Daily Tech News - 2026-07-17

The tech world buzzed today with significant advancements in artificial intelligence, cloud-native infrastructure, and developer tooling. AI innovation continues to accelerate, with major players like Netflix detailing their sophisticated in-house LLM serving strategies. The ecosystem for AI development is also maturing rapidly, highlighted by LiteLLM's feature-rich AI Gateway for LLMs and new tools enabling the fine-tuning of video and image models at scale with NVIDIA NeMo Automodel and Hugging Face Diffusers.

The concept of AI agents gained further traction, with discussions ranging from building durable agents on platforms like Supabase to understanding the crucial 'agent harness' layer that ensures their practical functionality. FastAPI emerged as a key technology for serving these agents as APIs, addressing a critical gap in AI engineering. Noteworthy innovations included a local-first agent running offline on NVIDIA DGX Spark, a 3D foundation model (Lingbot-map) for reconstructing scenes from streaming data, and new demos on Google's Gemini Enterprise Agent Platform. Even browser security is getting an AI boost, with AI-assisted hacking for Chrome making headlines. The shift from simple chatbots to robust AI workflows for incident investigation also marks a significant evolution in AI application, alongside new principles for 'AI Tokenomics' to optimize software engineering.

In cloud and infrastructure, the KubeCon + CloudNativeCon India 2026 recap showcased the vibrant cloud-native community, while discussions on software architecture in OutSystems ODC provided insights into modern development paradigms. Tools for secure and collaborative work saw an update with migration options to ngrok from Cloudflare Tunnels, and performance benchmarking for Red Hat Data Grid on OpenShift 4 was detailed. On the developer front, the first Linux Assembly X server, Frame, pushed the boundaries of operating system components. Homebrew enhanced its security by plumbing into the vulnerability ecosystem, and new techniques for exporting PDFs with embedded annotations were unveiled, demonstrating ongoing refinements in core software utilities. Overall, the day painted a picture of a rapidly evolving technological landscape, driven heavily by practical applications of AI. The tech world buzzed today with significant advancements in artificial intelligence, cloud-native infrastructure, and developer tooling. AI innovation continues to accelerate, with major players like Netflix detailing their sophisticated in-house LLM serving strategies. The ecosystem for AI development is also maturing rapidly, highlighted by LiteLLM's feature-rich AI Gateway for LLMs and new tools enabling the fine-tuning of video and image models at scale with NVIDIA NeMo Automodel and Hugging Face Diffusers.

Featured Articles

In-House LLM Serving at Netflix

-2615bd06b42e---4 crawled_date: 2026-07-17T22:33:34.932451+00:00 feed_url: https://netflixtechblog.com/feed published: Fri, 17 Jul 2026 21:32:39 GMT

In-House LLM Serving at Netflix By AI Platform...

Keywords: apis netflix, ml netflix, netflix technology, netflix production, scale netflix, netflix ai, serving netflix, netflixtechblog, netflix fronted, changes netflix
Source: netflixtechblog.com

LiteLLM: AI Gateway for LLMs — Features Overview

-5b301f10ddcd---4 crawled_date: 2026-07-17T08:29:41.472470+00:00 feed_url: https://itnext.io/feed published: Fri, 17 Jul 2026 08:30:24 GMT

LiteLLM: AI Gateway for LLMs — Features Overview In the...

Keywords: kubernetes traces, llm monitoring, monitoring llm, kubernetes metrics, tracing llms, llms observability, gateway kubernetes, kubernetes integration, tracing observability, opentelemetry tracing
Source: itnext.io

Fine-tune video and image models at scale with NVIDIA NeMo Automodel and 🤗 Diffusers

Fine-tune video and image models at scale with NVIDIA NeMo Automodel and 🤗 Diffusers Diffusion models power some of the most exciting open-source releases of the last two years — such as FLUX.1-dev fo...

Keywords: automodel diffusers, nvidia nemo, collaboration nvidia, nemo framework, nvidia hugging, library nvidia, video diffusers, models nemo, image models, model support
Source: huggingface.co

Running goose fully offline on a DGX Spark

A hands-on look at running a local-first agent on the NVIDIA DGX Spark: zero cloud calls, on a 128 GB box that fits on a desk. Almost every agent demo you see talks to a frontier model over a cloud AP...

