<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Cloudflare on BriefTechNews.com</title><link>https://brief-tech-news.com/tags/cloudflare/</link><description>Recent content in Cloudflare on BriefTechNews.com</description><generator>Hugo</generator><language>en</language><copyright>Copyright &amp;copy;2021 &lt;br&gt; Designed &amp;amp; Developed by [Gethugothemes](https://gethugothemes.com/)</copyright><lastBuildDate>Mon, 05 Oct 2026 12:43:30 +0000</lastBuildDate><atom:link href="https://brief-tech-news.com/tags/cloudflare/index.xml" rel="self" type="application/rss+xml"/><item><title>SaaS vendors add agent tollbooths while foundation models break seat pricing</title><link>https://brief-tech-news.com/blogs/2026-10-05-founders/</link><pubDate>Mon, 05 Oct 2026 12:43:30 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-10-05-founders/</guid><description>Salesforce, HubSpot, and Atlassian are aggressively moving away from per-seat subscriptions, hitting automated agent workflows with annual API surcharges reaching $240,000. Flat-rate pricing models are collapsing across generative AI, with legal startup Harvey recording -50% gross margins as heavy inference calls outpace fixed contracts. Meanwhile, Simon Willison makes the case for hard spend caps by default as autonomous code-generation tools create silent, five-figure cloud bills. Cognition hit $1B ARR just 24 months after launching Devin, rewriting enterprise deployment velocity.</description></item><item><title>Kubernetes Patches Node Exploits, Cloudflare Opens Full Tracing, and Meta Open-Sources Rebalancer</title><link>https://brief-tech-news.com/blogs/2026-10-05-devops/</link><pubDate>Mon, 05 Oct 2026 11:48:28 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-10-05-devops/</guid><description>Kubernetes shipped out-of-band security releases for four minor branches to fix a confused-deputy flaw in StatefulSets and an NTLM coercion vulnerability on Windows nodes. Cloudflare moved Traces into open beta, unlocking automatic OpenTelemetry collection across edge rules, cache layers, and origin routes without code instrumentation. AWS integrated its DevOps Agent with OpenSearch over the Model Context Protocol to automate log triage during incidents. Meta open-sourced Rebalancer, a C++ solver library for shard, server, and traffic allocation scaling up to one million objects.</description></item><item><title>BootLoops reframes Claude for science as the decision model ecosystem goes open-source</title><link>https://brief-tech-news.com/blogs/2026-10-02-ai/</link><pubDate>Fri, 02 Oct 2026 14:00:25 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-10-02-ai/</guid><description>Anthropic and Matthew Schwartz rolled out BootLoops, aligning LLMs directly to quantitative calculations to map mathematical overlaps across ecology and physics. OpenAI fired three safety researchers over external leaks while shelving GPT-6.1 Astra amid agent security failures. Meanwhile, the decision model race erupted: Cloudflare, Jared Palmer, Perplexity, and Strands all shipped open-weight models compatible with TypeSafe&amp;rsquo;s Jev API, while Microsoft dropped sub-100ms streaming speech-to-text.</description></item><item><title>OpenAI stalls IPO, memory crunch looms, and quantum TLS rolls out</title><link>https://brief-tech-news.com/blogs/2026-10-02-it/</link><pubDate>Fri, 02 Oct 2026 12:54:52 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-10-02-it/</guid><description>OpenAI postponed its $852 billion IPO citing safety gaps after agents breached external sites, while Micron warned that over 75 % of its 2027 DRAM capacity is already booked, tightening memory supply through 2028. Microsoft and Google backed Apache Ossie to standardize semantic model exchange, and Cloudflare unveiled zero‑overhead quantum‑safe TLS certificates. Meanwhile, Meta’s consumer AI agent Muse launches, Google’s threat intel shows AI‑driven vulnerability disclosures doubling, and CoreWeave rolls out Forge to tighten AI model ops.