# Persistence

In computer science, persistence is the characteristic of system state outliving the process that created it, typically by storing the state in durable data storage.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## How we scaled metrics aggregation by moving durable state into storage

DevFeed: [How we scaled metrics aggregation by moving durable state into storage](<https://devfeed.tech/articles/how-we-scaled-metrics-aggregation-by-moving-durable-state-into-storage-138310.md>)

Original publisher: [Read original article](<https://www.datadoghq.com/blog/engineering/how-we-scaled-metrics-aggregation/>)

Author: Khayyam Guliyev

Published: 2026-11-17T00:00:00Z

Content type: article

Language: en

Sources: [Datadog](<https://devfeed.tech/sources/datadog.md>), [Datadog](<https://devfeed.tech/sources/datadog-the-monitor-blog.md>)

Topics: [Logging and metrics](<https://devfeed.tech/topics/logging-and-metrics.md>), [Stateful](<https://devfeed.tech/topics/stateful.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [sharding](<https://devfeed.tech/topics/sharding.md>), [telemetry](<https://devfeed.tech/topics/telemetry.md>), [observability](<https://devfeed.tech/topics/observability.md>)

Tags: [1-5](<https://devfeed.tech/tags/1-5.md>), [account](<https://devfeed.tech/tags/account.md>), [added](<https://devfeed.tech/tags/added.md>), [after](<https://devfeed.tech/tags/after.md>), [aggregate](<https://devfeed.tech/tags/aggregate.md>), [aggregates](<https://devfeed.tech/tags/aggregates.md>), [aggregation](<https://devfeed.tech/tags/aggregation.md>), [all](<https://devfeed.tech/tags/all.md>), [allowed](<https://devfeed.tech/tags/allowed.md>), [allows](<https://devfeed.tech/tags/allows.md>), [also](<https://devfeed.tech/tags/also.md>), [datadog](<https://devfeed.tech/tags/datadog.md>), [fixed](<https://devfeed.tech/tags/fixed.md>), [metrics](<https://devfeed.tech/tags/metrics.md>), [moving](<https://devfeed.tech/tags/moving.md>), [ownership](<https://devfeed.tech/tags/ownership.md>), [partition](<https://devfeed.tech/tags/partition.md>), [reconciliation](<https://devfeed.tech/tags/reconciliation.md>), [state](<https://devfeed.tech/tags/state.md>), [storage](<https://devfeed.tech/tags/storage.md>), [tags](<https://devfeed.tech/tags/tags.md>)

### AI overview

Datadog's metrics aggregation scaled through three designs as durable state became a bottleneck. The first used fixed Kafka partition ownership, with leader and follower instances holding hours of RocksDB state; it provided failover but duplicated compute and required manual, risky repartitioning. The second dynamically reassigned partitions and uploaded incremental state to blob storage, restoring recent windows first and older state when late data required it. This enabled autoscaling and about 10x more volume, but recovery and state movement eventually constrained scale. The third makes streaming aggregation and deduplication best-effort: workers prioritize on-time traffic and can reset to the live edge, while storage reconciles partial aggregates, duplicates, late data, and backfills across its recent and long-term tiers. This removes hours of durable local state and snapshot recovery from the streaming path, while placing durable correctness in storage.

### Source excerpt

Learn how Datadog evolved metrics aggregation from fixed partition ownership to transferable state and, ultimately, storage-side reconciliation.

## SymfonyCon Warsaw 2026: Giving AI Agents a Durable Memory

DevFeed: [SymfonyCon Warsaw 2026: Giving AI Agents a Durable Memory](<https://devfeed.tech/articles/symfonycon-warsaw-2026-giving-ai-agents-a-durable-memory-136423.md>)

Original publisher: [Read original article](<https://symfony.com/blog/symfonycon-warsaw-2026-giving-ai-agents-a-durable-memory>)

Author: Mélyne FREUZE

Published: 2026-11-16T13:30:00Z

Content type: news

Language: en

Sources: [Symfony](<https://devfeed.tech/sources/symfony-blog.md>)

Topics: [Developer Events](<https://devfeed.tech/topics/developer-events.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [versioning](<https://devfeed.tech/topics/versioning.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [3](<https://devfeed.tech/tags/3.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [announcement](<https://devfeed.tech/tags/announcement.md>), [community](<https://devfeed.tech/tags/community.md>), [community-events](<https://devfeed.tech/tags/community-events.md>), [developer](<https://devfeed.tech/tags/developer.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [events](<https://devfeed.tech/tags/events.md>), [featured](<https://devfeed.tech/tags/featured.md>), [giving](<https://devfeed.tech/tags/giving.md>), [head-of-engineering](<https://devfeed.tech/tags/head-of-engineering.md>), [it-s](<https://devfeed.tech/tags/it-s.md>), [memory](<https://devfeed.tech/tags/memory.md>), [speaker](<https://devfeed.tech/tags/speaker.md>), [takes](<https://devfeed.tech/tags/takes.md>), [talks](<https://devfeed.tech/tags/talks.md>), [warsaw](<https://devfeed.tech/tags/warsaw.md>)

### AI overview

Clément Bertillon's announced SymfonyCon talk proposes preserving AI agent context across sessions and tools with user and project instructions that point to a small, indexed, versioned Markdown knowledge base. The session will also consider this approach's limits, compare it with native memory, and discuss when heavier retrieval systems may be a better fit. The source describes the planned talk, not its results.

### Source excerpt

It's almost time for our annual reunion! SymfonyCon Warsaw 2026 takes place Nov 26-27, featuring 3 tracks of developer talks and community events. 🎤 Speaker announcement Joining us as a featured speaker is Clément Bertillon, Head of Engineering,...

## Redis vs Memcached

DevFeed: [Redis vs Memcached](<https://devfeed.tech/articles/redis-vs-memcached-135573.md>)

Original publisher: [Read original article](<https://blog.levelupcoding.com/p/redis-vs-memcached>)

Author: Nikki Siapno

Published: 2026-11-16T11:51:16Z

Content type: article

Language: en

Sources: [Level Up Coding System Design Newsletter](<https://devfeed.tech/sources/level-up-coding-system-design-newsletter.md>)

Topics: [Memcached](<https://devfeed.tech/topics/memcached.md>), [Redis](<https://devfeed.tech/topics/redis.md>), [Cache](<https://devfeed.tech/topics/cache.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [consistency](<https://devfeed.tech/topics/consistency.md>)

Tags: [api](<https://devfeed.tech/tags/api.md>), [atomic](<https://devfeed.tech/tags/atomic.md>), [cache](<https://devfeed.tech/tags/cache.md>), [cas](<https://devfeed.tech/tags/cas.md>), [data](<https://devfeed.tech/tags/data.md>), [data-structures](<https://devfeed.tech/tags/data-structures.md>), [database](<https://devfeed.tech/tags/database.md>), [disposable](<https://devfeed.tech/tags/disposable.md>), [failover](<https://devfeed.tech/tags/failover.md>), [failure](<https://devfeed.tech/tags/failure.md>), [failures](<https://devfeed.tech/tags/failures.md>), [features](<https://devfeed.tech/tags/features.md>), [fields](<https://devfeed.tech/tags/fields.md>), [file](<https://devfeed.tech/tags/file.md>), [memcached](<https://devfeed.tech/tags/memcached.md>), [memory](<https://devfeed.tech/tags/memory.md>), [redis](<https://devfeed.tech/tags/redis.md>), [speed](<https://devfeed.tech/tags/speed.md>), [vs](<https://devfeed.tech/tags/vs.md>)

### AI overview

Both Redis and Memcached cache data in memory, but Redis supports richer data structures, atomic in-place updates, configurable eviction, persistence, replication, and failover. Memcached stores raw key-value values, automatically evicts items as memory fills, and expects lost cache data to be rebuilt; changing part of a value requires the application to rewrite it, though counters and CAS are available. Redis defaults to rejecting memory-requiring writes at its limit unless an eviction policy is configured, and its asynchronous replication can still lose acknowledged writes during failure. The article suggests Memcached for regenerable query results or API responses, Redis for structured data and costly-to-rebuild caches, and a durable database for critical records. Redis's added recovery and data features bring configuration and operational responsibilities.

### Source excerpt

Both cache data in memory. But the differences go far beyond speed.

## Durable Workflow: Long-Running Workflows for Laravel

DevFeed: [Durable Workflow: Long-Running Workflows for Laravel](<https://devfeed.tech/articles/durable-workflow-long-running-workflows-for-laravel-139541.md>)

Original publisher: [Read original article](<https://laravel-news.com/durable-workflow-laravel>)

Author: Yannick Lyn Fatt

Published: 2026-11-16T03:04:17Z

Content type: article

Language: en

Sources: [Laravel](<https://devfeed.tech/sources/laravel.md>)

Topics: [Laravel](<https://devfeed.tech/topics/laravel.md>), [workflow automation](<https://devfeed.tech/topics/workflow-automation.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>)

Tags: [amazon-sqs](<https://devfeed.tech/tags/amazon-sqs.md>), [background](<https://devfeed.tech/tags/background.md>), [background-work](<https://devfeed.tech/tags/background-work.md>), [cache](<https://devfeed.tech/tags/cache.md>), [database](<https://devfeed.tech/tags/database.md>), [laravel](<https://devfeed.tech/tags/laravel.md>), [laravel-packages](<https://devfeed.tech/tags/laravel-packages.md>), [long-running](<https://devfeed.tech/tags/long-running.md>), [packages](<https://devfeed.tech/tags/packages.md>), [progress](<https://devfeed.tech/tags/progress.md>), [queues](<https://devfeed.tech/tags/queues.md>), [workflow](<https://devfeed.tech/tags/workflow.md>), [workflows](<https://devfeed.tech/tags/workflows.md>)

### AI overview

Durable Workflow is an open-source Laravel package for processes that need progress preserved across worker restarts. It records activity results in your database, runs work through Laravel queues, and uses cache locks; it requires an asynchronous queue driver and a lock-capable cache. Workflows coordinate PHP activities, and replay reuses recorded results, though an operation may run again if a worker stops before recording its result, so activities should be safe to repeat. In v2, retries are opt-in and activities default to one attempt; an unhandled final failure does not automatically reopen after a fix. Signals and durable timers support waiting for external input. The article recommends ordinary Laravel jobs for single background tasks and positions workflows for multi-step processes or those that wait. Version 2.28.1 requires Laravel 12.72+ or 13.41+ and PHP 8.2+ (PHP 8.3+ for Laravel 13).