Keywords: agent nvidia, agent framework, agent spark, cloud calls, architecture spark, local agent, cloud machine, hosted agentic, zero cloud, exactly agent
Source: aaif.io

FastAPI for Agents

FastAPI for Agents FastAPI solved one of the biggest problems in software engineering: how to serve code as an API. That same problem is now the biggest gap in AI engineering: how to serve agents as a...

Keywords: fastapi agents, agents fastapi, runtime agents, api agentos, agents api, agentos fastapi, agent infrastructure, fastapi dev, agentos runtime, agentic workflows
Source: ashpreetbedi.com

Flipkart and LitmusChaos at KubeCon + CloudNativeCon India 2026: A recap

KubeCon + CloudNativeCon India 2026 brought the cloud native community to Mumbai on June 18-19. For LitmusChaos, this was not just another conference. It was one of our most significant events to date...

Keywords: keynote kubecon, flipkart litmuschaos, kubecon cloudnativecon, community flipkart, contest keynote, flipkart story, india flipkart, flipkart main, winning keynote, practice flipkart
Source: cncf.io

Journey to Root, Episode I: The Maglev King

Journey to Root, Episode I: The Maglev King Hacking Chrome with AI Table of Contents Introduction Welcome to the first installment of our series on AI-assisted browser bug hunting and exploit developm...

Keywords: chrome exploit, exploit chrome, chrome vulnerability, hacking chrome, exploitable bugs, finished exploits, reliable exploit, v8 exploits, vulnerability gpu, exploit google
Source: blog.calif.io

Lingbot-map: A 3D foundation model for reconstructing scenes from streaming data

teaser.mp4 🗺️ Meet LingBot-Map! We've built a feed-forward 3D foundation model for streaming 3D reconstruction! 🏗️🌍 LingBot-Map has focused on:

Geometric Context Transformer: Architecturally unifies...
Keywords: streaming 3d, flythrough mp4, streaming iterative, efficient streaming, efficiency streaming, benchmark pipeline, map accelerated, streaming framework, spires benchmark, streaming inference
Source: github.com

Durable workflows — and durable AI agents — on Supabase alone

Durable workflows — and durable AI agents — on Supabase alone One SQL file turns a stock Supabase project into a durable execution engine. We set it up, benchmarked it, crashed it on purpose, and ran...

Keywords: pg_cron durable, durable workflows, workflows durable, durable workflow, supabase durable, supabase runtime, pg_net supabase, durable execution, supabase projects, postgres runs
Source: journal.resonatehq.io

What Is an Agent Harness? The Layer That Makes AI Agents Actually Work

If you’ve shipped an LLM-powered feature beyond a simple chat interface, you’ve already built parts of an agent harness. The retry logic around a failed tool call, the code assembling a system prompt...

Keywords: agent harness, harness agent, harness engineering, harnesses execution, execution harness, command harness, management harness, harness manages, engineering harness, harnesses models
Source: live-comet-marketing-site.pantheonsite.io

Basic concepts of Software architecture in OutSystems ODC with comparisons to OutSystems O11

-5b301f10ddcd---4 crawled_date: 2026-07-17T15:31:36.576993+00:00 feed_url: https://itnext.io/feed published: Fri, 17 Jul 2026 14:59:08 GMT

Basic concepts of Software architecture in OutSystems OD...

Keywords: odc applications, odc architecture, architecture odc, outsystems odc, outsystems o11, odc application, applications kubernetes, o11 outsystems, odc version, outsystems application
Source: itnext.io

Migrate to ngrok from Cloudflare Tunnels

Jul 17, 2026 Latest PostJul 17, 2026 Latest PostSoftware teams need to work quickly, securely, and collaboratively. They need to focus on building solutions that make their products unique and worry a...

Keywords: ngrok infrastructure, ngrok cloudflare, services ngrok, ngrok application, using ngrok, utilize ngrok, manage ngrok, use ngrok, ngrok platform, ngrok tunnel
Source: ngrok.com

“Apple’s AI Is Free on Linux”: A Weekend of Install Logs and One Honest Caveat.