</description></item><item><title>Cloudflare launches post-quantum CA as S3 Vectors adds pre-filtering</title><link>https://brief-tech-news.com/blogs/2026-10-02-devops/</link><pubDate>Fri, 02 Oct 2026 11:46:51 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-10-02-devops/</guid><description>Cloudflare is launching a certificate authority using Merkle Tree Certificates targeting Chrome&amp;rsquo;s Quantum-resistant Root Store by 2027. The architecture batches certificates to avoid the 40x log storage penalty of post-quantum signature chains while running 9% faster than classical TLS. AWS rolled out metadata pre-filtering for Amazon S3 Vectors at no added cost, boosting recall up to 5x on selective queries across up to 100 constraints. Meanwhile, Cloudflare overhauled its Containers engine to slash cold starts from four seconds to 648 milliseconds for AI workloads, and Datadog shipped sandboxed code execution in its MCP Server to cut input tokens by 73%.</description></item><item><title>Starlink micro-passes, 37signals kills manual code, and $60B for Anthropic silicon</title><link>https://brief-tech-news.com/blogs/2026-10-02-tech/</link><pubDate>Fri, 02 Oct 2026 10:47:32 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-10-02-tech/</guid><description>Starlink is building out micro-billing infrastructure to sell hourly and weekly satellite passes through decentralized third-party hosts. Anthropic is securing custom silicon capacity through a massive $60 billion debt facility being raised by Broadcom and Wall Street syndicates. Meanwhile, Fervo Energy put the first commercial block of its 4 GW enhanced geothermal plant online in 23 months to power hyperscaler data centers. In software, 37signals officially abandoned manual coding as default practice following Claude Opus 4.5, while AWS embedded DuckDB inside Aurora PostgreSQL to query Iceberg data lakes in place.</description></item><item><title>AWS Can't Restore Bahrain Region Until 2027: Multi-AZ Is Dead</title><link>https://brief-tech-news.com/blogs/2026-09-30-it/</link><pubDate>Wed, 30 Sep 2026 12:53:24 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-09-30-it/</guid><description>AWS cannot restore its Bahrain region or one UAE data-hosting zone until early 2027 after March war damage, breaking the assumption that availability zones fail independently. NVIDIA&amp;rsquo;s Open Agent Safety Platform puts agent policy enforcement in BlueField-4 DPU hardware, arguing agents can&amp;rsquo;t govern themselves. Microsoft made WSLC Linux containers on Windows generally available, and Cloudflare launched free agentic threat intel plus a post-quantum certificate authority. Jev, a specialized model, hit F1 0.948 on identity resolution at $0.62 per 1,000 accounts—a tenth of Claude&amp;rsquo;s cost. Google confirmed ChromeOS support ends mid-2034.</description></item><item><title>Cloudflare’s “cf” CLI opens 3,000‑plus API ops; Modal shatters Kubernetes scaling limits</title><link>https://brief-tech-news.com/blogs/2026-09-30-devops/</link><pubDate>Wed, 30 Sep 2026 11:48:13 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-09-30-devops/</guid><description>Cloudflare’s open‑beta “cf” CLI lets agents tap into more than 3,000 Cloudflare API calls, dwarfing Wrangler’s 280‑command surface and promising JSON‑first automation. Modal unveiled a decentralized sandbox scheduler that spins up a million concurrent sandboxes in under a minute, sidestepping etcd bottlenecks. Meanwhile Pulumi’s GeoDeploy pairs TypeScript cost logic with TypeSafe AI’s Jev model to auto‑place Kubernetes clusters, Uber publishes a context‑aware retry guard that cut a recent storm by 9.5 million calls, and Bolna slashes analytics spend six‑fold by moving from BigQuery to ClickHouse. The day also saw releases from NVIDIA (OpenShell), Vectify AI (PageIndex), and Databricks (Lakebase Search), plus Terraform’s GA 8.0 provider for GCP.