### Source excerpt

Durable Workflow runs long-running Laravel workflows on your own database, queues, and cache. Learn how it records progress and coordinates background work. The post Durable Workflow: Long-Running Workflows for Laravel appeared first on Laravel News. Join the Laravel Newsletter to get Laravel articles like this directly in your inbox.

## Buckets guide to what's in store with AWS Storage at re:Invent 2026

DevFeed: [Buckets guide to what's in store with AWS Storage at re:Invent 2026](<https://devfeed.tech/articles/buckets-guide-to-what-s-in-store-with-aws-storage-at-re-invent-2026-131282.md>)

Original publisher: [Read original article](<https://aws.amazon.com/blogs/storage/buckets-guide-to-whats-in-store-with-aws-storage-at-reinvent-2026/>)

Author: Sean White

Published: 2026-11-13T00:36:07Z

Content type: article

Language: en

Sources: [AWS Storage Blog](<https://devfeed.tech/sources/aws-storage-blog.md>)

Topics: [Amazon S3](<https://devfeed.tech/topics/amazon-s3.md>), [Amazon Elastic Block Store (Amazon EBS)](<https://devfeed.tech/topics/amazon-elastic-block-store-amazon-ebs.md>), [Learning](<https://devfeed.tech/topics/learning.md>), [data lake](<https://devfeed.tech/topics/data-lake.md>), [data-engineering](<https://devfeed.tech/topics/data-engineering.md>), [Network Attached Storage (NAS)](<https://devfeed.tech/topics/network-attached-storage-nas.md>), [big-data](<https://devfeed.tech/topics/big-data.md>), [Data pipelines](<https://devfeed.tech/topics/data-pipelines.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [Amazon SageMaker](<https://devfeed.tech/topics/amazon-sagemaker.md>), [Amazon GuardDuty](<https://devfeed.tech/topics/amazon-guardduty.md>), [sharding](<https://devfeed.tech/topics/sharding.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-ml](<https://devfeed.tech/tags/ai-ml.md>), [amazon-bedrock](<https://devfeed.tech/tags/amazon-bedrock.md>), [amazon-elastic-block-store-amazon-ebs](<https://devfeed.tech/tags/amazon-elastic-block-store-amazon-ebs.md>), [amazon-elastic-file-system-efs](<https://devfeed.tech/tags/amazon-elastic-file-system-efs.md>), [amazon-fsx](<https://devfeed.tech/tags/amazon-fsx.md>), [amazon-fsx-for-lustre](<https://devfeed.tech/tags/amazon-fsx-for-lustre.md>), [amazon-fsx-for-netapp-ontap](<https://devfeed.tech/tags/amazon-fsx-for-netapp-ontap.md>), [amazon-fsx-for-openzfs](<https://devfeed.tech/tags/amazon-fsx-for-openzfs.md>), [amazon-simple-storage-service-s3](<https://devfeed.tech/tags/amazon-simple-storage-service-s3.md>), [announcements](<https://devfeed.tech/tags/announcements.md>), [aws](<https://devfeed.tech/tags/aws.md>), [aws-backup](<https://devfeed.tech/tags/aws-backup.md>), [aws-cloud-storage](<https://devfeed.tech/tags/aws-cloud-storage.md>), [aws-re-invent](<https://devfeed.tech/tags/aws-re-invent.md>), [aws-storage](<https://devfeed.tech/tags/aws-storage.md>), [re-invent](<https://devfeed.tech/tags/re-invent.md>), [storage](<https://devfeed.tech/tags/storage.md>), [uncategorized](<https://devfeed.tech/tags/uncategorized.md>)

### AI overview

The AWS Storage track at re:Invent 2026 offers 79 sessions across 107 time slots, with roughly three quarters of slots interactive and 73 sessions at level 300 or higher. Topics span AI agent storage, performance, data lakes, migration, security, and recovery. The guide contrasts services for particular needs: S3 Express One Zone for latency-sensitive object access, FSx for Lustre for high throughput, EBS for low-latency block storage, and EFS/FSx for file workloads. Sessions include hands-on workshops, chalk talks, and builders' sessions; these are not recorded, while breakouts are posted to AWS Events YouTube after the event. Reserved seats are open to registered attendees, though walk-up seating may be available; many sessions run twice. The event is in Las Vegas November 30-December 4, 2026, and attendees should allow at least 30 minutes between venues.

### Source excerpt

Wondering what is in store at re:Invent this year? A lot. I would know. Storing things is my whole job. Every great story starts with a single object. This one starts with me. Hi, I am Buckets. The official AWS Storage mascot, self-appointed guardian of all things durable, scalable, and correctly permissioned. At re:Invent, I [...]

## Long-horizon agents and durable execution

DevFeed: [Long-horizon agents and durable execution](<https://devfeed.tech/articles/long-horizon-agents-and-durable-execution-128970.md>)

Original publisher: [Read original article](<https://martinfowler.com/articles/durable-execution-agents.html>)

Author: Martin Fowler (martin@martinfowler.com)

Published: 2026-11-10T14:10:00Z

Content type: article

Language: en

Sources: [Martin Fowler](<https://devfeed.tech/sources/martin-fowler.md>)

Topics: [long-running](<https://devfeed.tech/topics/long-running.md>), [Multi-step AI processes](<https://devfeed.tech/topics/multi-step-ai-processes.md>), [agentic workflows](<https://devfeed.tech/topics/agentic-workflows.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [automate](<https://devfeed.tech/tags/automate.md>), [behavior](<https://devfeed.tech/tags/behavior.md>), [can](<https://devfeed.tech/tags/can.md>), [complete](<https://devfeed.tech/tags/complete.md>), [complex](<https://devfeed.tech/tags/complex.md>), [crashes](<https://devfeed.tech/tags/crashes.md>), [days](<https://devfeed.tech/tags/days.md>), [durable-execution](<https://devfeed.tech/tags/durable-execution.md>), [execution](<https://devfeed.tech/tags/execution.md>), [external](<https://devfeed.tech/tags/external.md>), [failure](<https://devfeed.tech/tags/failure.md>), [failures](<https://devfeed.tech/tags/failures.md>), [helps](<https://devfeed.tech/tags/helps.md>), [horizon](<https://devfeed.tech/tags/horizon.md>), [model](<https://devfeed.tech/tags/model.md>), [more](<https://devfeed.tech/tags/more.md>), [multiple](<https://devfeed.tech/tags/multiple.md>), [network](<https://devfeed.tech/tags/network.md>), [pattern](<https://devfeed.tech/tags/pattern.md>)

### AI overview

Durable execution lets long-running agents recover from crashes by logging each external step's inputs and results, replaying the program to rebuild state, and skipping completed steps. Temporal and Cloudflare Workflows both support durable steps and waits; the article's code uses Temporal, while Cloudflare uses `step.do` and `step.waitForEvent`. Retries are at least once, so writes need stable idempotency keys to avoid duplicates. The harness must also handle uncertain outcomes, stale approvals and changing prices, code changes during suspended workflows, growing memory, and untrusted text. The article contrasts runtime behavior during rule changes: Temporal continued an existing task under its earlier rule, while Cloudflare failed when replayed decisions violated current rules. It recommends versioning rules and treating the log as a published interface. Durable execution supports recovery, but does not ensure decisions are correct or secure.

### Source excerpt

Long-horizon agents can automate complex tasks that take days or weeks to complete and involve multiple exchanges with people and external systems. Many things can go wrong during this time, including process crashes, network failures, and unanticipated real-world behavior. Kiran Prakash explains durable execution: an emerging pattern that helps agents recover from these failures and move towards their goal. In this programming model, the runtime keeps a record of the result of each step that interacts with the outside world so that, when a failure occurs, the program runs again, completed steps are skipped, and the program's state is rebuilt from the record. more...

## Running Apache Polaris with MongoDB: NoSQL Persistence, Bootstrap, and Maintenance

DevFeed: [Running Apache Polaris with MongoDB: NoSQL Persistence, Bootstrap, and Maintenance](<https://devfeed.tech/articles/running-apache-polaris-with-mongodb-nosql-persistence-bootstrap-and-maintenance-123569.md>)

Original publisher: [Read original article](<https://www.dremio.com/blog/running-apache-polaris-with-mongodb-nosql-persistence-bootstrap-and-maintenance/>)

Author: Alex Merced

Published: 2026-11-09T13:42:07Z

Content type: tutorial

Language: en

Sources: [Dremio](<https://devfeed.tech/sources/dremio.md>)

Topics: [MongoDB](<https://devfeed.tech/topics/mongodb.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [test](<https://devfeed.tech/topics/test.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>)

Tags: [apache-iceberg](<https://devfeed.tech/tags/apache-iceberg.md>), [apache-polaris](<https://devfeed.tech/tags/apache-polaris.md>), [bootstrap](<https://devfeed.tech/tags/bootstrap.md>), [data-lakehouse](<https://devfeed.tech/tags/data-lakehouse.md>), [dremio-blog-open-data-insights](<https://devfeed.tech/tags/dremio-blog-open-data-insights.md>), [maintenance](<https://devfeed.tech/tags/maintenance.md>), [mongodb](<https://devfeed.tech/tags/mongodb.md>), [nosql](<https://devfeed.tech/tags/nosql.md>), [persistence](<https://devfeed.tech/tags/persistence.md>)

### AI overview

Apache Polaris's MongoDB metastore is documented as experimental, while PostgreSQL is recommended for production. The text describes MongoDB as a possible path for evaluation if teams verify release-specific support, configure a replica set, and bootstrap with an admin-tool matching the server version before starting Polaris. It calls for automated daily health checks and selective maintenance, plus tested MongoDB-level backups and end-to-end restores. Treat persistence migration to PostgreSQL as a separate planned operation, potentially requiring downtime; do not assume it can happen during a rolling server upgrade. The operational guidance and version details are tied to Polaris 1.8.0 and documentation checked November 9, 2026, so verify them for the release in use.

