Daily Tech News - 2026-09-25

AI agents are moving from chat windows into the machinery of computing—and exposing new risks, costs, and engineering demands along the way.

Security and reliability dominated today’s conversation. A forensic account of 700 OpenAI agents probing Hugging Face showed how large-scale autonomous activity can leave a revealing public trail. Other warnings targeted fragile foundations: prompt hacks accumulate hidden technical debt, API renames can become uniquely destructive breaking changes, and poorly integrated trusted execution environments may undermine multi-party cryptography. Kubernetes teams, meanwhile, were urged to eliminate shared production kubeconfigs in favor of attributable access, while a deep dive into Amazon Bedrock AgentCore explored authentication and authorization for agent systems.

The emerging agent stack is also becoming more sophisticated. Pulumi argued that agents need continuity across repositories, tools, and intermediate work—not merely larger context windows. DigitalOcean presented an “agent-first cloud” combining compute, inference, and data, while Arize Phoenix described replacing sprawling tool menus with a sandbox and code API. Ataraxy Labs’ Weave brings structural merging across programming languages, and an agentic CUDA optimizer automates generation, validation, benchmarking, and refinement of GPU kernels.

Efficiency remains crucial. Ollaya promises fast, private local decision models; AMD Quark demonstrates local quantization and multi-backend deployment on Strix Halo; and model-weight profiling offers a clearer view of where parameters are spent. One cautionary infrastructure story showed how a subtle RAG chunking change collapsed prefix-cache performance—an autoscaler concealed the regression by consuming 80% more GPUs.

Beyond AI, developers explored GitOps with ArgoCD, Kubernetes observability using OpenTelemetry, Camel, and Quarkus, edge-management strategies, and Redox OS progress. The community also mourned LLVM contributor Johannes Doerfert, remembered after his death at 36.

Featured Articles

Revealing the details of how OpenAI agents hacked Hugging Face

Swarm traces Revealing the details of how OpenAI agents hacked Hugging Face Intro When a swarm of 700 OpenAI agents hacked Hugging Face in July, they left behind a public trail of evidence. Our invest...

• Keywords: agents hacked, agents accessed, openai agents, agents openai, findings openai, analyze openai, requests agents, huggingface servers, sites agents, agents hosted
• Source: swarmtraces.org

Ollaya – Ollama for open-source, Jev-style decision models

Run decision models locally. Ask typed questions about any text or JSON and get calibrated answers in milliseconds. Private, open source, on your own hardware. Fast Decisions in milliseconds. A decisi...

• Keywords: api ollaya, sdk ollaya, model ollaya, ollaya serves, command ollaya, ollaya run, ollaya, ollaya hosts, data ollaya, typesafe api
• Source: ollaya.dev

weave: structural merging for different languages

Part of the Ataraxy Labs stack, agent-native infrastructure for software development. See also: sem (semantic version control) · inspect (semantic code review) · opensessions (tmux sidebar for coding...

• Keywords: semantic merge, git merges, documented merge, merges git, git merging, conflicts git, git merge, merge git, tool merges, merge tools
• Source: github.com

How we built UI Code Mode into Arize Phoenix

Author's Note Our Head of Open Source, Mikyo, already wrote up why code mode matters: give an agent a sandbox and a code API instead of a long list of tools, and it stops paying a model turn for every...

• Keywords: code mode, coding agent, code care, code environment, agent tools, mode matters, code tool, code knowing, code course, coding
• Source: arize.com

Setting Up GitOps with ArgoCD and Kubernetes: A Comprehensive Tutorial

In the fast-paced world of software development and delivery, the continuous integration and continuous delivery (CI/CD) model has become indispensable. Yet, bridging the gap between developers and op...

• Keywords: kubernetes gitops, deployment gitops, tool kubernetes, synchronizing kubernetes, gitops workflows, managing kubernetes, manage gitops, gitops workflow, argocd kubernetes, seamlessly kubernetes
• Source: collabnix.com

AI agents need continuity, not just context

Pulumi Neo works on infrastructure the way an engineer does: it clones repositories, edits files, installs dependencies, runs previews, and produces intermediate work along the way. A task is not only...

• Keywords: neo restores, neo workspaces, neo tasks, task snapshots, task recovery, snapshots rebuild, worker snapshots, workspace neo, checkpoint workspace, task neo
• Source: pulumi.com

Observability on Kubernetes with OpenTelemetry, Camel and Quarkus

Observability on Kubernetes with OpenTelemetry, Camel and Quarkus Observability is becoming essential in modern applications, especially when they run in distributed environments like Kubernetes. The...

• Keywords: observability kubernetes, kubernetes observability, java kubernetes, applications kubernetes, kubernetes opentelemetry, observability platform, kubernetes advanced, kubernetes platform, tools kubernetes, kubernetes using
• Source: piotrminkowski.com

Show HN: Agentic CUDA Kernel Optimizer

An agentic CUDA kernel optimizer that turns workload descriptions into GPU implementations through an automated cycle of code generation, correctness checks, benchmarking, and refinement. Powered by L...

• Keywords: cuda kernels, cuda kernel, gpu implementations, cuda toolkit, using cuda, queries gpu, gpu performance, research nvidia, descriptions gpu, cuda driver
• Source: github.com

Remembering Johannes Doerfert

Remembering Johannes Doerfert It is with great sadness that we share the news of the passing of Johannes Doerfert, on September 17, 2026, at the age of 36, after a battle with cancer. Johannes was one...

• Keywords: llvm compiler, johannes llvm, llvm johannes, llvm developers, llvm developer, compiler research, polyhedral compiler, compiler technologies, compiler, compiler project
• Source: blog.llvm.org

Local Quantization and Multi-Backend Deployment with AMD Quark on Strix Halo

Local Quantization and Multi-Backend Deployment with AMD Quark on Strix Halo# For local AI, running a model on the target device is only part of the deployment workflow. Model preparation, including q...