</description></item><item><title>16,000 Supabase Databases Left Wide Open: The Config Mistake That Exposed PII</title><link>https://brief-tech-news.com/blogs/2026-09-29-it/</link><pubDate>Tue, 29 Sep 2026 12:41:32 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-09-29-it/</guid><description>Over 16,000 Supabase databases are exposed due to misconfigured row-level security, leaking PII, plaintext passwords, and auth tokens - including 100,000 records from a US valet service. Microsoft pauses the KB5002907 update after it deactivates perpetual Office 2016/2019 licenses. Cloudflare patches a cross-tenant container flaw that let paid customers recover residual data from shared storage. Anthropic launches a 2,000+ plugin marketplace, and Google Meet rolls out new &amp;ldquo;Take notes for me&amp;rdquo; admin controls. Plus: OneDrive gets pay-as-you-go pricing and BitLocker suspension best practices.</description></item><item><title>Alibaba's OpenCodeReview, Bun's Rust Rewrite, and the Week's Biggest DevOps Shifts</title><link>https://brief-tech-news.com/blogs/2026-09-28-devops/</link><pubDate>Mon, 28 Sep 2026 11:49:23 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-09-28-devops/</guid><description>Alibaba open-sourced OpenCodeReview, an Apache-2.0 AI code review CLI that matches Claude Code&amp;rsquo;s precision using roughly one-ninth the tokens, though independent benchmarks show recall as low as 20%. Bun&amp;rsquo;s creator rewrote 535K lines of Zig into Rust in four months with AI agents, costing $165K in tokens. AWS launched CloudWatch Omni, an AI-first observability platform, and GitHub migrated github.com from CSS-in-JS to CSS Modules, cutting SSR time by 55%. Cloudflare fixed a cross-tenant data exposure vulnerability in Containers, and Google added flexible VMs to its Spark service.</description></item><item><title>GKE Pod Snapshots Slash AI Cold Starts, and More Agent Infrastructure News</title><link>https://brief-tech-news.com/blogs/2026-09-25-devops/</link><pubDate>Fri, 25 Sep 2026 11:50:04 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-09-25-devops/</guid><description>GKE&amp;rsquo;s new Pod snapshots cut AI inference startup by up to 89%, loading a 70B parameter model in 37 seconds, a direct hit on the cold-start problem that forces overprovisioning. Cloudflare proposes replacing the SDLC with an agent-driven lifecycle. OpenAI&amp;rsquo;s improved prompt caching for GPT-6 offers up to 90% discounts on cached tokens. Datadog fine-tuned a 9B model to investigate alerts at $0.003 per run, 20x cheaper than a frontier API. Plus: Docker&amp;rsquo;s Sandbox Kit spec, NVIDIA&amp;rsquo;s open-sourced Model Optimizer, and ClusterMAX 3.0&amp;rsquo;s GPU cloud rankings.</description></item><item><title>Kubernetes finally tells you which PVCs are dead weight</title><link>https://brief-tech-news.com/blogs/2026-09-23-devops/</link><pubDate>Wed, 23 Sep 2026 11:48:21 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-09-23-devops/</guid><description>Kubernetes v1.37 promotes the PersistentVolumeClaimUnusedSinceTime feature gate to Beta, so the control plane now flags PVCs no running pod references - no more custom cleanup scripts for orphaned storage. Cloudflare makes Python Workers generally available, letting FastAPI, Django, and Flask run on the edge without TypeScript glue code. A Prometheus survey shows the pain of mixing it with OpenTelemetry is easing, with the &amp;ldquo;hard to use&amp;rdquo; share dropping from 29% to 10%. Microsoft open-sources TauGrid to unify GPU workload scheduling on Kubernetes, and Google&amp;rsquo;s Ax aims to orchestrate billions of agent workloads with declarative manifests.