### Source excerpt

Configure Apache Polaris with MongoDB persistence, bootstrap the metastore, run required NoSQL maintenance, and evaluate the current production tradeoffs. The post Running Apache Polaris with MongoDB: NoSQL Persistence, Bootstrap, and Maintenance appeared first on Dremio.

## Inside GitHub's new Git storage architecture

DevFeed: [Inside GitHub's new Git storage architecture](<https://devfeed.tech/articles/inside-github-s-new-git-storage-architecture-119704.md>)

Original publisher: [Read original article](<https://github.blog/engineering/architecture-optimization/inside-githubs-new-git-storage-architecture/>)

Author: Karthik Raman

Published: 2026-11-05T20:23:21Z

Content type: article

Language: en

Sources: [GitHub Blog](<https://devfeed.tech/sources/github-engineering.md>), [GitHub Engineering](<https://devfeed.tech/sources/github.md>)

Topics: [Architecture & optimization](<https://devfeed.tech/topics/architecture-optimization.md>), [scalable-architecture](<https://devfeed.tech/topics/scalable-architecture.md>), [GitHub](<https://devfeed.tech/topics/github.md>), [decoupling](<https://devfeed.tech/topics/decoupling.md>), [Git](<https://devfeed.tech/topics/git.md>), [Cloud Architecture](<https://devfeed.tech/topics/cloud-architecture.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [Multi-tenancy](<https://devfeed.tech/topics/multi-tenancy.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-tasks](<https://devfeed.tech/tags/agent-tasks.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-development](<https://devfeed.tech/tags/agentic-development.md>), [agents](<https://devfeed.tech/tags/agents.md>), [api](<https://devfeed.tech/tags/api.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [architecture-optimization](<https://devfeed.tech/tags/architecture-optimization.md>), [atomic](<https://devfeed.tech/tags/atomic.md>), [availability](<https://devfeed.tech/tags/availability.md>), [azure](<https://devfeed.tech/tags/azure.md>), [balancing](<https://devfeed.tech/tags/balancing.md>), [blob](<https://devfeed.tech/tags/blob.md>), [blob-storage](<https://devfeed.tech/tags/blob-storage.md>), [branch](<https://devfeed.tech/tags/branch.md>), [cache](<https://devfeed.tech/tags/cache.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [git](<https://devfeed.tech/tags/git.md>), [github](<https://devfeed.tech/tags/github.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [repository](<https://devfeed.tech/tags/repository.md>), [storage](<https://devfeed.tech/tags/storage.md>), [storage-architecture](<https://devfeed.tech/tags/storage-architecture.md>), [technical-deep-dive](<https://devfeed.tech/tags/technical-deep-dive.md>)

### AI overview

GitHub says its new "Rubicon" architecture demonstrated 7,200 pushes per second from 5,000 concurrent agents on one repository during internal testing, while maintaining low read latency. It separates durable Git objects in Azure Blob Storage from reference state in Ref-CAS, a replicated log that orders reference updates with quorum commits; virtual replicas handle reads and writes, and separate compaction nodes merge packs in the background. This design avoids ordering immutable object uploads and lets reads, writes, and maintenance scale independently. Git clients keep the same push, fetch, and branch behavior and protections. GitHub says the architecture is in limited private preview, with broader rollout planned for early next year. The performance figures are internal test results, and the article describes several further scaling improvements as planned.

### Source excerpt

GitHub's new Git storage architecture is built so thousands of agents can push to a single repository at once. Here's how it works. The post Inside GitHub's new Git storage architecture appeared first on The GitHub Blog.

## Why AI Agents Need State Machines to Coordinate Humans

DevFeed: [Why AI Agents Need State Machines to Coordinate Humans](<https://devfeed.tech/articles/why-ai-agents-need-state-machines-to-coordinate-humans-116248.md>)

Original publisher: [Read original article](<https://hackernoon.com/why-ai-agents-need-state-machines-to-coordinate-humans?source=rss>)

Author: Akshay Panchavati

Published: 2026-11-04T07:42:10Z

Content type: article

Language: en

Sources: [HackerNoon](<https://devfeed.tech/sources/hackernoon.md>)

Topics: [Finite-state machine](<https://devfeed.tech/topics/finite-state-machine.md>), [human in the loop agent](<https://devfeed.tech/topics/human-in-the-loop-agent.md>), [state-management](<https://devfeed.tech/topics/state-management.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>)

Tags: [agentic-ai-architecture](<https://devfeed.tech/tags/agentic-ai-architecture.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent-architecture](<https://devfeed.tech/tags/ai-agent-architecture.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [autonomous-ai-agents](<https://devfeed.tech/tags/autonomous-ai-agents.md>), [dispatch](<https://devfeed.tech/tags/dispatch.md>), [finite-state-machine](<https://devfeed.tech/tags/finite-state-machine.md>), [human-in-the-loop](<https://devfeed.tech/tags/human-in-the-loop.md>), [llm](<https://devfeed.tech/tags/llm.md>), [proptech](<https://devfeed.tech/tags/proptech.md>), [state-machines](<https://devfeed.tech/tags/state-machines.md>)

### AI overview

For coordinating a tenant, plumber, and landlord who may reply late or change their minds, the article recommends a deterministic finite state machine (FSM) that stores job state in a database and controls transitions and actions. The LLM translates incoming messages into defined inputs, including `UNCLEAR`, and renders code-determined messages in natural language; it does not decide the next step or own the workflow state. This keeps approvals, scheduling, and follow-ups governed by readable rules even when human replies are delayed or ambiguous. The excerpt describes the design and a leaking-sink scenario, but does not specify the nine states or detail the surrounding structures for handling unreliable human inputs.

### Source excerpt

Most human-in-the-loop designs assume the human is an employee. Here is an FSM design where an LLM coordinates tenants, contractors, and landlords instead.

## The Agent Board: Kanban for Machines; Durable Coordination for Agents That Outlive the Conversation

DevFeed: [The Agent Board: Kanban for Machines; Durable Coordination for Agents That Outlive the Conversation](<https://devfeed.tech/articles/the-agent-board-kanban-for-machines-durable-coordination-for-agents-that-outlive-the-conversation-116481.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.04100>)

Author: Tyler Hunt, Donald Nguyen, Gurbinder Gill

Published: 2026-11-04T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Distributed Computing](<https://devfeed.tech/sources/arxiv-distributed-computing.md>)

Topics: [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>), [agent to agent communication](<https://devfeed.tech/topics/agent-to-agent-communication.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [Agent Harness](<https://devfeed.tech/topics/agent-harness.md>), [async](<https://devfeed.tech/topics/async.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [block](<https://devfeed.tech/tags/block.md>), [board](<https://devfeed.tech/tags/board.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-dc](<https://devfeed.tech/tags/cs-dc.md>), [events](<https://devfeed.tech/tags/events.md>), [fields](<https://devfeed.tech/tags/fields.md>), [handler](<https://devfeed.tech/tags/handler.md>), [harness](<https://devfeed.tech/tags/harness.md>), [kanban](<https://devfeed.tech/tags/kanban.md>), [kernels](<https://devfeed.tech/tags/kernels.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [large-language-models](<https://devfeed.tech/tags/large-language-models.md>), [mirrors](<https://devfeed.tech/tags/mirrors.md>), [policy](<https://devfeed.tech/tags/policy.md>), [post](<https://devfeed.tech/tags/post.md>), [roles](<https://devfeed.tech/tags/roles.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [single-threaded](<https://devfeed.tech/tags/single-threaded.md>), [substrate](<https://devfeed.tech/tags/substrate.md>), [team](<https://devfeed.tech/tags/team.md>)

### AI overview

The Agent Board is a durable shared task table for coordinating agents across turns, agents, and external systems. Agents can post or watch tasks and block without holding a process; the Board Keeper brings in external events, matches them to waiting tasks by interpreting their prose, and can allocate agents by role. Loadable policy kernels assign meaning to statuses and roles, but the board itself does not enforce those conventions. In three small simulated pilots, board wait cost stayed flat as wait length increased, while naive polling grew about 165-fold; a plain external watcher was cheaper. The Board Keeper produced no false wakes but missed 3.6% of real wakes.