-a648dc4ecb66---4 crawled_date: 2026-07-17T12:30:35.985675+00:00 feed_url: https://towardsdev.com/feed published: Fri, 17 Jul 2026 12:02:34 GMT

Member-only story “Apple’s AI Is Free on Linux”: A...

Keywords: mlx apple, worked gpu, gpus actually, gpus, apple ai, nvidia gpus, runs mlx, mlx runs, runs nvidia, gpu
Source: towardsdev.com

We're sunsetting Tinybird Classic for Free and Developer workspaces

We're sunsetting Tinybird Classic for Free and Developer plan workspaces. If you have an eligible Classic workspace, you need to migrate it to Tinybird Forward by September 15, 2026. Forward is now th...

Keywords: migrate tinybird, tinybird enterprise, tinybird runtime, run tinybird, apply tinybird, deploy tinybird, support tinybird, migrating early, using tinybird, migrating deadline
Source: tinybird.co

13 hands-on demos to build on Gemini Enterprise Agent Platform

13 hands-on demos to build on Gemini Enterprise Agent Platform Shubham Saboo Senior AI Product Manager, Google Cloud AI Earlier this year, we introduced Gemini Enterprise Agent Platform, where you can...

Keywords: build agents, agent build, build agent, agent codelab, agent development, agent platform, agents prototype, ai agents, build ai, runtime agents
Source: cloud.google.com

Deploy Multichain Dapps on Hedera in 60 Seconds with scaffold-hbar

TL;DR. One command spins up a working Hedera dapp with Next.js, a choice of Hardhat or Foundry, and AI agent context already wired in. Eight built-in templates at launch with more shipping through the...

Keywords: scaffold hbar, template hedera, dev scaffold, scaffolding cli, scaffolds projects, tools scaffold, hedera builders, built scaffold, contracts scaffolds, hedera built
Source: hedera.com

Frame – the first Linux Assembly X server

On my quest to own my software, one foundational piece kept itching… the X server. The underlying graphics engine, the thing that puts pixels on the screen. X11 is 4 million lines of code, a beast ver...

Keywords: frame xorg, screen x11, xorg, xorg burns, gdm x11, itching server, xorg pull, x11 i3, x11, runs desktop
Source: isene.org

Plumbing Homebrew into the vulnerability ecosystem

Back in January I released brew-vulns, a gem that added CVE scanning to Homebrew. It shelled out to brew info --json=v2 , derived a source repository URL and version tag for each installed formula, an...

Keywords: cve homebrew, brew vulns, homebrew vulns, homebrew versions, homebrew package, homebrew advisory, version homebrew, releases brew, repository homebrew, gem homebrew
Source: nesbitt.io

A Road to Lisp: Which Lisp

Most programming languages evolve as a single language. Python, Java, Javascript, C++, have new versions and standards, multiple implementations, but they still remain the same language. C and C++ can...

Keywords: learn lisp, lisp language, lisp beginners, choose lisp, lisp dialects, lisp programmers, lisp development, learning lisp, lisp dialect, compared lisp
Source: scotto.me

Beyond Chat: Why AI Workflows Beat Chatbots for Incident Investigation

OpsWorker CTO Nune Isabekyan spoke at WeAreDevelopers World Congress 2026 in Berlin. This is the argument she made - and the live demo that proved it. Watch the full talk: https://youtu.be/9SwVKRBfG2I...

Keywords: opsworker ai, immediately opsworker, frontend kafka, systems opsworker, kafka pipeline, opsworker cto, opsworker structured, kafka consumer, production communicate, opsworker
Source: opsworker.ai

Claude Code: Anatomy of a Misfeature

Claude Code: Anatomy of a Misfeature contents "Mechanical egg timer" by Hustvedt is licensed under CC BY-SA 3.0 . Padded to a wider frame; this adaptation is likewise licensed CC BY-SA 3.0. On Canada...