</description></item><item><title>Anthropic Picks Accenture to Police Its Own AI Slowdown, Codex Sandbox Falls, and Cloudflare Cuts 100TB</title><link>https://brief-tech-news.com/blogs/2026-09-21-it/</link><pubDate>Mon, 21 Sep 2026 12:52:39 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-09-21-it/</guid><description>Anthropic named Accenture&amp;rsquo;s Faculty unit as its first embedded evaluator for CEO Dario Amodei&amp;rsquo;s three-step AI slowdown proposal, with both firms investing at least $1 billion over five years. Researchers escaped OpenAI Codex&amp;rsquo;s strictest read-only sandbox mode via a Heapjack exploit, achieving remote code execution on host machines within eight days of the report. Cloudflare reclaimed over 100TB of RAM by tweaking the consistent-hashing library in its Pingora Backend Router. Windows 11&amp;rsquo;s Cloud Rebuild tool now supports remote drive sanitization for secure device decommissioning. Plus, Google Workspace Studio gains custom starters and webhooks, and Supabase Auth hits snags with unsupported SMS providers.</description></item><item><title>Salesforce stumbles, Kubernetes gets pod-level resource control, and the cost of your coding agent's harness</title><link>https://brief-tech-news.com/blogs/2026-09-18-devops/</link><pubDate>Fri, 18 Sep 2026 11:45:55 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-09-18-devops/</guid><description>Salesforce spent hours stumbling through a global outage that stalled requests on an internal login service and exhausted server resources across hundreds of instances. Kubernetes v1.37 graduates Pod-Level Resource Managers to Beta, letting the Kubelet place hardware based on pod-level declarations so sidecars stop wasting dedicated NUMA-aligned cores. A new analysis across 21 model-harness combinations on SWE-bench Lite and Terminal-Bench 2.0 found harness choice changes cost by up to 5x, and the model vendor&amp;rsquo;s own harness is not always the cheapest. OpenBao gets a fully open-source secrets backend on Kubernetes with CloudNativePG, using TLS certificate auth and synchronous replication. Cloudflare adds four granular roles for Workers so agents and teammates get least-privilege access to individual applications.</description></item><item><title>K8s Memory QoS Hits Beta, Cilium 1.20 Lands, and Cloudflare Draws a Line on AI Crawlers</title><link>https://brief-tech-news.com/blogs/2026-09-16-devops/</link><pubDate>Wed, 16 Sep 2026 11:47:12 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-09-16-devops/</guid><description>Kubernetes v1.37 ships with Memory QoS graduating to beta, meaning cluster operators can finally stop hand-tuning cgroup knobs for latency-sensitive workloads. Cilium 1.20 goes GA with Gateway API ExternalAuth and TCPRoute/UDPRoute support, pushing Ingress NGINX further toward legacy status. Cloudflare&amp;rsquo;s new Disallow AI Training setting lets sites block AI crawlers without sacrificing search indexing, solving a dilemma that has been poisoning robots.txt negotiations. Alibaba has open-sourced its internal AI code review tool, which claims millions of defects caught across tens of thousands of developers. And a Databricks blog digs into demand-based state repartitioning for Spark Structured Streaming, tackling the 200-partition wall that eventually hits every stateful pipeline.</description></item><item><title>OpenAI's Agents API Hits Public Beta, Cloudflare Kills TLS Handshake Lag</title><link>https://brief-tech-news.com/blogs/2026-09-11-devops/</link><pubDate>Fri, 11 Sep 2026 11:48:01 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-09-11-devops/</guid><description>OpenAI dropped the Agents API into public beta, letting developers build cloud agents on the same harness that powers Codex and pay only for tokens and tools used. Cloudflare&amp;rsquo;s Automatic Key Exchange cut TLS 1.3 HelloRetryRequests from 52% to 3.7%, shaving 150ms off p90 handshake latency across 45 billion daily connections. Kubernetes v1.37 reserves NodeConditionType constants for draining, maintenance, and graceful shutdown behind an alpha feature gate. GitHub Copilot reworked its cost model to optimize for task outcomes, not token count per call. PagerDuty&amp;rsquo;s PD Automation Runner finally reaches on-prem stacks with outbound-only connectivity.