How to build a knowledge factory for AI agents
The knowledge factory is how LangChain keeps its agents' company knowledge current: agents flag what's missing, and a curator agent fixes it in the background.
Search headlines, summaries and articles to connect the dots.
The knowledge factory is how LangChain keeps its agents' company knowledge current: agents flag what's missing, and a curator agent fixes it in the background.
Targets Go to use the toolchain and the package system you already know and love Comments URL: https://news.ycombinator.com/item?id=50595828 Points: 2 # Comments: 0
Article URL: https://keepplain.com/blame Comments URL: https://news.ycombinator.com/item?id=50594433 Points: 1 # Comments: 0
Offline password manager for Android: a fork of Passdroid (last release 2012), brought up to date. The app has no internet permission.Two key new features: - 2FA codes (TOTP) - fingerprint unlockPlus many smaller changes and fixes, and security hardening: AES-256 with PBKDF2 at 600k iterations, AES-GCM encrypted exports. Threat model and design are in INTERNALS.md.Built largely with Claude Code; the design and security decisions are mine.Feedback welcome, especially on the security side. Comments URL: https://news.ycombinator.com/item?id=50594645 Points: 1 # Comments: 0
Self-hosting LLMs on GPUs means paying to recompute the same context on every turn. Distributed KV caching with Amazon ElastiCache can cut inference cost by up to 40% and time to first token by up to 80%. This post walks through the distributed KV caching skill in the Agent Toolkit for AWS, which evaluates and sets up the cache from your coding agent.
Standard Bank Group automated the evidence behind its generative AI approval gates. Working with AWS Professional Services, the bank built an assessment in its own AWS account that scores each use case against international AI standards and its own policies using LLM-as-a-judge on Amazon Bedrock, giving risk committees consistent, traceable evidence.
Amazon SageMaker HyperPod now offers managed cluster extensions for observability and user management, bringing declarative, API- and console-driven configuration to the software running on Slurm cluster nodes. This post explains how the extensions work and follows a platform team through a quarter of running a shared Slurm cluster with them.
Enforcing physics-based constraints prevents AI models from generating unreliable predictions for complex nonlinear systems in weather, climate, and turbulent flows.
AWS Quick is the AI workspace built only for work. Today we're releasing updates that help your team get more done: build and run business-process flows in chat, triage your day in the updated web app, get more complete answers from multiple agents at once, and connect new data sources like Meta ads and Amplitude.
Hugging Face Transformers v5.24.0 introduces several new models including Inkling 2 Mini, Apertus 1.5, WavTokenizer, Dory, HY4-VL, Omnilingual ASR, Lara Mini, SarvamMLA, SarvamMoE, and EXAONE. The release features a major breaking change renaming `past_key_values` and `cache_params` to `cache` across the library, standardizing cache handling. Multimodal models now return shared output classes instead of model-specific ones. Additional updates include improvements to generation, continuous batching, chat templates, pipelines, vision and audio processing, quantization, parallelization, exporters, and numerous bug fixes.
Building a harness with a context that's bulk-addressable via familiar pattern primitives empowers local models to succeed with a fraction of the context size and helps all models avoid amnesia entirely on long-running sessions. Comments URL: https://news.ycombinator.com/item?id=50579653 Points: 1 # Comments: 0
A personal project for digging deeper into Linux system programming, ObFUSEcate implements a file backed virtual block device that automatically performs compression and encryption as blocks are accessed. This virtual device can be mounted via FUSE (requiring no root privileges) to provide, effectively, a "live" compressed, encrypted archive.Written in C++, and intended to have minimal dependencies, it has custom 32 bit arithmetic encoding/decoding, ChaCha20-Poly1305 authenticated encryption, and CRC8 + XOR parity error-correction. Comments URL: https://news.ycombinator.com/item?id=50585222 Points: 1 # Comments: 0
Article URL: https://github.com/JustVugg/nostromo Comments URL: https://news.ycombinator.com/item?id=50579736 Points: 1 # Comments: 0
I've seen too many slow-log queries with a GROUP BY that makes Postgres blow up, so for the last two weeks or so I've been working on Straum. You declare the aggregates as a view, and Straum answers over HTTP (other transports might be added). I've been using the New York taxi dataset as a benchmark. On the heavier GROUP BYs I see about 1,000x to 5,000x. It's written in Rust (of course).It leans on Postgres replication slots, logical or physical, to get its data. Postgres stays the source of truth.Benchmark: https://janbjorge.github.io/straum/benchmark.html Repo: https://github.com/janbjorge/straum Comments URL: https://news.ycombinator.com/item?id=50578332 Points: 1 # Comments: 0
AI operational readiness requires more than selecting the right servers, storage and computing capacity. Those considerations are particularly consequential for financial institutions, where reliability, regulatory obligations and the need to demonstrate returns on technology investments shape deployment decisions. Meanwhile, AI is changing the threat landscape, giving less sophisticated attackers access to capabilities that once required […] The post Three insights you may have missed from theCUBE’s coverage of the Dell Technologies New York Symposium appeared first on SiliconANGLE.
Amazon Bedrock AgentCore harness now runs on the open-source Strands harness, bringing a recommended model, task-oriented defaults, automatic prompt caching, proactive context management, and new built-in tools. It also adds lifecycle hooks so you can run deterministic allow or deny checks at defined points in the agent loop.
This second installment reviews 13 more Amazon SageMaker AI inference launches from 2026, spanning configuration recommendations, fitting more models onto each GPU, serving models too large for one instance, cross-Region routing, and post-launch tuning, bringing the year's total to 26.
Today we are releasing DuckDB 2.0.1, the first patch release of our 2.0 LTS line. This release ships bugfixes, performance improvements, and security patches. In addition, a capability is added to define custom extension repositories with only their hostname.
The update introduces a new tensor API conv_2d kernel for Metal, enabling split-K optimizations across all chips, including M5. It improves efficiency by utilizing tensor cores, reducing idle GPU time during long reductions. The kernel now splits reductions into partial sums stored in buffers, with a follow-up kernel summing them deterministically. Address math is optimized using 32-bit byte offsets, reducing computational overhead. A fallback to the generic kernel is implemented for images exceeding 4 GiB. Performance benchmarks show significant speed improvements on M5 Max devices, with reduced execution times across various convolution configurations.
The llama.cpp v1.25.49 release includes fixes for Inkling 2 attention mask handling, making the relative-position input a hybrid iswa input, enabling optional expert weight scale and final logit softcap, and improving KV rows mask generation for hybrid iswa memory. The update also standardizes code formatting in inkling2.cpp and ensures compatibility across multiple platforms including macOS, Linux, Windows, Android, and openEuler.