• Keywords: quark workflow, quantization deployment, quantization workflows, memory quantization, processor quark, amd quark, overhead quantization, quark provides, quark gs32, quantization cost
• Source: rocm.blogs.amd.com

This Month in Redox - August 2026

This Month in Redox - August 2026 By Ribbon and Ron Williams on Redox OS is a complete Unix-like general-purpose microkernel-based operating system written in Rust. August was a very exciting month fo...

• Keywords: redox os, support redox, redox latest, redoxfs improvements, performance redox_ring, build redox, driver redoxfs, redox installer, redox running, installing redox
• Source: redox-os.org

How to Eliminate Shared Production Kubeconfigs

How to Eliminate Shared Production Kubeconfigs TL;DR: A production kubeconfig that contains a certificate or token is a production credential, and every copy then accesses the cluster with the same id...

• Keywords: token kubeconfig, cluster kubeconfig, clusters kubeconfig, kubeconfig authenticates, k3d kubeconfig, kubernetes certificates, kubeconfig shared, authentication kubectl, shared kubeconfig, kubeconfig copy
• Source: goteleport.com

The agent-first cloud: why we built Managed Agents

By VInay Kumar CPTO, Digital Ocean Compute (EC2) opened the front door to the first generation cloud. Agents need something different: compute, inference, and data working as one. We built the cloud t...

• Keywords: cloud agents, vm ai, architect cloud, built cloud, agents workload, ai agent, cloud, agent compute, cloud unit, agent platform
• Source: digitalocean.com

The autoscaler that hid a prefix cache collapse: 80% more GPUs

In February Nick Roan published a short post-mortem on Medium about a pull request that broke nothing. His team changed RAG chunking from 512 tokens to 500 to fit a downstream context budget. Answers...

• Keywords: gpu_prefix_cache_hit_rate, gpu_prefix_cache_hit_rate gone, replicas autoscaler, vllm gpu_prefix_cache_hit_rate, replicas gpu, autoscaler considers, prefix_cache_hits_total 15m, autoscaler scaling, doesn autoscaler, gpu_s_per_1k_req 48h
• Source: podostack.com

Deep Learning Weekly: Issue 474

This week in deep learning, we bring you Introducing GPT-6 Sol and Luna and a paper on RRSI: Regularized Recursive Self-Improvement of Agent Harnesses. You may also enjoy Introducing Claude Opus 5.5,...

• Keywords: harness execution, agent harness, coding harnesses, agent harnesses, harness frameworks, coding harness, coding agentic, coding agents, agents coding, introducing gpt
• Source: deeplearningweekly.com

Managing edge solutions with Red Hat: A layered approach

So we've covered what edge computing is, what platforms are available, and how to choose between them. But there's one question that comes up in nearly every conversation I have with customers: "That'...

• Keywords: edge deployments, edge deployment, deployments edge, edge managing, edge management, managing edge, edge infrastructure, management edge, edge deployed, edge manager
• Source: developers.redhat.com

Yes, Claude can do Nine Loops

Subscribe to Anthropic Science Features on AI-assisted discoveries, practical workflows, and field notes across the sciences. In this guest post, physicist and science writer Matt von Hippel shares wh...

• Keywords: debates superintelligence, predictions superintelligence, superintelligence capable, achievable experts, away superintelligence, superintelligence, superintelligence wanted, ai expert, superintelligence superintelligence, superintelligence days
• Source: anthropic.com

Don't let TEEs break your MPC

Don't let TEEs break your MPC Threshold signature schemes, a form of multi-party computation (MPC) that lets a set of parties sign together without any one of them holding the key, are increasingly de...

• Keywords: mpc tees, mpc security, signatures tees, tees mpc, tee confidentiality, mpc vulnerabilities, tee trusted, tee attestation, mpc cryptographic, security mpc
• Source: blog.trailofbits.com

The Hidden Technical Debt of Prompt Hacks

-5517fd7b58a6---4 crawled_date: 2026-09-25T15:36:08.965346+00:00 feed_url: https://levelup.gitconnected.com/feed published: Fri, 25 Sep 2026 15:10:17 GMT

The Hidden Technical Debt of Prompt Hacks...

• Keywords: enterprise ai, catastrophic failure, semantic failure, architectural instability, failures cause, systems fail, ai agent, ai microsoft, cascading failures, agents rely
• Source: levelup.gitconnected.com

The most dangerous change to a published API is a rename

Hi Friends, Welcome to the 192nd issue of the Polymathic Engineer newsletter. This week's article continues our API design series. Last time, we focused on the core rules for creating intuitive and ea...

• Keywords: openapi specification, api design, api openapi, integrating apis, api needs, apis, implementation openapi, enforce openapi, rules openapi, api
• Source: newsletter.francofernando.com

Bedrock Agentcore Authentication and Authorization Deep Dive

-35e7a49c6df5---4 crawled_date: 2026-09-25T15:36:08.965346+00:00 feed_url: https://aws.plainenglish.io/feed published: Fri, 25 Sep 2026 14:38:09 GMT

Bedrock Agentcore Authentication and Authoriza...

• Keywords: bedrock_agentcore_authentication_and_authorization tree, bedrock_agentcore_authentication_and_authorization, vinycoolguy2015 bedrock_agentcore_authentication_and_authorization, agentcore authentication, bedrock agentcore, bedrock agent, authorization flow, authorization deep, authentication authorization, authorization mechanisms
• Source: aws.plainenglish.io

OCR as a Document, Not Just a String

-5517fd7b58a6---4 crawled_date: 2026-09-25T15:36:08.965346+00:00 feed_url: https://levelup.gitconnected.com/feed published: Fri, 25 Sep 2026 15:09:14 GMT

OCR as a Document, Not Just a String I ha...