### Source excerpt

arXiv:2611.04100v1 Announce Type: cross Abstract: Harnesses for large language models have converged on a single-threaded shape: one agent, one history, one machine, one sitting. That shape fails work that spans many agents, outlives any one model call, or waits on systems that are not agents. We describe the Agent Board, the coordination substrate in the Corvic harness. The board is a durable, shared table of semi-semantic tasks: a few exact fields the system routes on, and a free-form body that only language models read. Agents post tasks, watch tasks, and block on them, ending their turn and holding no process until a matching post wakes them. A system agent, the Board Keeper, acts as the interrupt handler for the board: it carries events from non-agent parts of the harness onto the board, matches them to waiting tasks by reading their prose, and allocates new agents for tasks that name a kind of agent rather than an instance. Loadable coordination policy kernels give statuses and roles meaning for a given pattern of work, without the board enforcing them. The design mirrors the kanban board a human team puts on a wall. We argue it is what persistent agent products such as Meta Muse, OpenAI dots and SpaceXAI Grok Bot need once an agent must wait, hand off work, or work with agents from another vendor. We view the board as an operating-system scheduler for agents and state where the analogy breaks. We separate what the mechanism guarantees from what it leaves to convention, and report a trace from a live board and three small simulated pilots. In the pilots, the wait cost of the board was flat in wait length while naive polling grew about 165-fold, but a plain external watcher was cheaper than the board. The Board Keeper matched prose waiters with no false wakes but missed 3.6 percent of real wakes.

## Durable Execution means deciding what survives failure

DevFeed: [Durable Execution means deciding what survives failure](<https://devfeed.tech/articles/durable-execution-means-deciding-what-survives-failure-117163.md>)

Original publisher: [Read original article](<https://temporal.io/blog/durable-execution-means-deciding-what-survives-failure>)

Author: Muhammad Abu Bakkar Bin Akmal

Published: 2026-11-04T00:00:00Z

Content type: article

Language: en

Sources: [Temporal](<https://devfeed.tech/sources/temporal-blog.md>)

Topics: [Persistence](<https://devfeed.tech/topics/persistence.md>), [long-running](<https://devfeed.tech/topics/long-running.md>), [reliability-engineering](<https://devfeed.tech/topics/reliability-engineering.md>), [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>), [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>), [ReAct agent Python](<https://devfeed.tech/topics/react-agent-python.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-workflows](<https://devfeed.tech/tags/agent-workflows.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [community](<https://devfeed.tech/tags/community.md>), [durable-execution](<https://devfeed.tech/tags/durable-execution.md>), [failure](<https://devfeed.tech/tags/failure.md>), [kubernetes](<https://devfeed.tech/tags/kubernetes.md>), [progress](<https://devfeed.tech/tags/progress.md>), [recovery](<https://devfeed.tech/tags/recovery.md>), [retries](<https://devfeed.tech/tags/retries.md>), [temporal](<https://devfeed.tech/tags/temporal.md>), [workflows](<https://devfeed.tech/tags/workflows.md>)

### AI overview

Temporal can restore Workflow progress after Worker restarts by replaying Event History, but application policy still determines retries, business status, cancellation, and what counts as success. In the Kubernetes scan, six analyzers run concurrently but are awaited in slice order, so failures are not observed fail-fast; the review also flags status cleanup and ambiguous database-record/Workflow-start outcomes. Analysis results are persisted before the optional assessment, allowing them to survive an assessment failure. In the AI recommendation loop, proposals carry prior feedback and budget across rounds, with explicit exits for consensus, approved score, repeated concerns, or round limit. Only accepted decisions should be cached; unresolved results can be structurally valid. The excerpt recommends defining start and ID reuse semantics, making external effects retry-safe through stable keys and deduplication, putting replay-sensitive configuration in Workflow input, and testing business outcomes and Worker-loss recovery. Its reviewed tests did not cover the listed failure scenarios or a live Nexus end-to-end path.

### Source excerpt

See how Temporal preserves progress across failures in Kubernetes analysis and AI agent Workflows, while keeping retries and recovery explicit.

## Scaling AI Data Storage and Delivery with NVIDIA AIStore and B2

DevFeed: [Scaling AI Data Storage and Delivery with NVIDIA AIStore and B2](<https://devfeed.tech/articles/scaling-ai-data-storage-and-delivery-with-nvidia-aistore-and-b2-116449.md>)

Original publisher: [Read original article](<https://www.backblaze.com/blog/scaling-ai-data-storage-and-delivery-with-nvidia-aistore-and-b2/>)

Author: Jeronimo De Leon

Published: 2026-11-03T17:29:46Z

Content type: article

Language: en

Sources: [Backblaze](<https://devfeed.tech/sources/backblaze-blog-cloud-storage-cloud-backup.md>)

Topics: [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [Cache](<https://devfeed.tech/topics/cache.md>), [Amazon S3](<https://devfeed.tech/topics/amazon-s3.md>), [big-data](<https://devfeed.tech/topics/big-data.md>), [data lake](<https://devfeed.tech/topics/data-lake.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [sharding](<https://devfeed.tech/topics/sharding.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-data](<https://devfeed.tech/tags/ai-data.md>), [ai-ml](<https://devfeed.tech/tags/ai-ml.md>), [ai-workloads](<https://devfeed.tech/tags/ai-workloads.md>), [b2cloud](<https://devfeed.tech/tags/b2cloud.md>), [backblaze](<https://devfeed.tech/tags/backblaze.md>), [backblaze-b2](<https://devfeed.tech/tags/backblaze-b2.md>), [capacity](<https://devfeed.tech/tags/capacity.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [cloud-storage](<https://devfeed.tech/tags/cloud-storage.md>), [compute](<https://devfeed.tech/tags/compute.md>), [curation](<https://devfeed.tech/tags/curation.md>), [data](<https://devfeed.tech/tags/data.md>), [data-storage](<https://devfeed.tech/tags/data-storage.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [delivery](<https://devfeed.tech/tags/delivery.md>), [fast](<https://devfeed.tech/tags/fast.md>), [featured](<https://devfeed.tech/tags/featured.md>), [featured-cloud-storage](<https://devfeed.tech/tags/featured-cloud-storage.md>), [local](<https://devfeed.tech/tags/local.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [scaling](<https://devfeed.tech/tags/scaling.md>), [scaling-ai](<https://devfeed.tech/tags/scaling-ai.md>), [storage](<https://devfeed.tech/tags/storage.md>)

### AI overview

Backblaze B2 holds the durable dataset record while NVIDIA AIStore caches and stages active data on storage near compute, balancing retention cost with faster access. For S3-based processing such as NeMo Curator, applications can use AIStore through configuration, prefetch data, and reuse cached reads. For Isaac GR00T training, AIStore stages a dataset revision to GPU-node disk and checks it against a manifest, leaving the trainer's data loader unchanged; checkpoints publish back through AIStore to B2. The post recommends sizing AIStore for concurrent working sets and reuse, using prefetch when predictable startup matters, keeping dataset revisions in write-once prefixes, publishing checksummed manifests, and measuring loading, warm reads, processing, and publication separately. The approach suits repeated or shared reads; a single sequential pass may be simpler streamed directly from B2. AIStore is infrastructure teams operate, and write-through adds local-write overhead before upload.

### Source excerpt

AI workloads need durable data retention and fast access to active datasets. See how Backblaze B2 and NVIDIA AIStore pair a persistent capacity tier with local storage near compute for video curation, robot training, and repeated reads across shared datasets. The post Scaling AI Data Storage and Delivery with NVIDIA AIStore and B2 appeared first on Backblaze Blog | Cloud Storage & Cloud Backup

## Eliminating Servers in State Machine Replication

DevFeed: [Eliminating Servers in State Machine Replication](<https://devfeed.tech/articles/eliminating-servers-in-state-machine-replication-113645.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.01363>)

Author: Dohyun Park, Wenjie Hu, Xiangyao Yu, Mahesh Balakrishnan, Yongjoo Park

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Databases](<https://devfeed.tech/sources/arxiv-databases.md>), [arXiv: Distributed Computing](<https://devfeed.tech/sources/arxiv-distributed-computing.md>)

Topics: [Raft](<https://devfeed.tech/topics/raft.md>), [etcd](<https://devfeed.tech/topics/etcd.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [Stateful](<https://devfeed.tech/topics/stateful.md>)

Tags: [availability](<https://devfeed.tech/tags/availability.md>), [client](<https://devfeed.tech/tags/client.md>), [cs-db](<https://devfeed.tech/tags/cs-db.md>), [cs-dc](<https://devfeed.tech/tags/cs-dc.md>), [distributed](<https://devfeed.tech/tags/distributed.md>), [distributed-databases](<https://devfeed.tech/tags/distributed-databases.md>), [etcd](<https://devfeed.tech/tags/etcd.md>), [faster](<https://devfeed.tech/tags/faster.md>), [interface](<https://devfeed.tech/tags/interface.md>), [key-value-store](<https://devfeed.tech/tags/key-value-store.md>), [latency](<https://devfeed.tech/tags/latency.md>), [materialization](<https://devfeed.tech/tags/materialization.md>), [metadata](<https://devfeed.tech/tags/metadata.md>), [order](<https://devfeed.tech/tags/order.md>), [ordering](<https://devfeed.tech/tags/ordering.md>), [paper](<https://devfeed.tech/tags/paper.md>), [provisioning](<https://devfeed.tech/tags/provisioning.md>), [raft](<https://devfeed.tech/tags/raft.md>), [replication](<https://devfeed.tech/tags/replication.md>), [serverless](<https://devfeed.tech/tags/serverless.md>), [state-machine](<https://devfeed.tech/tags/state-machine.md>), [storage](<https://devfeed.tech/tags/storage.md>)

### AI overview

Serverless SMR (S²MR) stores both the durable command order and materialized state on disaggregated storage, with clients concurrently maintaining a single shared state copy. The authors argue the shared log can support distributed state materialization as well as command ordering, making it the interface for SMR without server-based state storage and snapshot serving. Their demonstration, AirDB, is a transactional key-value store. In the reported evaluation, AirDB over S²MR was more available than existing SMR systems such as etcd over Raft, remained unaffected by server failures and bad client activity, and started 10-45 times faster from a cold start.