Keywords: claude_afk_timeout_ms void, claude_afk_timeout_ms claim, egg timer, claude_afk_timeout_ms appears, timeout control, setting claude_afk_timeout_ms, claude_afk_timeout_ms milliseconds, claude_afk_timeout_ms message, claude_afk_timeout_ms claude_afk_countdown_ms, claude_afk_timeout_ms
Source: olafalders.com

Export PDFs with embedded annotations via PDF.js and annotpdf

Export PDFs with embedded annotations via PDF.js and annotpdf Table of contents annotpdf library. This tutorial covers mapping app annotation types to the PDF specification, color conversion, and trig...

Keywords: pdf annotations, need pdfjs, js annotpdf, annotated pdf, pdf annotation, downloadjs annotations, annotpdf annotationfactory, pdfjs, annotations pdf, annotpdf anddownloadjs
Source: nutrient.io

Benchmark Red Hat Data Grid in OpenShift 4 using Hyperfoil

This article covers benchmarking with the Red Hat Data Grid operator and Helm charts. The container-awareness concept is the foundation here, as the Data Grid heap size is 50% of the container size. T...

Keywords: hyperfoil benchmarking, hyperfoil benchmark, benchmarking framework, benchmark hyperfoil, framework benchmarking, benchmark configuration, benchmarking performance, running benchmark, benchmarking tool, benchmark deployment
Source: developers.redhat.com

Put OPA in Front of Your Quarkus MCP Tools

Put OPA in Front of Your Quarkus MCP Tools Use structured policy input and caller context to decide which MCP tools a client can see and call. opa_decision={"allow":false,"skillId":"pptx_export","reas...

Keywords: quarkus enforcement, enforce policyversion, rego execution, rego policy, inspection regex, string policyversion, quarkus opa, opa quarkus, quarkus securityidentity, opa policy
Source: the-main-thread.com

Guide to AI Tokenomics: Eleven Principles for Token Efficient Software Engineering

Guide to AI Tokenomics: Eleven Principles for Token Efficient Software Engineering Alex "Sandu" Astrum Developer Relations, Antigravity Luke Schlangen Developer Advocate, Google Cloud Optimizing token...

Keywords: ai tokenomics, optimizing token, tokenomics principles, token optimization, tokenomics, keeping ai, token efficient, guide ai, ai coding, ai sessions
Source: cloud.google.com

🧠 Moonshot AI launches Kimi K3

Good morning! Here's what's happening in AI today: Moonshot AI launches Kimi K3, a 1 million token open frontier model Grok adds Automations for scheduled AI tasks that run on their own tldraw launche...

Keywords: build ai, ai launches, ai released, scheduled ai, moonshot ai, building ai, ai app, ai tools, ai agent, ai agents
Source: simplifyingai.co

The Human-in-the-Loop Is Tired

Yet another thought piece about LLMs. I know. Bear with me. This is an attempt to put words around something I think most developers are experiencing right now but haven't had time to make sense of. P...

Keywords: programming llm, programming llms, llms genuinely, making llm, working llms, llms, plan llm, guiding llms, replace programmers, llm intensely
Source: pydantic.dev

How a 60% speedup on Lucene vector search came from the CPU, not the algorithm

Based on the OpenSearchCon Europe 2026 talk by Shubham Chaudhary, Amazon | Watch the full session → When people talk about making vector search faster, the conversation almost always goes to algorithm...

Keywords: search performance, speedup lucene, search faster, vector search, search computationally, search workloads, opensearch lucene, search latency, like lucene, cpu vectorized
Source: opensearch.org

Inside Cursor’s agent factory: how it verifies AI-written code

Key takeaways • Cursor treats verification as an integrated architecture across CI, security review, risk scoring, behavioral artifacts, and specialized review agents. • Agents work inside developer-l...

Keywords: trust cursor, cursor ecosystem, cursor experience, evaluations cursor, cursor uses, cursor bugbot, engineering cursor, cursor production, evidence cursor, called cursor
Source: arize.com

Minikotlin

class instance → struct.new Every class becomes a GC struct type; properties are real struct fields. Allocation is struct.new , not a hand-rolled heap of bytes. minikotlin is written from scratch in C...