• Keywords: epub reader, ocr document, reads epub, ocr library, resources ocr, build ocr, ocr researcher, pdf text, ocr directly, pdfs
• Source: levelup.gitconnected.com

[email protected]

[email protected] Patch Changes

e80b141: Only a reader opening an <AccordionItem> or<Expandable> writes its id to the URL hash. AdefaultOpen item on the page a client-side navigation lands on no longer re...

• Keywords: mcpget_page search, links agent, skills oragents, skills index, agent skills, mcpget_page, agent cli, txt agent, agent markdown, skill mcp
• Source: useblume.dev

Dependency Mocking Approach That Gets More Accurate as Your Services Deploy More Often

TL;DR — Key Takeaways

• Traditional dependency mocks can become less accurate as upstream services deploy more frequently, creating a growing gap between test behavior and production reality.

• Traffi...

• Keywords: dependency mocks, downstream mocks, deployment mock, determines mocks, mocks depend, dependency mocking, mocks affected, mocks observed, regenerating mocks, mocking primarily

• Source: devops.com

Model Weight Profiles: Where Do the Parameters Go?

Model Weight Profiles: Where Do the Parameters Go?# This blog describes model weight profiles, which show how many model parameters are devoted to embeddings, attention, dense layers, and other compon...

• Keywords: billions parameters, model size, 685b parameters, billion parameters, memory footprint, parameters footprint, model weight, model parameters, parameters weight, memory weights
• Source: rocm.blogs.amd.com

When to choose x86-64 vs aarch64

Not all cloud vCPUs are created equal. When you create a PlanetScale Postgres or Neki database, you have to choose between aarch64 (ARM) and x86-64 . Two clusters on different architectures can have t...

• Keywords: processors cloud, cloud vcpus, x86 clusters, x86 processor, vcpus memory, x86 instances, arm processor, arm processors, architectures vcpu, replicas cpu
• Source: planetscale.com

Daily Reading List – September 24, 2026 (#874)

When chatting with someone yesterday, I realized one reason this newsletter exercise helps me so much. We all notice a ton of information each day, but don’t always take a moment to imprint any of it...

• Keywords: agent skills, agents google, agents, agent, agent design, agent create, architecture agentic, ai collaborator, solution agent, build ai
• Source: seroter.com

Deploying Windows Containers on Amazon ECS (Step-by-Step Guide)

-35e7a49c6df5---4 crawled_date: 2026-09-25T15:36:08.965346+00:00 feed_url: https://aws.plainenglish.io/feed published: Fri, 25 Sep 2026 14:39:35 GMT

Deploying Windows Containers on Amazon ECS (St...

• Keywords: ecs_cluster windows, windows ec2, ecs windows, windows ecs, windows container, containers ecs, windows containers, scalable aws, ecs cloudformation, deployment ecs
• Source: aws.plainenglish.io

Jevmem – automatic project memory for Claude Code, built on Jev

Automatic project memory for Claude Code. Also works with Cursor and Codex. jevmem-launch-readme-v2.mp4

• Saves decisions, constraints, bugs and todos from your Claude Code chats into JEVMEM.md , auto...

• Keywords: sqlite locks, memory line, sqlite, store sqlite, add_memory goes, memory lines, add_memory, memories jevmem, saved jevmem, use sqlite

• Source: github.com

New in LangSmith: Engine v2, Managed Deep Agents, Fine-Tuning, and more

Key Takeaways

• Proactively detect agent issues. LangSmith Engine v2 introduces red teaming, expanded issue detection, and automated fix validation.

• Run agents with identity-scoped auth and memory....

• Keywords: agent tools, improving agents, agent development, agent automates, improve agents, agents need, agent memory, ai tools, built agent, running agents

• Source: langchain.com

Grieving the loss of details

I’ve been thinking about where I stand in respect to the current state of the industry. This is more of a journal note than a post, apologies for that. I usually leave this stuff private, but I though...

• Keywords: vibecoding coined, makes programming, programmer, programmers, programmers primarily, term vibecoding, vibecoding, programming, primarily designing, increasingly retrocomputing
• Source: purplesyringa.moe

Is Seedance Open Source?

Is Seedance Open Source? OpenRouter · We found no published weights and no model license for any Seedance model. ByteDance’s Seedance 2.5 launch post and model page describe access through Jimeng AI,...

• Keywords: seedance bytedance, bytedance seedance, bytedance seed, seedance repository, seedance hosted, seedance github, repository seedance, seedance release, seed repositories, source seedance
• Source: openrouter.ai

Object Storage branch reset, workload automation with Functions, and Realtime coming to Neon release - Sep 25, 2026

Reset a branch's Object Storage from its parent Reset from parent now rolls back a branch's Object Storage along with its Postgres data. When you reset a branch, its buckets and objects return to the...

• Keywords: lakebase search, updated lakebase, updated lakebase_vector, lakebase, soon lakebase, reset branch, includes lakebase, storage bucket, lakebase_vector lakebase_text, lakebase_vector
• Source: neon.com

Pencils Down, Eyes Open: A Rails Developer After Rails World

Pencils Down, Eyes Open: A Rails Developer After Rails World DHH’s Rails World keynote made me uncomfortable in a way I think is useful. I’ve spent more than a decade building Rails apps, and I’m not...

• Keywords: rails developer, developer rails, rails agent, agent draft, dhh rails, agents typing, better workflows, rails dhh, built rails, building rails
• Source: caio.ca

What is Chrome DevTools Protocol (CDP)?

Chrome DevTools Protocol (CDP) is a low-level protocol for inspecting and controlling Chromium-based browsers. It exposes browser capabilities such as JavaScript execution, DOM inspection, network mon...