### Source excerpt

arXiv:2611.01363v1 Announce Type: new Abstract: State Machine Replication (SMR) underpins many stateful systems, including distributed databases, coordination services, and metadata stores. Fundamentally, SMR requires state to be materialized from a durable total order of commands. Existing SMR systems rely on servers to materialize, locally store, and serve snapshots of state; while the durable total order either resides on the same servers, or on an external, disaggregated shared log. Unfortunately, servers act as brittle, expensive choke-points in such systems, causing service unavailability when they fail; incurring provisioning latency on cold starts; and red-lining when exposed to bad clients. In this paper, we propose Serverless SMR (S^2MR), where both the materialized state and the durable total order are stored on disaggregated storage; clients concurrently maintain a single copy of materialized state on external storage. Our key technical insight is that the shared log abstraction is useful not just for consistently ordering commands in SMR, but also for enabling distributed materialization, thus becoming the only interface for ac hieving SMR. We demonstrate S^2MR by developing AirDB, a transactional key-value store. In our evaluation, AirD B over S^2MR achieves higher availability than existing SMR-based systems like etcd over Raft: it is oblivious to server failures and bad client activity; and spins up 10x-45x faster from a cold start.

## Approve the Revolut counterparty before the first payment

DevFeed: [Approve the Revolut counterparty before the first payment](<https://devfeed.tech/articles/approve-the-revolut-counterparty-before-the-first-payment-109712.md>)

Original publisher: [Read original article](<https://www.swytchcode.com/content/revolut-business-api>)

Author: Swytchcode Team

Published: 2026-10-30T18:56:52Z

Content type: tutorial

Language: en

Sources: [Swytchcode](<https://devfeed.tech/sources/swytchcode.md>)

Topics: [human in the loop agent](<https://devfeed.tech/topics/human-in-the-loop-agent.md>), [API Economy](<https://devfeed.tech/topics/api-economy.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [API Logging](<https://devfeed.tech/topics/api-logging.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [api](<https://devfeed.tech/tags/api.md>), [app](<https://devfeed.tech/tags/app.md>), [business](<https://devfeed.tech/tags/business.md>), [draft](<https://devfeed.tech/tags/draft.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [id](<https://devfeed.tech/tags/id.md>), [idempotency](<https://devfeed.tech/tags/idempotency.md>), [payment](<https://devfeed.tech/tags/payment.md>), [request](<https://devfeed.tech/tags/request.md>), [revolut](<https://devfeed.tech/tags/revolut.md>), [state](<https://devfeed.tech/tags/state.md>)

### AI overview

A Revolut counterparty marked `created` only confirms the recipient was stored; it does not mean a payment was approved. To keep a person in the loop, create a payment draft and have an owner, admin, or designated approver send it and clear any `Pending review` in the Revolut Business app. A token with `WRITE` but not `PAY` can prepare drafts without initiating direct transfers; `PAY` can bypass that draft queue. For `POST /pay` retries after a timeout, reuse the same `request_id` (Revolut recommends UUID v4): duplicate prevention applies for about two weeks, after which reuse may create another transaction. The excerpt also cautions that recipient fields and account or card identifiers must match the counterparty, and sandbox Revtag tests must use Revolut's published test users.

### Source excerpt

A Revolut counterparty in state created is not a paid transfer. Payment drafts wait for approval in the Business app, and request_id dedupes for two weeks.

## Fujifilm pours $64 million into magnetic tape production as AI storage demand surges

DevFeed: [Fujifilm pours $64 million into magnetic tape production as AI storage demand surges](<https://devfeed.tech/articles/fujifilm-pours-64-million-into-magnetic-tape-production-as-ai-storage-demand-surges-109127.md>)

Original publisher: [Read original article](<https://www.techspot.com/news/114394-fujifilm-pours-64-million-magnetic-tape-production-ai.html>)

Author: Alfonso Maruccia

Published: 2026-10-30T17:38:00Z

Content type: news

Language: en

Sources: [TechSpot](<https://devfeed.tech/sources/techspot.md>)

Topics: [recording](<https://devfeed.tech/topics/recording.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [big-data](<https://devfeed.tech/topics/big-data.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-storage](<https://devfeed.tech/tags/ai-storage.md>), [capacity](<https://devfeed.tech/tags/capacity.md>), [data-center](<https://devfeed.tech/tags/data-center.md>), [data-storage](<https://devfeed.tech/tags/data-storage.md>), [demand](<https://devfeed.tech/tags/demand.md>), [factory](<https://devfeed.tech/tags/factory.md>), [fujifilm](<https://devfeed.tech/tags/fujifilm.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [industry](<https://devfeed.tech/tags/industry.md>), [investment](<https://devfeed.tech/tags/investment.md>), [japan](<https://devfeed.tech/tags/japan.md>), [lto](<https://devfeed.tech/tags/lto.md>), [magnetic](<https://devfeed.tech/tags/magnetic.md>), [production](<https://devfeed.tech/tags/production.md>), [storage](<https://devfeed.tech/tags/storage.md>), [tape](<https://devfeed.tech/tags/tape.md>)

### AI overview

Fujifilm plans to invest 10 billion yen (about $64 million) from fiscal 2026 through 2028 to expand magnetic-tape production at its Kanagawa Factory in Odawara, with new positions expected. The company says cloud providers' and data centers' growing storage needs, driven in part by generative AI services, mean output must rise to about five times its current level. The plant will focus on LTO-10 tapes, which hold 40 TB natively or up to 100 TB compressed, at a transfer rate of 400 MB/s. The article describes tape as a lower-cost option for long-term storage than SSDs, while noting that tapes are typically used for cold-storage workloads.

### Source excerpt

Fujifilm recently announced a significant new investment in its magnetic tape manufacturing operations. The Japanese giant will spend 10 billion yen, or around $64 million, to expand production capacity at its Kanagawa Factory in the city of Odawara. The site manufactures magnetic tapes for data storage workloads, and the IT... Read Entire Article

## The Illusion of Local Privacy: Confidentiality Boundary Failures in Consumer LLM Serving Systems

DevFeed: [The Illusion of Local Privacy: Confidentiality Boundary Failures in Consumer LLM Serving Systems](<https://devfeed.tech/articles/the-illusion-of-local-privacy-confidentiality-boundary-failures-in-consumer-llm-serving-systems-108999.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.18526>)

Author: Youssef Hamdi Zafan Ibrahim, Muhammad Ikram, Mohammed Khalaf Salama

Published: 2026-10-30T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Cryptography and Security](<https://devfeed.tech/sources/arxiv-cryptography-and-security.md>)

Topics: [LLM security](<https://devfeed.tech/topics/llm-security.md>), [llama.cpp](<https://devfeed.tech/topics/llama-cpp.md>), [ai security](<https://devfeed.tech/topics/ai-security.md>), [tenant data protection](<https://devfeed.tech/topics/tenant-data-protection.md>), [Local AI](<https://devfeed.tech/topics/local-ai.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [LLM Techniques](<https://devfeed.tech/topics/llm-techniques.md>), [interface](<https://devfeed.tech/topics/interface.md>)

Tags: [authorization](<https://devfeed.tech/tags/authorization.md>), [boundary](<https://devfeed.tech/tags/boundary.md>), [consumer](<https://devfeed.tech/tags/consumer.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [isolation](<https://devfeed.tech/tags/isolation.md>), [llm](<https://devfeed.tech/tags/llm.md>), [llm-serving](<https://devfeed.tech/tags/llm-serving.md>), [local](<https://devfeed.tech/tags/local.md>), [memory](<https://devfeed.tech/tags/memory.md>), [oracle](<https://devfeed.tech/tags/oracle.md>), [persistence](<https://devfeed.tech/tags/persistence.md>), [prefix-caching](<https://devfeed.tech/tags/prefix-caching.md>), [privacy](<https://devfeed.tech/tags/privacy.md>), [prompt](<https://devfeed.tech/tags/prompt.md>), [storage](<https://devfeed.tech/tags/storage.md>)

### AI overview

Keeping LLM inference on-device does not by itself ensure prompt confidentiality: serving software can expose prompts at several stages. The authors use LLAnalyzer to examine model loading, runtime memory, wrapper persistence, and serving interfaces across four open-weight model families and two consumer platforms. In a 24-hour AFL++ campaign exceeding 12 million executions, they found no parser crashes or successful malformed GGUF loads within the explored state space. By contrast, plaintext prompt remnants persisted in allocator-managed memory after inference; sanitisation reduced but did not remove them, and wrappers could prolong prompt lifetime through plaintext storage. They also report a llama.cpp authorization flaw that let one authenticated client restore another tenant's saved conversation in 200/200 controlled trials, plus a shared-prefix-cache timing oracle distinguishable over WAN. The authors call for explicit guarantees on prompt lifetime, persistent storage, and tenant isolation.

### Source excerpt

arXiv:2609.18526v3 Announce Type: replace Abstract: Running large language models (LLMs) locally is often considered more private than cloud-hosted inference because user prompts remain on the device. We ask whether keeping inference local is, by itself, sufficient to keep those prompts confidential. Our results show that it is not: prompt confidentiality also depends on how the surrounding serving software handles prompt data before, during, and after inference. We examine four boundaries at which prompt confidentiality can fail in consumer local-LLM serving systems: model loading, runtime memory, wrapper-level persistence, and the serving interface. To study these boundaries, we develop LLAnalyzer, a measurement framework that tests each boundary separately and traces observed failures to the responsible software component. Applying LLAnalyzer to four open-weight model families and two consumer deployment platforms, we find markedly different behaviour across boundaries. In a 24-hour AFL++ campaign with more than 12 million executions, we observe no parser crashes or successful malformed GGUF loads within the explored state space. Runtime memory tells a different story: we recover prompts after inference because multiple plaintext representations survive in allocator-managed memory, and sanitisation reduces this residue without eliminating it. We also find that consumer wrappers can extend prompt lifetime through plaintext persistence. At the serving boundary, we uncover a previously undocumented authorization flaw in llama.cpp that allows one authenticated client to restore another tenant's saved conversation state; the attack succeeds in 200/200 controlled trials. Separately, shared prompt-prefix caching exposes a remote timing oracle that remains distinguishable under WAN conditions. We argue that local LLM systems need explicit guarantees for prompt lifetime, persistent storage, and tenant isolation.