Keywords: kt compile, wasm kt, main kt, compiled wasm, kt files, gc struct, write kt, kt greeter, wasm module, wasm gc
Source: minikotlin.run

NVIDIA Vera Rubin Maximizes Intelligence per Dollar for Post-Training Workloads – a Key Metric for Agentic AI

Think of a professional athlete. What separates elite performers is what happens between games: continuous refinement, adjusting to new opponents and sharpening skills based on what the last game expo...

Keywords: training agentic, agentic ai, agentic models, agentic model, agentic, stop agentic, training cycles, competing ai, agentic post, training ai
Source: blogs.nvidia.com

One RAG Pipeline, Four Very Different PDFs: Same Four Bricks, Every Answer Typed and Cited

This article is the second of two parts on the upgraded pipeline in Part III of Enterprise Document Intelligence, a series that builds an enterprise RAG system from four bricks: document parsing, ques...

Keywords: document pipelines, document intelligence, pipeline pdfs, pdf parsers, document parsing, document parser, answer pipeline, document answering, parse pdfs, parsing document
Source: towardsdatascience.com

Topcoat: The full full-stack framework for Rust

Topcoat is a modular, batteries-included Rust framework for building fullstack apps. It prioritizes simplicity and productivity. See the Getting started guide to set up a new project. Early-stage and...

Keywords: topcoat renders, rust topcoat, topcoat evaluates, use topcoat, allows topcoat, topcoat optionally, main topcoat, topcoat, topcoat serve, topcoat start
Source: github.com

Kimi K3: Moonshot AI’s 2.8T Open-Weight Model — Release, Specs & Pricing (2026)

Quick answer. Kimi K3 is Moonshot AI's frontier model, released July 16, 2026. It's a 2.8-trillion-parameter open-weight, multimodal reasoning model — the largest open-weight model shipped to date — w...

Keywords: models k3, k3 kimi, kimi k3, weights kimi, k3 moonshot, k3 benchmarks, cost kimi, kimi k2, weight moonshot, k3 priced
Source: codersera.com

UIUC AI Teaching Assistant

We're live on HuggingFace: 👉 https://huggingface.co/spaces/kastan/ai-teaching-assistant 👈 Our system runs 11 separate models in parallel for text/image retrieval, generation, moderation and ranking an...

Keywords: retrieval generation, search retrieval, semantic search, retrieval, retrieval rlhf, retrieval augmented, huggingface datasets, image retrieval, evaluation dataset, search
Source: github.com

OpenFeature: swap your feature-flag vendor in one line

OpenFeature: swap your feature-flag vendor in one line Evaluation API, providers, hooks, LaunchDarkly migration, CNCF incubating spec The renewal quote landed in March and the flag vendor wanted rough...

Keywords: openfeature vendor, feature flags, openfeature standardizes, feature flag, openfeature swap, openfeature backed, openfeature, flag vendor, fair openfeature, openfeature exists
Source: podostack.com

Intel-Powered Medical Exoskeleton Helps Patients Get Back on Their Feet

Intel-Powered Medical Exoskeleton Helps Patients Get Back on Their Feet From muscle signals to movement: AI reimagines and speeds recovery. People with lower-limb injuries who cannot stand or walk oft...

Keywords: powered exoskeleton, rehabilitation intel, medical exoskeleton, wearable exoskeleton, exoskeletons powered, exoskeleton therapist, rehab exoskeleton, exoskeleton, bionics exoskeletons, walk exoskeletons
Source: newsroom.intel.com

Open Book Touch: open-source e-reader

Oddly Specific Objects Books KiCad Back this project to help bring it into existence. Funding ends on Aug 20, 2026 at 04:59 PM PDT. Open Book Touch is the device I’ve been trying to build for six year...

Keywords: book device, touch book, book touch, touch reads, open books, books kicad, open book, book reader, books wall, book cover
Source: crowdsupply.com

Optimized CDN caching and deploying of immutable static assets

Vercel now reuses static files across deployments for frameworks that output content-addressed assets. It's zero-config: Vercel uses Framework-defined infrastructure to manage these immutable files al...