• Keywords: chrome devtools, chrome protocol, chrome cdp, browser cdp, devtools protocol, browser infrastructure, commands chrome, browser automation, communicate browser, cdp client
• Source: browser-use.com

Shuu12121/Owl-ph2-base-len2048 — public verification state: pending

Shuu12121/Owl-ph2-base-len2048 Shuu12121/Owl-ph2-base-len2048 is an open-weights language model for masked-language modeling under the apache-2.0 license (602 MB), released Feb 16, 2026. Hosted open-w...

• Keywords: shuu12121 owl, metadata shuu12121, owl ph2, shuu12121 java, file sha, download shuu12121, shuu12121 python, model shuu12121, len2048 shuu12121, safetensors files
• Source: huggingbay.xyz

Tool Calling vs. Code Execution for AI Agents: Choosing the Right Action Primitive

In this article, you will learn what tool calling and code execution are as agent action primitives, how they differ mechanically, and when to choose one over the other. Topics we will cover include:...

• Keywords: tool calls, agent framework, aggregation tasks, calls tools, tools callable, tool calling, agent handles, tool use, agents choosing, execution agent
• Source: machinelearningmastery.com

Understanding NvPCRs in systemd v262

systemd answers TPM PCR scarcity with additional PCR-like registers allocated in the TPM’s NV memory, with an anchoring design that was reworked in v262. In this hands-on deep dive, we rebuild a syste...

• Keywords: systemd pcrs, tpm pcrs, systemd tpm2, pcrs systemd, secure systemd, tpm pcr, systemd pcr, pcrs tpm2, systemd security, tpm2 pcr
• Source: katexochen.aro.bz

What to Log Before Your Agent Fails in Production

-5517fd7b58a6---4 crawled_date: 2026-09-25T15:36:08.965346+00:00 feed_url: https://levelup.gitconnected.com/feed published: Fri, 25 Sep 2026 15:09:49 GMT

What to Log Before Your Agent Fails in Pr...

• Keywords: agent failures, agent logging, agent wait, agent fails, log agent, agent incidents, agent tool, agent loop, slowing agent, agent continue
• Source: levelup.gitconnected.com

File Notification Attacks: Side-Channel Leakage from the File-Notification System on Linux, Android, Windows, and macOS

Side-Channel Leakage from the File-Notification System on Linux, Android, Windows, and macOS. Accepted at The ACM Conference on Computer and Communications Security (CCS), November 15-19, 2026 — The H...

• Keywords: file notifications, file notification, files leaks, files leak, filesystem events, files leaked, notification attacks, filesystemwatcher, files watched, notifications files
• Source: inoti.fyi

Launching FreeBSD/EC2 desktop AMIs

Launching FreeBSD/EC2 desktop AMIs I'm excited to announce — oh don't worry, I'm not doing that again — a new feature for the FreeBSD/EC2 platform: Desktop AMIs.Most people think of FreeBSD as being a...

• Keywords: freebsd desktop, ec2 desktop, freebsd amis, freebsd ec2, ami freebsd, launching freebsd, desktop aws, ec2 platform, running freebsd, desktop amis
• Source: daemonology.net

Agent of the Day – September 25, 2026

Agent of the Day – September 25, 2026 Most security scanners are optimized for one thing: finding something to report. That incentive quietly rewards noise — every string concatenation near a shell ca...

• Keywords: scan vulnhunter, vulnhunter scan, agent scan, daily vulnhunter, vulnhunter interesting, vulnhunt methodology, vulnhunter runs, vulnhunter, makes vulnhunter, agent day
• Source: github.github.com

Best practices guide for customizing Gemini models via Reinforcement Learning (RL)

Best practices guide for customizing Gemini models via Reinforcement Learning (RL) Jiaqi Pan Senior Software Engineer, Google Kunal Jha Senior Product Manager, Google Reinforcement learning (RL) has b...

• Keywords: gemini reinforcement, reward gemini, gemini reward, gemini service, customizing gemini, gemini autorater, gemini models, gemini google, model reward, google reinforcement
• Source: cloud.google.com

CVE-2025-13032: Entering and Breaking the Avast Antivirus Sandbox Part 2

CVE-2025-13032: Entering and Breaking the Avast Antivirus Sandbox Part 2 This post is the second and final part of our Avast Antivirus research, detailing the full exploitation of CVE-2025-13032 on an...

• Keywords: vulnerability avast, breaking avast, avast kernel, article pool_overflow_exploitation_since_windows_10_19h1, pool_overflow_exploitation_since_windows_10_19h1, avast antivirus, ensure avast, final avast, discovered avast, actes pool_overflow_exploitation_since_windows_10_19h1
• Source: safateam.com

DuckDB and Hugging Face: Querying Datasets Directly

DuckDB and Hugging Face: Querying Datasets Directly TL;DR: Hugging Face hosts hundreds of thousands of datasets, and DuckDB can read them directly, exactly where they are and without downloading anyth...

• Keywords: datasets duckdb, data duckdb, using duckdb, parquet duckdb, downloading duckdb, face duckdb, duckdb guide, query duckdb, duckdb infers, duckdb reads
• Source: duckdb.org

Harness vs. Framework vs. SDK: How to Choose a Coding Agent’s Runtime

-5517fd7b58a6---4 crawled_date: 2026-09-25T14:35:48.056830+00:00 feed_url: https://levelup.gitconnected.com/feed published: Fri, 25 Sep 2026 14:28:43 GMT

Member-only story Harness vs. Framework v...