## Compaction as Epistemic Failure: How Agentic LLM Tools Fabricate Confirmed Results from Killed Processes

DevFeed: [Compaction as Epistemic Failure: How Agentic LLM Tools Fabricate Confirmed Results from Killed Processes](<https://devfeed.tech/articles/compaction-as-epistemic-failure-how-agentic-llm-tools-fabricate-confirmed-results-from-killed-processes-101717.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2607.13071>)

Author: Hiroki Tamba

Published: 2026-10-27T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [Reliable AI outputs](<https://devfeed.tech/topics/reliable-ai-outputs.md>), [Claude Code](<https://devfeed.tech/topics/claude-code.md>), [Agent Harness](<https://devfeed.tech/topics/agent-harness.md>), [agentic-coding](<https://devfeed.tech/topics/agentic-coding.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [reliability-engineering](<https://devfeed.tech/topics/reliability-engineering.md>), [LLM security](<https://devfeed.tech/topics/llm-security.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [compaction](<https://devfeed.tech/tags/compaction.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [failure](<https://devfeed.tech/tags/failure.md>), [llm](<https://devfeed.tech/tags/llm.md>), [llm-tools](<https://devfeed.tech/tags/llm-tools.md>), [results](<https://devfeed.tech/tags/results.md>), [tools](<https://devfeed.tech/tags/tools.md>)

### AI overview

The paper reports that Claude Code can turn partial output from timed-out commands (exit code 143) into "confirmed" results in compaction summaries. Later sessions may inherit those false positives as ground truth without re-verification. The described mechanism is a conflation of seeing information in terminal output with confirming it was durably stored. The authors say this threatens workflows that depend on session continuity for data processing, scientific computation, or multi-step automation. The supplied text is an abstract, so it does not provide further evidence or mitigation steps.

### Source excerpt

arXiv:2607.13071v2 Announce Type: replace Abstract: Agentic LLM coding tools compress long session histories into compaction summaries that subsequent sessions inherit as ground truth. This paper documents a failure mode in Claude Code where partial standard output from timed-out commands (exit code 143) is recorded in compaction summaries as confirmed results, propagating false positives across sessions and model versions without re-verification. The underlying mechanism is a conflation of observation and persistence, where information that appeared in the terminal is treated as equivalent to information written to durable storage. This finding extends the analysis of LLM self-evaluation failures reported in prior work on non-determinism in LLM-as-judge grading by showing that agentic tools exhibit analogous reliability deficits when reporting on their own operational outcomes. The failure has direct implications for any workflow that relies on agentic session continuity for data processing, scientific computation, or multi-step automation.

## LatentStore: An Efficient Latent-First Storage System for AI-Generated Images

DevFeed: [LatentStore: An Efficient Latent-First Storage System for AI-Generated Images](<https://devfeed.tech/articles/latentstore-an-efficient-latent-first-storage-system-for-ai-generated-images-96705.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2605.19385>)

Author: Zirui Wang, Yunjia Zheng, Tingfeng Lan, Zhaoyuan Su, Haoran Ni, Juncheng Yang, Yue Cheng

Published: 2026-10-23T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Databases](<https://devfeed.tech/sources/arxiv-databases.md>), [arXiv: Distributed Computing](<https://devfeed.tech/sources/arxiv-distributed-computing.md>)

Topics: [Image](<https://devfeed.tech/topics/image.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [Cache](<https://devfeed.tech/topics/cache.md>), [big-data](<https://devfeed.tech/topics/big-data.md>), [systems](<https://devfeed.tech/topics/systems.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-generated-images](<https://devfeed.tech/tags/ai-generated-images.md>), [bandwidth](<https://devfeed.tech/tags/bandwidth.md>), [cache](<https://devfeed.tech/tags/cache.md>), [cluster](<https://devfeed.tech/tags/cluster.md>), [cs-db](<https://devfeed.tech/tags/cs-db.md>), [cs-dc](<https://devfeed.tech/tags/cs-dc.md>), [generative](<https://devfeed.tech/tags/generative.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [images](<https://devfeed.tech/tags/images.md>), [latency](<https://devfeed.tech/tags/latency.md>), [prototype](<https://devfeed.tech/tags/prototype.md>), [request](<https://devfeed.tech/tags/request.md>), [storage](<https://devfeed.tech/tags/storage.md>), [storage-infrastructure](<https://devfeed.tech/tags/storage-infrastructure.md>), [sustainability](<https://devfeed.tech/tags/sustainability.md>), [system](<https://devfeed.tech/tags/system.md>), [tensors](<https://devfeed.tech/tags/tensors.md>), [trace](<https://devfeed.tech/tags/trace.md>), [trade](<https://devfeed.tech/tags/trade.md>), [user](<https://devfeed.tech/tags/user.md>)

### AI overview

LatentStore reduces persistent storage for AI-generated images by storing compressed, model-native latents and reconstructing images on demand with GPUs. Guided by a 35-month production trace covering 2 billion requests, its design keeps frequently accessed images in decoded pixel form for fast hits, stores less-active objects as latents to expand cache capacity, and continuously adjusts the balance between the two caches to manage user-perceived latency. In a prototype evaluated against the trace, LatentStore cut persistent storage by 78.7% while achieving competitive or lower mean and tail latency than pure image-based storage.

### Source excerpt

arXiv:2605.19385v3 Announce Type: replace-cross Abstract: The explosive growth of AI-generated images has created a sustainability challenge for storage infrastructure. Platforms like Midjourney and Adobe Firefly already host billions of generative images, yet conventional object stores persist them as blobs with full-resolution pixels, consuming huge amounts of storage capacity and bandwidth. Unlike natural photos, however, AI-generated images can be deterministically reconstructed from compact, model-native latent tensors, making persistent image storage fundamentally redundant. This paper presents LatentStore, a latent-first storage system for AI-generated images. LatentStore treats compressed latents as durable storage objects and uses on-demand GPU reconstruction on the read path to trade inexpensive compute for large persistent storage savings. Our design is guided by the first large-scale analysis of AI-generated image access we are aware of, based on a 35-month, 2-billion-request production trace from a major generative-content platform. Motivated by the trace analysis, LatentStore keeps frequently accessed images in decoded pixel format for fast hits, stores less-active objects as compressed latents to expand effective cache capacity, and continuously adjusts the splits between the image and latent cache to optimize user-perceived access latency. We build a LatentStore prototype and evaluate it with the production trace. LatentStore reduces persistent storage by 78.7% with competitive or even lower mean and tail latency over a pure image-based storage.

## AfterDeleteBench: Auditing Post-Deletion Information Persistence in LLM Agents

DevFeed: [AfterDeleteBench: Auditing Post-Deletion Information Persistence in LLM Agents](<https://devfeed.tech/articles/afterdeletebench-auditing-post-deletion-information-persistence-in-llm-agents-96608.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.31366>)

Author: Ehsan Saleh, Roozbeh Nahavandi, Sama Salarian, Raef Bassily, Andrew Perrault

Published: 2026-10-23T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Cryptography and Security](<https://devfeed.tech/sources/arxiv-cryptography-and-security.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [LLM security](<https://devfeed.tech/topics/llm-security.md>), [harvestbench](<https://devfeed.tech/topics/harvestbench.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [auditing](<https://devfeed.tech/tags/auditing.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [information](<https://devfeed.tech/tags/information.md>), [llm](<https://devfeed.tech/tags/llm.md>), [llm-agents](<https://devfeed.tech/tags/llm-agents.md>), [memory](<https://devfeed.tech/tags/memory.md>), [persistence](<https://devfeed.tech/tags/persistence.md>), [persistent-memory](<https://devfeed.tech/tags/persistent-memory.md>), [primary](<https://devfeed.tech/tags/primary.md>), [provenance](<https://devfeed.tech/tags/provenance.md>), [query](<https://devfeed.tech/tags/query.md>), [qwen](<https://devfeed.tech/tags/qwen.md>), [recovery](<https://devfeed.tech/tags/recovery.md>), [secret](<https://devfeed.tech/tags/secret.md>), [storage](<https://devfeed.tech/tags/storage.md>), [synthetic](<https://devfeed.tech/tags/synthetic.md>)

### AI overview

AfterDeleteBench tests whether fixed-weight LLM agents can recover information after its original memory record is deleted. It compares agents exposed to synthetic secrets and then cleared with matched agents never exposed. In the primary Qwen2.5-7B-Instruct setting, one query recovered the deleted secret in roughly two thirds of cases; controls had zero recovery. When only secret components remained under a public composition rule, exact reconstruction succeeded in 62-99% of cases across Qwen, Mistral, and Phi-3. The authors note that format and parser mismatches can score target-bearing replies as failures. Provenance-aware closure eliminated observed recovery in tested settings with positive baseline recovery, while an exact-literal write filter blocked reinsertion but could not erase information already present. The findings support evaluating deletion behavior across the agent and its memory system, rather than storage alone.