Keywords: vercel uses, vercel supported, config vercel, vercel reuses, faster deployments, vercel, additional frameworks, deployments frameworks, reduction vercel, cached deployments
Source: vercel.com

Learning a few things about running SQLite

Learning a few things about running SQLite Hello! I’ve been working on a Django site recently, and I decided to use SQLite as the database. When I was getting started with using SQLite as database for...

Keywords: sqlite database, used sqlite, use sqlite, running sqlite, sqlite, using sqlite, sqlite fts5, sqlite web, sqlite production, sqlite3
Source: jvns.ca

A scorecard for the AI age

A scorecard for the AI age The question I hear from CFOs everywhere is simple: how do we get more value from our AI spend? For years, the market measured the success of software through adoption: seat...

Keywords: value ai, ai spend, ai dollar, ai demands, scorecard ai, ai age, ai useful, ai tasks, improves ai, ai work
Source: openai.com

Cloudflare WAF protects WordPress applications from two high-severity vulnerabilities

Cloudflare WAF protects WordPress applications from two high-severity vulnerabilities Cloudflare has deployed new Web Application Firewall (WAF) protections for two critical vulnerabilities affecting...

Keywords: vulnerabilities cloudflare, waf protections, cloudflare waf, cloudflare deployed, ensure cloudflare, wordpress cloudflare, cloudflare rules, fixes cloudflare, wordpress protections, cloudflare infrastructure
Source: blog.cloudflare.com

Embedded Routing Stack: Give your customers control over model routing

Embedded Routing Stack: Give your customers control over model routing Today we’re launching Embedded Routing Stack in Merge Gateway. It gives your customers dedicated LLM routing controls to manage t...

Keywords: embedded routing, routing stack, models routing, model routing, routing controls, routing policies, includes routing, routing, stack model, llm routing
Source: merge.dev

Google Kills Custom Search API on Jan 1, 2027

Google Kills Custom Search API on Jan 1, 2027. You Have 9 Months Google Custom Search died on January 1, 2027. Here’s the playbook. As of this writing in August 2026, the API is still technically up....

Keywords: search api, api discontinued, google deprecated, cse google, cheap google, google developers, paying google, query google, google limit, google cse
Source: thenextgennexus.com

How to Build an AI Agent That Gets Better the More You Use It

I've been teaching programming for over twenty years. And the single most consistent thing I see in my students isn't a lack of intelligence. It's a lack of feedback. They read the material once. They...

Keywords: improving agents, improving agent, ai agents, agent improves, making agents, design agents, build ai, agents write, agent tools, agents aren
Source: simplifyingai.co

The Creator of Claude Code Doesn’t Prompt It Anymore

-a648dc4ecb66---4 crawled_date: 2026-07-17T12:30:35.985675+00:00 feed_url: https://towardsdev.com/feed published: Fri, 17 Jul 2026 12:04:57 GMT

The Creator of Claude Code Doesn’t Prompt It Anymor...

Keywords: loop agent, agent loops, loop engineer, loops engineer, loops prompt, agent loop, instruction agent, coding agent, loops run, prompt agents
Source: towardsdev.com

ArtfulType: Markdown Writing App for the 68K Macintosh

A distraction-free Markdown writing app for classic 68k Macintosh computers (System 6/7), built to run from a BlueSCSI device on a Mac Plus or similar compact Mac.

Writer mode — live Markdown-to-ric...
Keywords: markdown files, markdown writing, use bluescsi, free markdown, markdown raw, raw markdown, live markdown, markdown mode, plain markdown, markdown rich
Source: github.com

PennyLane is an open-source quantum software platform for quantum

PennyLane is an open-source quantum software platform for quantum computing, quantum machine learning, and quantum chemistry. Create meaningful quantum algorithms, from inspiration to implementation....

Keywords: quantum software, photonics pennylane, quantum computing, chemistry pennylane, research pennylane, innovate pennylane, features pennylane, quantum hardware, platform quantum, using pennylane
Source: github.com

Agentic AI Security: Defending Against Prompt Injection and Tool Misuse

In this article, you will learn what prompt injection and tool misuse are in the context of agentic AI systems, and which defense strategies experts recommend to mitigate them. Topics we will cover in...