• Keywords: agent runtime, agent project, agent framework, coding agents, coding agent, agent architecture, new agent, agent, agents, import agent
• Source: levelup.gitconnected.com

UltraQuant on AMD Instinct: More Efficient Agentic Serving for Qwen3.8-MXFP4

UltraQuant on AMD Instinct: More Efficient Agentic Serving for Qwen3.8-MXFP4# UltraQuant can raise decode throughput and cut latency under load without a measurable drop in reasoning accuracy. In this...

• Keywords: cache ultraquant, cache agentic, concurrency ultraquant, runs ultraquant, cache throughput, gpus qwen3, replay ultraquant, ultraquant requests, kv cache, latency ultraquant
• Source: rocm.blogs.amd.com

Redox OS Adds IO_uring-Like API, NUMA & Gets QEMU Working

Redox OS Adds IO_uring-Like API, NUMA & Gets QEMU Working The Redox OS open-source, Rust-based operating system has seen a lot of activity in recent weeks with many notable improvements landing. Redox...

• Keywords: redoxos qemu, qemu works, qemu usage, qemu emulator, redox os, os redox, qemu working, arm64 redox, qemu, redox currently
• Source: phoronix.com

Building Production Agents with Jev and LangGraph

Last week, TypeSafe AI released Jev, a new kind of model. Unlike traditional LLMs, Jev doesn't generate text. It makes decisions your code can act on directly, for what TypeSafe calls AI-powered softw...

• Keywords: evaluations jev, typesafe ai, jev decisions, decisions jev, jev handling, jev designed, building ai, god jev, like jev, ai production
• Source: langchain.com

Improving site performance by shipping more CSS

Josh Black Josh Black is a Software Engineer based in Austin, Texas. He loves working on Design Systems, creating accessible experiences, and eating chilaquiles. A look at how a design system shipped...

• Keywords: primer css, css primer, migrating css, css migration, benefits css, support css, native css, css modules, primer styled, primer design
• Source: github.blog

Kamalu211/llama-3-8b-Instruct-kamal-treeofthoughts-game24 — public verification state: pending

Kamalu211/llama-3-8b-Instruct-kamal-treeofthoughts-game24 Kamalu211/llama-3-8b-Instruct-kamal-treeofthoughts-game24 is an open-weights language model for general under the apache-2.0 license (4.9 GB),...

• Keywords: download kamalu211, file kamalu211, game24 kamalu211, install huggingbay, model kamalu211, check download, kamalu211 llama, kamalu211 released, treeofthoughts game24, game24 files
• Source: huggingbay.xyz

Migrating a JSF 2 application to Vaadin 25 with AI agents

I spent more than ten years writing JSF. JSF 1.2 first, then years of JSF 2 and PrimeFaces in a corporate environment, and teaching other people to do the same. We wrote a lot of our own components. T...

• Keywords: application jsf, jsf application, jsf jsf, jsf primefaces, jsf, writing jsf, original jsf, assess jsf, form jsf, jsf migration
• Source: blog.vaadin.com

OpenSearch Kubernetes Operator 3.0 is now generally available

In February, the 3.0 alpha of the OpenSearch Kubernetes Operator was announced and the community was asked to try it in lower environments. Many did. Today, 3.0 is generally available and is the recom...

• Keywords: opensearch kubernetes, kubernetes manages, kubernetes kubernetes, kubernetes cluster, driven kubernetes, kubernetes, pod readiness, kubernetes node, kubernetes operator, real kubernetes
• Source: opensearch.org

Real-Time Telemetry with Eventlog Live

Well-Typed are happy to announce Eventlog Live, a program that streams real-time telemetry from any Haskell application to any observability platform that supports the OpenTelemetry protocol, such as...

• Keywords: ghc_eventlog_heaplive metric, ghc_eventlog_heaplive, values ghc_eventlog_heaplive, named ghc_eventlog_heaplive, ghc_eventlog_heaplive selected, live eventlog, built eventlog, eventlog live, recording heap, samples eventlog
• Source: well-typed.com

Getting Started with Local MCP Servers on Claude Desktop

The Model Context Protocol (MCP) is an open protocol that enables seamless integration between LLM applications and external data sources and tools. With the introduction of desktop extensions, instal...

• Keywords: mcp clients, mcp server, llm applications, extensions desktop, local mcp, mcp integrations, desktop extensions, desktop extension, mcp servers, plugin remote
• Source: support.claude.com

What we learned running seven AI coding models on real production work

What we learned running seven AI coding models on real production work Seven AI models read the same ticket and reached the same answer. They were all wrong. The ticket said add a 30-second timeout, s...

• Keywords: timeout example, timeout obvious, timeout engineer, story timeout, timeout type, timeout, timeout wrote, timeout fired, size timeout, process catch
• Source: hostinger.com

Who Is Open Source About?

Open source is, at least in part, about you, where “you” refers to the user. Open Source Is Not About “Open Source Is Not About You” In other words: Rich Hickey was wrong when he wrote “Open Source Is...

• Keywords: open source, freeloading stakeholders, source open, maintainers free, free software, free support, benefit users, openness, external sources, freeloading
• Source: blog.glyph.im

A Type Stronger than the Sum of its Components

A Type Stronger than the Sum of its Components 24 Sep 2026Have you ever written a type that you appreciated so much you still think about it? Like eating a really good meal, where if you try hard enou...

• Keywords: component types, component enum, enum variants, structs component, components type, type variants, paths interface, struct normalcomponent, normalcomponent struct, type useful
• Source: schneems.com

Qualcomm Posts Patches For Enabling GPU Support In Their Snapdragon X2 Mahua SoC

Qualcomm Posts Patches For Enabling GPU Support In Their Snapdragon X2 Mahua SoC Yesterday Qualcomm posted open-source GPU driver patches for their Adreno 850 GPU one day after being announced with th...

