimized indexed access, and zero-cost NkiTensor view APIs. The release adds a public beta Neuron UltraServer Operator for Amazon EKS to automate Trainium UltraServer workload management. Additionally, the Neuron Compiler features a redesigned backend now default on Trn2 and Trn3, while the Runtime simplifies configuration with contiguous shared scratchpad support.

• NKI 0.5.0 adds MX FP8 scale dtype support and tensor indirection for fewer instructions.
• UltraServer Operator for EKS automates discovery and resource claims for Trainium workloads.
• Redesigned Neuron Compiler backend is now default on Trn2 and Trn3 for better performance.
• Neuron Runtime supports contiguous shared scratchpad, removing manual page size configuration.
• NkiTensor view APIs enable zero-cost tensor layout transformations in custom kernels.

Amazon SageMaker Feature Store now supports BatchWriteRecord, allowing data scientists to ingest multiple records across feature groups in a single request for higher throughput. The update also introduces ListRecords to discover stored data without pre-known identifiers and enables the creation of databases and tables with custom names in the offline store. These changes target reduced API call volume and lower latency for high-volume feature ingestion.

• Use BatchWriteRecord to reduce API calls and latency for high-throughput feature ingestion.
• Discover stored features via ListRecords without needing specific record identifiers.
• Create custom-named databases and tables in the offline store for better organization.
• Optimize training pipelines by batching writes to the online store simultaneously.

Amazon SageMaker Unified Studio Workflows now includes 19 new operators covering Amazon Bedrock, S3 Tables, S3 Vectors, AWS Glue Data Catalog, and MWAA Serverless. These additions enable users to orchestrate tasks like managing Bedrock guardrails, provisioning S3 resources, and triggering MWAA runs directly through the visual workflow creator. This expansion reduces the need to write custom integration code or switch between different AWS consoles.

• New operators allow visual orchestration of Bedrock guardrails without custom code.
• S3 Tables and Vectors resources can be provisioned and deleted via workflow tasks.
• Glue Data Catalog tables and databases are now manageable within SageMaker workflows.
• MWAA Serverless workflow runs can be triggered directly from the visual editor.
• Available in all AWS Regions, reducing context switching for data builders.

Prior research indicates that chain-of-thought reasoning is often unfaithful, meaning stated steps do not reflect the model's actual process. Since detecting unfaithfulness requires controlled interventions impossible to apply to existing transcripts, this work focuses on logical consistency instead. The authors introduce a reusable framework to assess whether stated reasoning aligns with the final answer using only transcript data.

• Detects unfaithful CoT by checking logical consistency between reasoning and answer
• Requires no experimental interventions, enabling post-hoc analysis of evaluation transcripts
• Provides a reusable method for auditing AI safety evaluation data
• Addresses the gap where faithfulness detection is impractical for static logs

This paper addresses latency waste in production LLM agents caused by regenerating procedural code on every request. It introduces an agentic tool-making pipeline that compiles repeated standard operating procedure steps into validated, versioned tools before deployment. By grounding synthesis in live execution traces and backend schemas, the system generates and repairs candidate tools, allowing runtime agents to call them directly instead of relying on inference-time coding.

• Replaces inference-time code regeneration with pre-compiled, versioned tools for repeated tasks.
• Tool synthesis uses live execution traces and backend schemas to ensure accuracy.
• Agents fallback to code generation only when pre-built tools are insufficient.
• Validated in a Fulfillment Center alarm-triage system with a 44-node SOP.

The AWS MCP Server now supports direct AI agent connections via AWS Sign-In and standard OAuth protocols, eliminating the need for custom authentication middleware. This update allows agents to leverage existing IAM permissions and governance controls while supporting both interactive browser-based and programmatic headless authorization flows. Administrators gain enhanced control through new IAM condition keys, token introspection APIs, and CloudTrail audit logging.

• Agents connect via standard OAuth without extra auth software
• Leverages existing IAM policies and AWS identities
• Supports interactive and headless authorization modes
• New IAM condition keys and token revocation APIs
• Full CloudTrail audit coverage for OAuth access

Researchers introduce DeepSearch-Evolve, a self-distillation framework designed to train tool-use agents that improve through their own experience. The system relies on DeepSearch-World, a deterministic environment featuring reproducible search and page-reading tools alongside 420K multi-hop QA tasks. This setup addresses the limitations of fixed teacher-distilled trajectories and sparse-reward reinforcement learning by enabling progress verification, grounded reflection, and failure recovery.

• DeepSearch-World provides a deterministic, verifiable sandbox for training web agents with reproducible tools.
• The dataset includes 420K multi-hop QA tasks generated via entity-level random walks.
• Framework supports self-evolution behaviors like progress verification and failure recovery.
• Bypasses weak supervision from sparse rewards by using iterative trajectory generation.
• Addresses long-horizon interaction challenges in agentic AI training pipelines.

This paper argues that aggregate safety metrics for planner-executor LLM pipelines conflate harmful intent reframing, planner refusal, and executor delegation. The authors introduce a five-condition controlled contrast design to isolate these mechanisms. Evaluated on synthetic and benchmark scenarios, the study reveals that overall pipeline safety is not a stable architectural property.

• Aggregate 'pipeline effect' metrics mask distinct failure modes in multi-agent systems.
• New five-condition design separates reframing, refusal, and delegation factors.
• Results show pipeline safety is unstable, not a fixed architectural trait.
• LLM-judged compliance used for evaluation across synthetic and benchmark sets.

This research demonstrates that biased LLM judges can effectively turn off the skill retirement mechanism in self-evolving agents. While retirement relies on accurate failure signals to prune poor skills, reference-free tasks force the use of judges that introduce systematic bias rather than random noise. The study proves this bias prevents the agent from correctly identifying and discarding bad skills, leading to uncontrolled drift in the skill library.