Keywords: ai security, dealing chatbots, ai agents, misuse prompt, chatbots, chatbots accidentally, permissions agents, compromise ai, malicious instructions, prompt injection
Source: machinelearningmastery.com

Kimi K3 Benchmarks: How It Stacks Up vs Fable 5, GPT-5.6 Sol & Opus 4.8 (2026)

Quick answer. On the Artificial Analysis Intelligence Index, Kimi K3 scores about 57 — fourth overall, behind Claude Fable 5 (~60) and GPT-5.6 Sol (~59) and narrowly ahead of Claude Opus 4.8 (~56). It...

Keywords: index kimi, benchmark k3, specs kimi, agentic benchmark, intelligence k3, intelligence index, compare kimi, kimi k3, coding kimi, intelligence open
Source: codersera.com

Show HN: On-chain bond market where the issuers are AI agents

Have your AI agent raise money for you by selling bonds

1 Just tell your agent to raise money using sellbonds.now
2 Whenever your agent needs money for something, it will fundraise and repay on it...
Keywords: agent invest, ai agents, ai agent, funding agent, agents core, lenders agent, future agents, issuing debt, live agents, making investments
Source: selbonds.now

How Has Roman Concrete Lasted for Millennia? 1,900-Year-Old Latrine Offers Clues

How Has Roman Concrete Lasted for Millennia? A 1,900-Year-Old Latrine Offers New Clues About the Material’s Impressive Durability A chemical process called carbonation, which helps seal cracks, could...

Keywords: concrete longevity, roman concrete, modern concrete, concrete lasted, durability concrete, concrete durability, concrete materials, resilient concrete, concrete environmental, italy concrete
Source: smithsonianmag.com

Patroni and logical replication fo PostgreSQL: Preventing data loss during failovers

Everyone knows that asynchronous replication failovers — whether performed by Patroni or another tool — can result in lost data. This occurs in situations where certain transactions committed on the l...

Keywords: postgresql replication, replication failovers, failover psql, postgresql replica, replica failover, patroni failover, failover replica, using pg_replication_slot_advance, replica psql, haproxy replication
Source: palark.com

Building Managed Agents That Use GitHub Without Exposing Your Token

Building Managed Agents That Use GitHub Without Exposing Your Token July 17, 20266 minute read You can run the GitHub CLI (gh) inside a Gemini API Managed Agents without exposing your GitHub Personal...

Keywords: github token, api github, github agent, access github, exposing github, github exposing, token git, github personal, github uses, gh github
Source: philschmid.de

GLM-5.2 vs Claude Sonnet 5: how they compare on coding

Table of contents GLM-5.2 vs Claude Sonnet 5: how they compare on coding GLM-5.2 has become one of the most talked-about coding models of the summer, while Claude Sonnet 5 is the model most teams alre...

Keywords: glm worth, sonnet glm, glm vs, glm pricing, tokens glm, glm finished, netting glm, glm ai, better glm, stronger glm
Source: merge.dev

GrapheneOS recommended for domestic abuse victims

In 2026, technology can be a lifeline – or a weapon. For people experiencing domestic violence (DV), tech-facilitated abuse is now one of the most common forms of control. This includes GPS stalking,...

Keywords: phone safety, privacy phones, phone security, phones safety, phone dangerous, safe phones, safe phone, protected device, devices trusted, smartphones
Source: privacypros.com.au

The Good, the Bad and the Ugly in Cybersecurity – Week 29

The Good | Authorities Sanction Cybercriminals & Dismantle Russian Bulletproof Hosting Infrastructure The EU and the United Kingdom have jointly sanctioned multiple Russian individuals and entities fo...

Keywords: sanction cybercriminals, cybersecurity, global ransomware, security products, cybercriminals dismantle, security service, cybersecurity researchers, compromised systems, malware cybersecurity, malware investigators
Source: sentinelone.com

VulnHunter: Capital One's agentic AI code security tool

Announcing VulnHunter Capital One’s open-source, agentic AI code security tool. The rules of software security are changing faster than most defenders can keep pace. Advanced AI models have dramatical...