• Keywords: mahua gpu, gpu patches, patches qualcomm, support snapdragon, gpu support, qualcomm announced, enabling gpu, source gpu, snapdragon x2, gpu enablement
• Source: phoronix.com

Socket Joins New OpenJS Program to Fund Node.js Security Work

Socket Joins New OpenJS Program to Fund Node.js Security Work Socket is joining the OpenJS Security Stewardship Program to fund Node.js vulnerability research, maintainer remediation, and security rel...

• Keywords: js vulnerability, javascript infrastructure, js ecosystem, socket ceo, node js, javascript ecosystem, openjs security, according openjs, new openjs, js threat
• Source: socket.dev

Cross-Site Scripting (XSS) Cheat Sheet

Cross-site scripting (XSS) cheat sheet This cross-site scripting (XSS) cheat sheet contains many vectors that can help you bypass WAFs and filters. You can select vectors by the event, tag or browser...

• Keywords: scripting xss, script xss, xss compatibility, xss onmousewheel, xss onkeypress, xss onwebkitwillrevealbottom, frames xss, xss javascript, xss vbscript, xss onkeydown
• Source: portswigger.net

LAST CALL FOR ENROLLMENT: Rebuild YouTube with AI

Rebuild YouTube with AI, taught by a former YouTube engineer, kicks off on Saturday, September 26. Enrollment closes in 24 hours. What you will learn Scope a realistic MVP. Decide which YouTube featur...

• Keywords: youtube ai, youtube engineer, taught youtube, decide youtube, rebuild youtube, salesforce ai, youtube, ai engineering, ai development, youtube features
• Source: blog.bytebytego.com

Plan mode is dead

Plan mode is dead Earlier this year, I believed planning was going to become the most important part of building software with AI. My belief in this idea was so strong that I built and launched an ent...

• Keywords: plan modes, plan mode, planning plan, planning interface, understanding plan, planning execution, plan models, implement plan, abstraction planning, plan implementation
• Source: aymannadeem.com

AI is not “just a tool”

AI is not “just a tool” I am sick and tired of people saying that “AI is just a tool” because it is quite simply wrong. Framing the technology of LLMs as “just a tool” is meaningless and justifies not...

• Keywords: tool ai, ai industry, ai products, ai just, ai broader, ai product, ai meant, llms ai, saying ai, ai companies
• Source: brettcodes.com

Approaching A 10 Second Linux Kernel Build

Approaching A 10 Second Linux Kernel Build The coffee window is closing. Going back many years with the time it's taken to compile the Linux kernel has been a well known opportunity for a coffee break...

• Keywords: speeding kernel, kernel build, kernel builds, kernel building, faster linux, build time, speedy builds, build times, optimizing kernel, build linux
• Source: phoronix.com

How to add hybrid search to a Postgres app with Pinecone in 2026

How can you add hybrid search to Postgres with Pinecone? Use Pinecone for vector search, with Postgres as a system of record. Hybrid search with semantic and full-text search capabilities can be imple...

• Keywords: search hybrid, hybrid search, searchable store, pinecone postgres, postgres pinecone, search postgres, search inventory, snacks postgres, search semantic, search combining
• Source: pinecone.io

Meta's Muse appears to use an OpenAI model labeled muse-special

Is Meta’s Muse secretly running an OpenAI model? I found a model labeled azure/muse-special while Muse was building my website. So I dug deeper. This is Part 2 of digging through the Muse filesystem a...

• Keywords: azure muse, muse agent, muse secretly, api muse, called muse, muse actually, does muse, model muse, meta muse, muse
• Source: mouse.dev

mradermacher/Cyber-Ornith-1.5-9B-OBLITERATED-GGUF — public verification state: pending

mradermacher/Cyber-Ornith-1.5-9B-OBLITERATED-GGUF mradermacher/Cyber-Ornith-1.5-9B-OBLITERATED-GGUF is an open-weights language model for general under the apache-2.0 license (81.4 GB), released Sep 2...

• Keywords: download mradermacher, install huggingbay, check download, downloads explicit, files mradermacher, released apache, q8_0 gguf, download completeness, download works, unknown download
• Source: huggingbay.xyz

NVIDIA Files Patent for AI Tool to Help Developers Diagnose GPU Performance Issues

NVIDIA has filed a patent for an AI tool that could help developers diagnose and optimize GPU performance issues through a plain-language chat interface. The <a href="https://eureka.patsnap.com/patent...

• Keywords: nvidia ai, gpu profiling, nvidia project, nvidia planning, gpu, general gpu, gpu performance, nvidia, gpu optimization, optimize gpu
• Source: techpowerup.com

Security Slam 2026 – Fall edition

Security Slam 2026 – Fall Edition is a 30-day virtual event from October 5 through November 6, 2026. What Is the Security Slam? The Open Source Security Foundation (OpenSSF) is partnering with the Clo...

• Keywords: security slam, security milestones, openssf projects, openssf project, security foundation, 2026 security, award openssf, foundation openssf, openssf booth, openssf
• Source: cncf.io

Why Old Azure DevOps Releases Survive a Pipeline Cutover

TL;DR — Key Takeaways

• Azure DevOps Classic releases preserve the deployment tasks, Helm release names, image paths and variable values captured when the release was created.

• Updating a release def...

• Keywords: preserve deployment, release kubernetes, releases preserve, end deployments, devops migration, orphaned kubernetes, deployment tasks, release migration, stale deployment, devops deploy

• Source: devops.com

AI Game Show Host: Built With MeetStream

AI Game Show Host: Built With MeetStream "Good evening, and welcome! Let's begin. Question 1, worth 100 points: What is 7 plus 8?" That's the first thing you hear when our bot joins your Google Meet....