### Source excerpt

arXiv:2610.31366v1 Announce Type: new Abstract: Many LLM agents maintain external persistent memory and derive additional state from past interactions, such as summaries. Deleting an original memory record therefore need not remove information carried by its descendants. We introduce AfterDeleteBench, a controlled behavioral audit for this failure mode in fixed-weight LLM agents. For each synthetic secret, we compare an exposed-then-deleted agent with an otherwise matched agent that was never exposed to the secret, allowing post-deletion recovery to be attributed to the original exposure. In our primary setting, a single query recovers the deleted secret from Qwen2.5-7B-Instruct in roughly two thirds of cases, while matched never-exposed controls have zero recovery. Crucially, recovery does not require a surviving verbatim copy: when the complete secret string is absent and only separately stored components remain under a public composition rule, exact reconstruction succeeds in 62-99% of cases across Qwen, Mistral, and Phi-3. We also show that the response interface is part of the measurement system: target-bearing replies can be scored as failures when the requested format and parser are misaligned. Finally, we separate reconstruction from write-back and fresh-session recall. Provenance-aware closure eliminates observed recovery in the tested positive-baseline settings, whereas an exact-literal write filter blocks reinsertion but cannot erase information already present. Together, these results show that source deletion should be evaluated behaviorally at the level of the agent and its memory system, not only as a storage operation.

## The science behind TorchNeuron, Amazon's new PyTorch backend for AWS Trainium chips

DevFeed: [The science behind TorchNeuron, Amazon's new PyTorch backend for AWS Trainium chips](<https://devfeed.tech/articles/the-science-behind-torchneuron-amazon-s-new-pytorch-backend-for-aws-trainium-chips-96329.md>)

Original publisher: [Read original article](<https://www.amazon.science/blog/the-science-behind-torchneuron-amazons-new-pytorch-backend-for-aws-trainium-chips>)

Author: Robert Hundt

Published: 2026-10-22T19:36:42Z

Content type: article

Language: en

Sources: [Amazon Science homepage](<https://devfeed.tech/sources/amazon-science-homepage.md>)

Topics: [AWS Trainium](<https://devfeed.tech/topics/aws-trainium.md>), [PyTorch](<https://devfeed.tech/topics/pytorch.md>), [neuron](<https://devfeed.tech/topics/neuron.md>), [AWS AI chips](<https://devfeed.tech/topics/aws-ai-chips.md>), [Amazon EC2](<https://devfeed.tech/topics/amazon-ec2.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [Graviton](<https://devfeed.tech/topics/graviton.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-models](<https://devfeed.tech/tags/ai-models.md>), [amazon](<https://devfeed.tech/tags/amazon.md>), [aws](<https://devfeed.tech/tags/aws.md>), [aws-trainium](<https://devfeed.tech/tags/aws-trainium.md>), [backend](<https://devfeed.tech/tags/backend.md>), [chips](<https://devfeed.tech/tags/chips.md>), [cloud-and-systems](<https://devfeed.tech/tags/cloud-and-systems.md>), [efficiently](<https://devfeed.tech/tags/efficiently.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [models](<https://devfeed.tech/tags/models.md>), [optimizations](<https://devfeed.tech/tags/optimizations.md>), [pytorch](<https://devfeed.tech/tags/pytorch.md>), [runtime](<https://devfeed.tech/tags/runtime.md>), [torchneuron](<https://devfeed.tech/tags/torchneuron.md>), [trainium](<https://devfeed.tech/tags/trainium.md>)

### AI overview

TorchNeuron is an open-source backend that translates PyTorch code to run on AWS Trainium, generally with minimal or no code changes. It offers opt-in adaptive eager mode, which gathers operations into small graphs, compiles and caches them, and overlaps compilation with execution to reduce overhead while preserving source-code correlation for debugging. Generated executables (NEFFs) are cached in memory and on disk; across accelerators, compilation can be shared so identical code is compiled once and reused. In a test of a large language model spread across accelerators, these optimizations cut time to the first training step by about a third. For evaluation, AI agents surveyed open-source models, selected 841 representing varied architectures, and built a registry of tensor shapes. Fast shape-based tests screen revisions before tests on the representative set and then the full model distribution, enabling earlier regression detection.

### Source excerpt

Runtime optimizations and a thorough evaluation process driven by agentic AI help ensure that models written in PyTorch will execute efficiently and reliably on Trainium systems.

## SBD: We Have a Filesystem at Home

DevFeed: [SBD: We Have a Filesystem at Home](<https://devfeed.tech/articles/sbd-we-have-a-filesystem-at-home-95252.md>)

Original publisher: [Read original article](<https://fly.io/blog/sprites-block-device/>)

Author: Daniel Botha

Published: 2026-10-22T00:00:00Z

Content type: release

Language: en

Sources: [Fly](<https://devfeed.tech/sources/the-fly-blog.md>)

Topics: [Ext4](<https://devfeed.tech/topics/ext4.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [Disk image](<https://devfeed.tech/topics/disk-image.md>), [Integration testing](<https://devfeed.tech/topics/integration-testing.md>), [mount](<https://devfeed.tech/topics/mount.md>), [NVMe](<https://devfeed.tech/topics/nvme.md>)

Tags: [block](<https://devfeed.tech/tags/block.md>), [cdn](<https://devfeed.tech/tags/cdn.md>), [close-to-users](<https://devfeed.tech/tags/close-to-users.md>), [database](<https://devfeed.tech/tags/database.md>), [deploy-app-servers](<https://devfeed.tech/tags/deploy-app-servers.md>), [device](<https://devfeed.tech/tags/device.md>), [docker](<https://devfeed.tech/tags/docker.md>), [elixir](<https://devfeed.tech/tags/elixir.md>), [ext4](<https://devfeed.tech/tags/ext4.md>), [filesystem](<https://devfeed.tech/tags/filesystem.md>), [fly](<https://devfeed.tech/tags/fly.md>), [fly-io](<https://devfeed.tech/tags/fly-io.md>), [heroku-alternative](<https://devfeed.tech/tags/heroku-alternative.md>), [heroku-competitor](<https://devfeed.tech/tags/heroku-competitor.md>), [hosting](<https://devfeed.tech/tags/hosting.md>), [i](<https://devfeed.tech/tags/i.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [linux](<https://devfeed.tech/tags/linux.md>), [metadata](<https://devfeed.tech/tags/metadata.md>), [networking](<https://devfeed.tech/tags/networking.md>), [object](<https://devfeed.tech/tags/object.md>), [object-storage](<https://devfeed.tech/tags/object-storage.md>), [postgresql-clusters](<https://devfeed.tech/tags/postgresql-clusters.md>), [reliability](<https://devfeed.tech/tags/reliability.md>), [servers](<https://devfeed.tech/tags/servers.md>), [storage](<https://devfeed.tech/tags/storage.md>), [the-fly-blog](<https://devfeed.tech/tags/the-fly-blog.md>)

### AI overview

Sprites replaces its heavily modified JuiceFS setup with the Sprites Block Device (SBD): a daemon presents bucket-backed storage to Linux as an NBD block device, which Sprite formats as ext4. The prior design stored a 100 GB disk image and relied on separate JuiceFS metadata and checkpoint databases; writeback could leave metadata pointing to an object that was never uploaded, and cloned checkpoints could share that missing chunk. SBD keeps volume metadata with the blocks in the bucket, caches locally, uploads dirty blocks in the background, and flushes when durability must be guaranteed. Checkpoints record block snapshots and volume versions in SBD's catalog; restore rewinds the volume in place. SBD verifies blocks read from storage and restarts the Sprite if its daemon dies, favoring a clean restart over continued use of an untrusted disk. The post reports that stress testing and design partner feedback found SBD orders of magnitude more reliable, and says it will replace the old filesystem.

### Source excerpt

I have unkind things to say about people who make their own pasta. Here, I'll only use them as an example of how not to think when you're building infrastructure. It's unfairly easy to believe that the harder you work on something, the better it gets. But infra is about making the $1 bag of spaghetti: a simple thing that works well (and doesn't require you to spend a week scraping flour out of the corners of your kitchen). When work on Sprites started, it was a new kind of infra for a new kind of product. So when we couldn't find the right filesystem in the pasta section, we made our own. It was clever, and that's exactly why we didn't like it. This post introduces its replacement, the Sprites Block Device (SBD). It's the bag of spaghetti you grab off the shelf, and the only difference you'll notice is better reliability. It was always object storage Our stance on one core aspect of Sprites' storage hasn't changed from v1: the root of a Sprite's disk is object storage, not the NVMe in whatever server the Sprite happens to be running on. That's what lets a Sprite sleep, wake on a different server, and still find its files where it left them. The durable state of a Sprite is, more or less, a location in a bucket. We wrote about why, at length, with feeling. The thing about object stores is they can't change part of an object. To change one byte, you have to replace the whole thing. That's a bad cultural fit for Linux programs which rewrite a few bytes in the middle of a file all the time. Something has to handle that translation. At launch, that something was a heavily modded version of JuiceFS. In stock form, JuiceFS splits storage into the data (as chunks in the bucket) and the metadata (a map of which chunks make up which files) in a database. That's not how we used it, though. Our hot rodded version ran a rewritten SQLite metadata backend, kept durable with Litestream. It stored exactly one file: a 100 GB disk image, mounted on a loop device with ext4 inside. Linu

## Level up your AI-powered game backend with Amazon DynamoDB new features

DevFeed: [Level up your AI-powered game backend with Amazon DynamoDB new features](<https://devfeed.tech/articles/level-up-your-ai-powered-game-backend-with-amazon-dynamodb-new-features-93967.md>)

Original publisher: [Read original article](<https://aws.amazon.com/blogs/gametech/level-up-your-ai-powered-game-backend-with-amazon-dynamodb-new-features/>)

Author: Naveen Kantamneni

Published: 2026-10-21T23:05:45Z

Content type: article

Language: en

Sources: [AWS for Games Blog](<https://devfeed.tech/sources/aws-for-games-blog.md>)

Topics: [Amazon DynamoDB](<https://devfeed.tech/topics/amazon-dynamodb.md>), [Amazon SageMaker](<https://devfeed.tech/topics/amazon-sagemaker.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [AI for interactive applications](<https://devfeed.tech/topics/ai-for-interactive-applications.md>)