</description></item><item><title>Meta cuts ZippyDB connection sprawl by 19x, Cloudflare ships AI vulnerability hunters, ClickHouse pipes your SQL</title><link>https://brief-tech-news.com/blogs/2026-09-07-devops/</link><pubDate>Mon, 07 Sep 2026 11:47:06 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-09-07-devops/</guid><description>Meta&amp;rsquo;s ZGateway now fronts 40% of ZippyDB traffic at ~1 billion ops/sec, cutting persistent connections by 19x and turning a million-host TLS mesh into a controllable proxy tier. Cloudflare launched an invitation-only Vulnerability Discovery and Remediation service that runs OpenAI&amp;rsquo;s GPT-5.6 Cyber against customer code and prioritizes findings by real production exposure from WAF and traffic data. ClickHouse 26.8 added a |&amp;gt; pipelined SQL operator that chains FROM, WHERE, AGGREGATE, ORDER BY, and LIMIT as explicit stages with no new execution semantics. OpenTofu is researching built-in linting that unifies pre-plan checks and post-plan policy enforcement, with experimental rules landing in v1.13. AWS, NVIDIA, Fastly, YugabyteDB, Docker, and two notable GitHub repos round out the day.</description></item><item><title>BGP Hijack, HPA Scale-to-Zero, and Cloudflare's Zstd Cache Wins</title><link>https://brief-tech-news.com/blogs/2026-09-04-devops/</link><pubDate>Fri, 04 Sep 2026 11:41:30 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-09-04-devops/</guid><description>A BGP hijack diverted Softaculous IP space for 33 hours last week, letting an attacker obtain valid Let&amp;rsquo;s Encrypt certificates and serve a malicious Virtualizor update to a small number of installations. Meanwhile, Kubernetes v1.37 promoted HorizontalPodAutoscaler scale-to-zero to Beta, Cloudflare prototyped Zstandard cache transcoding that shrinks on-disk asset size to roughly one-third, and two new open-source tools - Magnitude for local LLM inference and Semantica for enterprise knowledge graphs - landed on GitHub.</description></item><item><title>The New Valley: Why $5M-$25M ARR Is Where AI Startups Go to Die</title><link>https://brief-tech-news.com/blogs/2026-09-02-founders/</link><pubDate>Wed, 02 Sep 2026 12:38:38 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-09-02-founders/</guid><description>Series A financing for AI startups has hit a wall. Carta data shows just 15.4% of Q1 2022 seed companies reached a Series A within two years, down from 30.6% for 2018 vintage companies, creating a &amp;ldquo;stall zone&amp;rdquo; between $5M and $25M in ARR that investors call the new valley. Meanwhile, a working CFO details how he&amp;rsquo;s churning SaaS products in 2026 by quoting double the price of Claude, and WorkOS launches Relay to keep credentials out of AI agent context. Cloudflare&amp;rsquo;s new Adaptive Intelligence engine reframes bot defense around economics, and a data-driven argument shows that adding more AI agents can make teams slower, with output dropping from 70k to 67k lines per month going from 4 to 8 agents.</description></item><item><title>Cloudflare sheds 100 TB from 1.1.1.1, Kubernetes 1.37 ships, Grafana adds shared saved queries</title><link>https://brief-tech-news.com/blogs/2026-08-28-devops/</link><pubDate>Fri, 28 Aug 2026 11:39:22 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-08-28-devops/</guid><description>Cloudflare cut 100 TB of RAM from its 1.1.1.1 DNS fleet by trimming per-entry cache size from 953 to 420 bytes across 250 billion entries, also boosting insert throughput 43% and cutting lookup latency 19%. Kubernetes 1.37 &amp;ldquo;Garhwal&amp;rdquo; landed with 67 enhancements, including the metrics.k8s.io API finally going Stable after nine years in Beta and HPA scale-to-zero reaching Beta. Grafana 13.2 made saved queries GA with RBAC, search, and Terraform provisioning. Harness, Datadog, and Google Cloud all published pieces arguing that the next platform problem isn&amp;rsquo;t code or compute, but governance for AI agents. DigitalOcean launched v5 Droplets on 5th Gen EPYC with per-resource billing, and Tailscale released Tailcat, a netcat-style tool that rides its WireGuard data plane without needing an account.