• Keywords: meetstream gameshow, game meeting, bot meeting, bots meetings, meeting bot, meeting game, chat play, chat meetstream, gameshow host, meetstream ai
• Source: meetstream.ai

Pentium II at 600Mhz with Voodoo 3 Emulated on 86Box with M6 Mac Mini

Mac Mini M6: Retro PC Emulation with 86Box (600MHz PII?!) The best platform for cycle-accurate retro PC emulation may just be the Mac. The M6 achieved a stable 600MHz clock using a 6.0 derived 86Box b...

• Keywords: emulation performance, emulation speed, core 86box, mac m6, m6 cores, cpu emulation, cores 86box, 500mhz macs, speed mac, emulated cpu
• Source: nyaa.sh

ThursdAI - Sep 24 - Opus 5.5 beats Fable for 40% less, OpenAI halves GPT-6 prices & more

Hey, it’s Alex 👋 What a freaking week! A week after every lab head agreed to “pace the frontier”, Anthropic and OpenAI shipped big new models within hours of each other, and both of them are CHEAPER....

• Keywords: pacing frontier, pace frontier, muse partners, connect muse, end frontier, frontier ai, frontier, muse connect, opus days, week audio
• Source: sub.thursdai.news

Jev and System One Models: Calibration Beats Accuracy

Jev and System One Models: Calibration Beats Accuracy An ML engineer's read on TypeSafe AI's Jev: what a non-autoregressive System One model changes for production classifiers, where it fits, and how...

• Keywords: jev models, models jev, classifier jev, jev marketing, marketing jev, jev reading, does jev, models calibration, features jev, jev task
• Source: kartikpansuriya.com

Muse Looks Cute, but Looks Are Deceiving

<p>Jason Aten, in a good follow-up to his previous column at Inc. (<a href="https://daringfireball.net/linked/2026/09/22/aten-muse">the one</a> where he described how Muse, running on his Mac, read hi...

• Keywords: chatbot, chatbot actually, virtual machines, muse interesting, machines permission, read messages, just chatbot, ai agent, millions devices, permission architectures
• Source: inc.com

Three Lenses on Coordination

Three Sets of Specs (for Staying in Spec) In the last post I talked about how systems formalisms are like sculpting with a chisel: we remove behaviors we don’t want, and what remains is correct. This...

• Keywords: staying specification, reconciliation deadlocks, specs staying, spec waiting, staying spec, systems formalisms, deadlock fundamental, specification, spec actively, unavailability deadlock
• Source: jhellerstein.github.io

Trading a Cloud Identity for Your Own: Workload Attestation on Managed Compute

-2615bd06b42e---4 crawled_date: 2026-09-25T16:36:09.368416+00:00 feed_url: https://netflixtechblog.com/feed published: Fri, 25 Sep 2026 16:01:03 GMT

Trading a Cloud Identity for Your Own: Workloa...

• Keywords: cloud identity, aws identity, identity managed, identity workload, identity metadata, cloud provider, identity provider, cloud credentials, identity service, managed compute
• Source: netflixtechblog.com

Discord Update: September 25, 2026

Here are the Discord updates from September 25, 2026. You can also find the most recent updates in Discord under Settings > More > What's New. We’ve put in system wide echo cancellation on Windows cli...

• Keywords: revamped discord, discord updates, updates discord, audio discord, discord patch, discord settings, discord town, discord calls, discord place, discord discord
• Source: discord.com

Every package is already installed

tl;dr; omnibin is a FUSE filesystem that puts every binary nixpkgs ever shipped on your $PATH . Nothing is installed. Nothing needs building. 0 bytes on disk until something actually reads a file. 😈 I...

• Keywords: omnibin nixosmodules, binary nixpkgs, single nixpkgs, built nixpkgs, nixpkgs available, version nixpkgs, nixpkgs, nixosmodules default, omnibin bin, nixosmodules
• Source: fzakaria.com

Back to the 90s: The Software Engineer Has All the Roles Again

When I started as a software engineer in the 1990s, I programmed COBOL on the IBM mainframe. Later, client/server applications came along. Many things were simpler back then. Not the technology, but t...

• Keywords: cobol ibm, architecture developers, cobol designed, customer architect, agile, functional teams, programmed cobol, idea cobol, systems teams, cobol programs
• Source: martinelli.ch

LLM Policies: Progress At All Costs

LLM Policies: Progress At All Costs GNOME and KDE have started to consider LLM policies, and we should talk about what this is really about. KDE caught everyone's attention first by igniting a flame w...

• Keywords: llm policies, llm productivity, llm contributions, ban llm, gnome projects, llm friendly, gnome proposed, war llm, externalities llms, llm simply
• Source: diegoe.be

The Efficiency-Throughput Gap with GitHub Copilot

The rapid adoption of AI-assisted tools in software development has rekindled interest in engineering productivity. However, a holistic understanding of these tools’ real-world impact on engineering p...

• Keywords: software productivity, engineering productivity, ai productivity, developer efficiency, productivity github, productivity, productivity improvements, productivity suggesting, productivity throughput, driven productivity
• Source: cacm.acm.org

Build golden path CI/CD workflows in Red Hat Developer Hub

In Build a unified CI/CD control plane with Red Hat Developer Hub, we set up a Red Hat Developer Hub instance on OpenShift, configured secrets, enabled dynamic plug-ins for Tekton, Argo CD, Quay, and...

• Keywords: developer hub, hub developer, openshift dev, scaffolding developers, developers create, path platform, workflows platform, application scaffolding, software template, rhdh kubernetes
• Source: developers.redhat.com

Dutch governments builds alternative for Microsoft based on NixOS

What DAWO stands for DAWO brings public values and technology together. The community pursues five primary goals. Digital autonomy Strengthening the digital autonomy of the Netherlands. Collaboration...