Tags: [agent-state](<https://devfeed.tech/tags/agent-state.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-inference](<https://devfeed.tech/tags/ai-inference.md>), [ai-powered](<https://devfeed.tech/tags/ai-powered.md>), [amazon-dynamodb](<https://devfeed.tech/tags/amazon-dynamodb.md>), [amazon-s3](<https://devfeed.tech/tags/amazon-s3.md>), [amazon-sagemaker-lakehouse](<https://devfeed.tech/tags/amazon-sagemaker-lakehouse.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [apache-iceberg](<https://devfeed.tech/tags/apache-iceberg.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [aws-glue](<https://devfeed.tech/tags/aws-glue.md>), [aws-lambda](<https://devfeed.tech/tags/aws-lambda.md>), [backend](<https://devfeed.tech/tags/backend.md>), [dynamodb](<https://devfeed.tech/tags/dynamodb.md>), [game](<https://devfeed.tech/tags/game.md>), [games](<https://devfeed.tech/tags/games.md>), [gaming](<https://devfeed.tech/tags/gaming.md>), [industries](<https://devfeed.tech/tags/industries.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [new-features](<https://devfeed.tech/tags/new-features.md>)

### AI overview

The text presents four DynamoDB changes as building blocks for AI game backends that need durable player and agent state: Multi-Region strong consistency (MRSC), zero-ETL integration with Amazon SageMaker Lakehouse, LangGraph checkpoint storage, and lower on-demand pricing. MRSC synchronously replicates writes within supported US, EU, or AP Region sets and supports current-version reads, with a stated single-digit-millisecond overhead. It excludes TTL, LSIs, and transaction operations; cross-continent strong consistency is unavailable. SageMaker Lakehouse replication uses CDC and Iceberg without consuming table read capacity, but setup requires PITR and AWS Glue export permissions. LangGraph DynamoDBSaver persists session checkpoints and offloads payloads above 350 KB to S3. The text reports a 50% on-demand throughput price cut and up to 67% lower replicated-write pricing; warm throughput can support pre-established peaks, though performance varies with workload and settings.

### Source excerpt

The most difficult engineering problem when employing AI for gaming isn't selecting the correct model--it's selecting the infrastructure that supports the functionality needed to create compelling user experiences. A generative non-player character (NPC) that forgets every conversation, an AI dungeon master that loses narrative context mid-quest, or an agentic loop that drops state across an AWS Lambda cold start all produce sub-par experiences for the same reason: no durable, globally consistent state storage layer.

## Vercel Open-Sources durable-harness for Crash-Proof Coding Agents

DevFeed: [Vercel Open-Sources durable-harness for Crash-Proof Coding Agents](<https://devfeed.tech/articles/vercel-open-sources-durable-harness-for-crash-proof-coding-agents-94009.md>)

Original publisher: [Read original article](<https://www.marktechpost.com/2026/10/21/vercel-durable-harness-crash-proof-coding-agents/>)

Author: Michal Sutter

Published: 2026-10-21T22:15:39Z

Content type: article

Language: en

Sources: [MarkTechPost](<https://devfeed.tech/sources/marktechpost.md>)

Topics: [Agent Harness](<https://devfeed.tech/topics/agent-harness.md>), [agentic-coding](<https://devfeed.tech/topics/agentic-coding.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [Vercel](<https://devfeed.tech/topics/vercel.md>)

Tags: [adapters](<https://devfeed.tech/tags/adapters.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai-shorts](<https://devfeed.tech/tags/ai-shorts.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [claude-code](<https://devfeed.tech/tags/claude-code.md>), [codex](<https://devfeed.tech/tags/codex.md>), [coding](<https://devfeed.tech/tags/coding.md>), [coding-agents](<https://devfeed.tech/tags/coding-agents.md>), [crash](<https://devfeed.tech/tags/crash.md>), [crashes](<https://devfeed.tech/tags/crashes.md>), [editors-pick](<https://devfeed.tech/tags/editors-pick.md>), [for-devs](<https://devfeed.tech/tags/for-devs.md>), [harness](<https://devfeed.tech/tags/harness.md>), [laptop](<https://devfeed.tech/tags/laptop.md>), [logs](<https://devfeed.tech/tags/logs.md>), [model](<https://devfeed.tech/tags/model.md>), [new-releases](<https://devfeed.tech/tags/new-releases.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [opencode](<https://devfeed.tech/tags/opencode.md>), [package](<https://devfeed.tech/tags/package.md>), [pi](<https://devfeed.tech/tags/pi.md>), [postgres](<https://devfeed.tech/tags/postgres.md>), [sessions](<https://devfeed.tech/tags/sessions.md>), [staff](<https://devfeed.tech/tags/staff.md>), [tech-news](<https://devfeed.tech/tags/tech-news.md>), [technology](<https://devfeed.tech/tags/technology.md>), [tool](<https://devfeed.tech/tags/tool.md>), [vercel](<https://devfeed.tech/tags/vercel.md>)

### AI overview

Vercel's experimental durable-harness logs model and tool call intents and outcomes, then resumes a turn from that log if its function crashes. Completed calls are read back rather than rerun; tools can be configured for safe replay or never to replay. Harness session files are mirrored too. Pi and OpenCode can run in the function, while Claude Code and Codex require a Vercel Sandbox. Workspaces include a networkless virtual filesystem, temporary in-memory scratch files, or a Sandbox with Drive storage. The package can run on Vercel, locally, or self-hosted with Postgres on Node.js 22+; the excerpt also lists MongoDB, Redis, and Turso. It is version 0.0.x, and Vercel has published no performance benchmarks; recovery claims are self-reported.

### Source excerpt

Vercel released durable-harness, an open-source package that logs every model and tool call so Pi, OpenCode, Claude Code and Codex sessions resume after crashes, timeouts and redeploys on Vercel, Postgres or a laptop. The post Vercel Open-Sources durable-harness for Crash-Proof Coding Agents appeared first on MarkTechPost.

## Building an AEO scoring pipeline with AWS Lambda durable functions

DevFeed: [Building an AEO scoring pipeline with AWS Lambda durable functions](<https://devfeed.tech/articles/building-an-aeo-scoring-pipeline-with-aws-lambda-durable-functions-91266.md>)

Original publisher: [Read original article](<https://aws.amazon.com/blogs/compute/building-an-aeo-scoring-pipeline-with-aws-lambda-durable-functions/>)

Author: Nick Beall

Published: 2026-10-20T17:45:29Z

Content type: tutorial

Language: en

Sources: [AWS Compute Blog](<https://devfeed.tech/sources/aws-compute-blog.md>)

Topics: [Amazon Web Services](<https://devfeed.tech/topics/aws.md>), [Amazon Bedrock](<https://devfeed.tech/topics/amazon-bedrock.md>), [function as a service](<https://devfeed.tech/topics/function-as-a-service.md>), [Persistence](<https://devfeed.tech/topics/persistence.md>), [Amazon EventBridge](<https://devfeed.tech/topics/amazon-eventbridge.md>)

Tags: [advanced-300](<https://devfeed.tech/tags/advanced-300.md>), [aeo](<https://devfeed.tech/tags/aeo.md>), [agent](<https://devfeed.tech/tags/agent.md>), [aws](<https://devfeed.tech/tags/aws.md>), [aws-lambda](<https://devfeed.tech/tags/aws-lambda.md>), [build](<https://devfeed.tech/tags/build.md>), [building](<https://devfeed.tech/tags/building.md>), [can](<https://devfeed.tech/tags/can.md>), [costs](<https://devfeed.tech/tags/costs.md>), [deploy](<https://devfeed.tech/tags/deploy.md>), [durable-functions](<https://devfeed.tech/tags/durable-functions.md>), [engine](<https://devfeed.tech/tags/engine.md>), [functions](<https://devfeed.tech/tags/functions.md>), [lambda](<https://devfeed.tech/tags/lambda.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [pipeline](<https://devfeed.tech/tags/pipeline.md>), [retries](<https://devfeed.tech/tags/retries.md>), [sam](<https://devfeed.tech/tags/sam.md>), [score](<https://devfeed.tech/tags/score.md>), [solution](<https://devfeed.tech/tags/solution.md>), [technical-how-to](<https://devfeed.tech/tags/technical-how-to.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

The article describes a five-step AEO audit pipeline built as an AWS Lambda durable function: fetch a page, score it, analyze the results, publish HTML and JSON reports to Amazon S3, and record metadata in DynamoDB. Its five signals--discoverability, parsability, token efficiency, capability signaling, and access control--use deterministic Python checks for a reproducible 0-100 score; Amazon Bedrock separately generates prose findings and recommendations. Checkpoints let failed runs resume at the failed step, avoiding repeated completed work, and the function suspends between stages without incurring additional costs. An optional callback can pause for human approval. The fetch uses plain HTTP GET, so pages dependent on client-side JavaScript may score lower. Deployment requires AWS Bedrock access, AWS SAM CLI v1.100+, Python 3.12+, and CloudFormation deployment permissions; the excerpt gives `sam build && sam deploy --guided` as the deployment command.

### Source excerpt

Learn how to build a five-step Agent Engine Optimization (AEO) scoring pipeline on AWS Lambda durable functions. Each checkpointed step suspends without incurring costs between stages and retries without re-spending on prior work, so you can score any page against five AEO signals and deploy the solution with an AWS SAM template.

[Next page](<https://devfeed.tech/topics/persistence.md?cursor=WyIyMDI2LTEwLTIwVDE3OjQ1OjI5KzAwOjAwIiwgImQxOWEwZjExLWU4ZDItNDQ3Ny1iMjZlLTEyODczZmFiYTgwMiJd>)