</description></item><item><title>Anthropic teaches agents to drive lab gear, Cloudflare sheds 100TB of DNS bloat</title><link>https://brief-tech-news.com/blogs/2026-08-28-tech/</link><pubDate>Fri, 28 Aug 2026 10:37:35 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-08-28-tech/</guid><description>Anthropic shipped the Model Hardware Standard, a driver layer that lets AI agents control microscopes, cameras and lasers over a common interface, aiming to collapse weeks of lab setup into minutes. Cloudflare cut the per-entry footprint of its 1.1.1.1 DNS cache by more than 50%, freeing roughly 100TB of fleet memory and trimming lookup latency by 19%. Stripe closed a $7B deal for OpenRouter, which routes over 10 trillion tokens a day across 400+ models. Meta is internally projected to spend as much as $10B a year on Anthropic despite Zuck publicly swiping at the lab. Anthropic&amp;rsquo;s revenue run rate reportedly hit $65B by end of July while OpenAI&amp;rsquo;s tripled from $13B to $40B in a year.</description></item><item><title>AI Agents Are Burning Out Founders; DOJ Probes a16z — Your Week in Tech</title><link>https://brief-tech-news.com/blogs/2026-08-24-founders/</link><pubDate>Mon, 24 Aug 2026 12:33:20 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-08-24-founders/</guid><description>The WSJ reports startup founders are logging brutal hours to keep pace with AI agents whose capabilities and model updates outrun human schedules, producing erratic sleep, inability to focus on anything else, and a pervasive sense that stopping isn&amp;rsquo;t an option. The DOJ is deep into an antitrust probe of Andreessen Horowitz over board seats in competing AI companies — Ben Horowitz at Databricks and Martin Casado at Fivetran named as examples — and the probe may force VC governance changes that reshape term sheets for anyone raising now. PostHog published a deep dive on building a semantic layer to fix the problem of AI agents returning different metric answers to the same question; it sits on top of existing warehouses without data copying and encodes definitions, trusted tables, and join logic. Cloudflare quietly shipped Bot Preference Sync across all plans, automatically writing AI crawler policies into robots.txt so that stated preferences actually match enforcement rules. Notion&amp;rsquo;s survey of 6,000 professionals found 88% still using AI superficially while the company also shipped a TypeScript workspace API for git-versioned, reproducible setups. Local LLM evaluation failures — 44.8 terabytes of test results showing identical models producing different answers across GPU configurations — are a quantization and numerics problem, not a weights problem.</description></item><item><title>Kubeflow graduates, Bun 1.4 lands, and a Cloudflare Spectre bug that bit 12 bits a second</title><link>https://brief-tech-news.com/blogs/2026-08-21-devops/</link><pubDate>Fri, 21 Aug 2026 11:33:48 +0000</pubDate><guid>https://brief-tech-news.com/blogs/2026-08-21-devops/</guid><description>CNCF has graduated Kubeflow, the Kubernetes-native stack for AI training, fine-tuning, and inference, after it crossed nearly 260 million PyPI downloads. Bun 1.4 ships with 1,500 more passing Node test cases, up to 5x lower idle CPU, and a steady migration from Zig to Rust. Cloudflare researchers showed a remote Spectre attack against Workers leaking up to 12 bits per second at 99% accuracy, exploiting a gap in Dynamic Process Isolation that a long-lived Durable Object WebSocket kept open for hours. Datadog revealed gitretriever, a Git mirror that absorbed 20x traffic growth to 100 million requests a week across 5,500 repos while holding 40 ms median latency.</description></item></channel></rss>