• Keywords: community dawo, digital workplace, dawo, dawo stands, stands dawo, strengthening digital, workplace cloud, open workplace, digital, dawo nixos
• Source: dawo.community

On-demand compute for scheduled Copy Pipes

Scheduled Copy Pipes can now run on on-demand compute. Add ON_DEMAND_COMPUTE true to a Copy Pipe datafile and each scheduled run uses temporary, dedicated ClickHouse® compute instead of the primary co...

• Keywords: api_usage_hourly copy_schedule, compute tinybird, tinybird workspace, node hourly_api_usage, copy_schedule copy_mode, compute schedule, copy_schedule, target_datasource api_usage_hourly, scheduled copy, hourly_api_usage sql
• Source: tinybird.co

Push images to Vercel Container Registry from GitHub Actions

You can now push container images from GitHub Actions to Vercel Container Registry (VCR) without storing long-lived registry credentials. The new vercel/vcr-action/login action authenticates your work...

• Keywords: vercel vcr, vcr vercel, authenticates workflow, vcr github, vercel team, images github, github actions, vercel_team_id build, vercel container, vercel token
• Source: vercel.com

CommonGrid: The open source registry of the US power grid

The open, connected registry of the U.S. power grid Built and maintained by a community of users who know every corner of the grid, CommonGrid connects energy infrastructure data into a public, sharea...

• Keywords: energy infrastructure, power grid, voltage infrastructure, connected utilities, utilities territories, power agencies, utilities state, power plants, infrastructure data, power plant
• Source: commongrid.info

Jev Based Code Review

Most PRs that are generated by agents today get YOLO merged because its hard for human mind to comprehend when your agent just suddenly shows up with 230 file changes. This is an attempt to reduce the...

• Keywords: jev reviewer, reviewer git, pr bundled, jev code, pr repository, pr files, reviewer npm, agent prs, prs generated, github watch
• Source: github.com

Proaction boosts sales 60% and saves 75+ hours with Codex

Proaction boosts sales 60% and saves 75+ hours with Codex With Codex, GPT‑Live‑1, and GPT‑6 Astra, Proaction builds, operates, and sells modern fleet management faster. Results 40–60 Engineering hours...

• Keywords: proaction building, proaction builds, proaction build, proaction built, proaction engineers, proaction time, demos proaction, proaction platform, sales proaction, proaction uses
• Source: openai.com

Special Projects (2016)

Special projects Impactful scientific work requires working on the right problems—problems which are not just interesting, but whose solutions matter. Impactful scientific work requires working on the...

• Keywords: ai breakthrough, breakthrough ai, ai research, ai development, consequences ai, advancing ai, use ai, ai break, ai world, ai
• Source: openai.com

Boards of Casio

While working away on the CZP-1, which was at least initially an effort to create a Casio CZ-101 on a web page I happened to see this YouTube video by oliveoil22: The immediate question was “How can I...

• Keywords: czp audio, czp saves, loading czp, czp initially, czp manual, shared cz, czp doesn, czp tabs, note czp, czp
• Source: ambionix.com

The Merge: Your tests pass. Can your agent finish the job?

In CodeRabbit’s latest episode of The Merge, I sat down with Francesco Bonacci, CEO and co-founder of Cua AI, which builds tools and environments that let AI agents operate computers. We talked about...

• Keywords: ai agents, cua tested, tests computer, agent tools, test software, ai builds, agents operate, cua technical, software doing, software examines
• Source: coderabbit.ai

Using an LLM to Automate the Process of Archiving New macOS App Icons

Using an LLM to Automate the Process of Archiving New macOS App Icons Whenever a new version of macOS drops, I find myself archiving the new app icons for my macOS icon gallery. In the past, the work...

• Keywords: application icons, app icons, icons macos, icons extracted, icon file, app icon, store icon, icon gallery, icon copied, apps artwork
• Source: blog.jim-nielsen.com

AI broke code review. What about knowledge sharing?

AI broke code review. What about knowledge sharing? The volume of AI-generated code has broken code review. Engineers are reviewing 500-line diffs they didn’t write, generated by models they don’t ful...

• Keywords: knowledge sharing, review knowledge, sharing review, instruct review, code review, reviewed ai, review agent, code practices, code supervising, review engineers
• Source: aviator.co

Configuring S3 Backup and Recovery with Clumio

Key Takeaways

• Amazon S3 durability helps protect against infrastructure failures, but additional backup protection is needed for accidental deletion, misconfiguration, compromised credentials, and d...

• Keywords: s3 durability, durability s3, backup s3, s3 protection, protects s3, securevault backup, protect s3, restore s3, rollback s3, backup protection

• Source: commvault.com

Figma to Code Before LLMs

Jev launched last week, and people are already sorting inboxes and even playing Doom with it. I recognized the pattern: a model making structured decisions inside a larger algorithm. When I built the...

• Keywords: figma design, figma designs, structures figma, layout figma, figma tree, figma nodes, trees figma, tree responsive, figma web, tree models
• Source: manualmeida.dev

What I learned at DevTalks Cluj-Napoca 2026

What I learned at DevTalks Cluj-Napoca 2026 Are you struggling with performance issues in your Spring, Jakarta EE, or Java EE application? Imagine having a tool that could automatically detect perform...

• Keywords: allowing performance, prevent performance, performance issues, java ee, ee java, future java, participants devtalks, java developers, performance, java applications
• Source: vladmihalcea.com

Vercel Sandbox now supports memory observability

Vercel Sandbox observability now includes memory usage data. You can access sandbox memory usage data in the dashboard and through the CLI via the vercel metrics command. Copy link to headingSandbox o...

• Keywords: sandbox memory, sandbox memory_used_bytes, memoryusedbytes measure, memory sandboxes, memory dashboard, memory usage, vercel sandbox, memory_used_bytes run, memory_used_bytes, measure sandbox
• Source: vercel.com