Keywords: ai security, ai attack, security tools, security tool, vulnerabilities attackers, vulnerabilities software, detect vulnerabilities, capabilities vulnhunter, cybersecurity, rigid cybersecurity
Source: capitalone.com

AI Meets Cryptography 2: What AI Found in OpenVM's ZkVM

This is the second post in the series. In case you have not read the first one on Cloudflare's CIRCL, it has more context on why we run these experiments and how our pipeline is set up. In this post,...

Keywords: openvm pairing, openvm reason, bug openvm, openvm avoided, critical openvm, openvm team, openvm ai, openvm maintainers, fixed openvm, openvm checked
Source: blog.zksecurity.xyz

Chat SDK adds native Slack agent support

You can now build native Slack agents with Chat SDK's Slack adapter. The adapter supports the full Slack agent messaging experience, from agent conversations in the Messages tab to suggested prompts,...

Keywords: chat sdk, agent chat, agents chat, createslackadapter chat, sdk slack, native slack, slack agent, agent messaging, slack native, slack createslackadapter
Source: vercel.com

GLM 5.2 is 35% off via Novita on AI Gateway

GLM 5.2 is 35% off on AI Gateway through July 24 when routed through Novita. To get the discounted rate, set the model to zai/glm-5.2 in the AI SDK and route requests through Novita: agent.ts import {...

Keywords: novita gateway, novita routing, zai glm, novita discounted, routing glm, routed novita, ai gateway, glm requests, novita agent, provideroptions gateway
Source: vercel.com

The U.S. is seeking a share in Korean semiconductor companies’ enormous profits from the global artificial intelligence (AI) chip boom, citing strong U.S. demand for Korean chips, according to an indu...

Keywords: firms korea, korean firms, korean companies, korean chipmakers, korea semiconductor, korean semiconductors, korea trade, korean semiconductor, korean chips, korea entitled
Source: koreatimes.co.kr

The Constitution Contradicts Itself, and Then Gets Translated

The Constitution Contradicts Itself, and Then Gets Translated Twelve models, four labs: the rules meant to align them contradict themselves. And the same model runs warmer in Hindi, stricter in Russia...

Keywords: constitution contradicts, refuse model, models disagreeing, models disagreed, arguing vaccines, principles viewpoint, viewpoint neutrality, models disagree, contradict model, scenarios violate
Source: robonaissance.com

Turn User Signal Into What You Build Next

Turn User Signal Into What You Build Next Fast coding hasn't made teams ship better because the delays live in handoffs, not in typing. Working from a single shared prototype closes that gap. Most sof...

Keywords: builds design, design work, handing design, designed workflow, build fast, interpret design, prototypes, feedback production, user testing, engineer user
Source: builder.io

2026.29: Mainframes and Main Characters

Welcome back to This Week in Stratechery! As a reminder, each week, every Friday, we’re sending out this overview of content in the Stratechery bundle; highlighted links are free for everyone. Additio...

Keywords: concern ibm, ibm announced, misses ibm, mainframe ibm, ibm, ibm mainframe, ibm stock, ibm misses, essential ibm, moat ibm
Source: stratechery.com

Depot is now available in Stripe Projects

We're thrilled to join the Stripe Projects developer preview. You and your agents can add Depot to a project and authenticate straight from the terminal using the Stripe CLI. This is one of the cooles...

Keywords: stripe projects, building stripe, stripe project, workflow stripe, install stripe, depot stripe, run stripe, stripe cli, credentials stripe, using stripe
Source: depot.dev

EEG shows brain can simultaneous encode two speech streams

Figures Abstract Successful speech communication in multi-talker scenarios requires a skillful combination of sustained attention and rapid attention switching. While the neurophysiology literature of...

Keywords: speech attention, attention switching, neurophysiology attention, speech attentional, attention switches, switching attention, switches attention, auditory attention, frequency attention, process attention
Source: journals.plos.org

Runtime logs now show cache reasons

Runtime logs now show a Cache Reason explaining why a request wasn't a fresh cache hit, for example a time-based or tag-based revalidation. Use cache reasons to debug misses and improve your hit rate....

Keywords: cache_reason time, debug caching, cache statuses, cache_reason, cache status, vercel logs, inspect cache, cache reason, cli cache, logs cache
Source: vercel.com