• Biased judges do not just add noise; they structurally disable the skill retirement mechanism.
• Reference-free task evaluation often forces the use of biased LLM judges.
• Skill drift occurs because agents fail to recognize and prune failing skills.
• Symmetric noise preserves retirement, but systematic bias breaks it.

AWS details a pipeline that captures PostgreSQL audit logs from CloudWatch Logs and converts them into structured CSV format for storage in Amazon S3. The solution processes log entries in near real time after generation to facilitate long-term analysis. This approach removes manual effort from the log extraction and formatting workflow.

• Extracts PostgreSQL audit logs directly from CloudWatch Logs for automated processing.
• Converts unstructured log entries into structured CSV format for easier analysis.
• Stores processed logs in Amazon S3 for scalable, long-term retention.
• Operates in near real-time, minimizing the window for audit data gaps.

CloudNativePG 1.30.0 introduces a DatabaseRole CRD to enable declarative, GitOps-friendly management of PostgreSQL roles within Kubernetes. The release also adds a lease-based primary election primitive to improve the safety of failover operations. Additionally, this update marks the End-of-Life for the 1.28.x series, urging operators to migrate to 1.29 or 1.30.

• Use the new DatabaseRole CRD to manage PostgreSQL users declaratively via GitOps workflows.
• Benefit from lease-based primary election logic to reduce split-brain risks during failover.
• Plan immediate upgrades from 1.28.x, which reached EOL on June 30, 2026.
• Maintenance releases 1.29.2 and 1.28.4 are available for current users.
• CloudNativePG remains the leading operator for PostgreSQL on Kubernetes.

Airbnb engineers released details on Sitar-agent, a Kubernetes sidecar designed to handle dynamic configuration delivery across tens of thousands of pods. The system processes updates multiple times per minute, leveraging Java and Amazon S3 for snapshot bootstrapping. A key architectural shift involved migrating from Sparkey to SQLite to enhance startup performance and reliability.

• Sitar-agent serves as a Kubernetes sidecar for real-time config delivery across massive pod fleets.
• Updates are processed several times per minute to maintain consistent service state.
• Migration from Sparkey to SQLite improves startup speed and configuration availability.
• Amazon S3 snapshot bootstrapping ensures reliable initial state loading on pod start.
• Java implementation supports the high-throughput requirements of dynamic config systems.

Amazon S3 Vectors is now live in AWS GovCloud (US-East) and (US-West), extending its vector storage capabilities to government and compliance-focused environments. The service provides purpose-built storage for AI agents, RAG, and semantic search, supporting billion-vector scales while maintaining S3's durability and elasticity. Users can store, access, and query vectors using dedicated APIs without provisioning underlying infrastructure.

• S3 Vectors is available in both US-East and US-West GovCloud regions.
• Supports AI agents, RAG, and semantic search at billion-vector scale.
• Uses dedicated APIs for storage and querying without infrastructure provisioning.
• Retains S3's core durability, availability, and elasticity guarantees.

Amazon Timestream for InfluxDB now publishes lifecycle event notifications to Amazon EventBridge whenever database instances or clusters change state. These events cover creation, deletion, scaling, parameter updates, maintenance windows, and reboots, including both successes and failures. This capability allows automation systems to react programmatically to operations without relying on API polling.

• Eliminates the need for polling APIs to track database lifecycle status changes.
• Enables automated workflows triggered by scaling, maintenance, or reboot completions.
• Supports immediate alerting for failed operations via EventBridge rules.
• Facilitates audit trails by routing events to CloudWatch Logs or S3.

Amazon EC2 U7i instances with 12TiB of DDR5 memory are now available in the AWS Europe (Zurich) region. Powered by fourth-generation Intel Xeon Scalable processors, these 896 vCPU nodes deliver up to 100Gbps network and EBS bandwidth. They are designed to support mission-critical in-memory databases such as SAP HANA, Oracle, and SQL Server.

• U7i-12tb.224xlarge instances provide 12TiB DDR5 memory for massive in-memory workloads.
• Zurich region now hosts these high-memory nodes, expanding European availability.
• Supports up to 100Gbps network and EBS for faster data loading and backups.
• Targeted at SAP HANA, Oracle, and SQL Server in-memory database use cases.

Amazon MSK Replicator extends its capabilities to allow data replication from external Apache Kafka clusters—including on-premises, self-managed, or third-party cloud environments—directly into Amazon MSK Standard brokers. This update complements existing support for MSK Express brokers, enabling broader migration paths, disaster recovery setups, and hybrid multi-cloud data distribution strategies. The feature automates replication tasks, removing the need for custom infrastructure or manual configuration of open-source tools.

• Replicate from external Kafka sources to MSK Standard brokers, not just Express.
• Simplifies migration of workloads to MSK Standard tier without custom tooling.
• Enables MSK Standard clusters to serve as failover or backup targets.
• Supports hybrid and multi-cloud data distribution via automated replication.
• Eliminates need to manage custom replication infrastructure or open-source configs.

The pglayers project publishes 53 PostgreSQL extensions as minimal Docker images built from scratch. These layers contain only shared libraries, control files, SQL scripts, and correct filesystem paths, allowing them to be stacked onto the official postgres image via COPY --from. This approach eliminates the need for compilation, apt-get, or build tools in the final container, enabling a straightforward CREATE EXTENSION workflow.

• Use COPY --from to stack extension layers onto official postgres images without bloat.
• Avoids compilation and build tools, reducing attack surface and image size.
• Supports 53 extensions including pgvector, PostGIS, and pg_cron.
• Maintains standard CREATE EXTENSION syntax for immediate usability.