Blogger drops Next.js after chasing perfect Lighthouse scores
The author says Next.js 16.1.6 still shipped a client runtime and about 230 KB of unused JavaScript on static pages.
The author says Next.js 16.1.6 still shipped a client runtime and about 230 KB of unused JavaScript on static pages.
Cache-read cuts matter most for coding agents, where cache reads are 96–99% of token spend, the author says.
The hub adds skill scanning, an agent VPN, Sage runtime hooks, and Norton monitoring between an agent’s decision and execution.
Crafted Brotli data ran code on Node.js 24.22.0 Linux arm64; 24.22.1 rejected it, and no wild attacks were reported.
Fixed builds are 22.23.4, 24.22.1, and 26.12.1; a public proof of concept did not achieve code execution on a standard build.
Inception's own tests give Mercury 3 eight of 12 wins and 1,068 tokens per second, with no independent verification yet.
Pwn.ai leaked host memory on Vercel production sandboxes and gained root on lab hosts; guest-only updates do not fix it.
Nebius, Nscale and Verda are building on Vera, with server makers and agent-sandbox vendors also named.
CVE-2026-99999 lets an x86 KVM guest alter already-validated IP packets, read host memory, and potentially run code.
Researchers show CVE-2026-99999 lets a KVM guest escape to the host via Linux checksum handling.
Vercel Labs published tsc-ts, a TypeScript compiler written in TypeScript.
Linux patched CVE-2026-99999, a networking flaw researchers used to escape VMs to the host.
pwn.ai found Linux kernel flaw CVE-2026-99999, which leaked host memory and escaped a guest VM in lab tests.
Linux kernel flaw CVE-2026-99999 can let a guest VM leak host memory and, in a lab, run code on the host.
Linux kernel flaw CVE-2026-99999 lets guest VMs read host memory, shown on Vercel sandboxes but not exploited in the wild.
Next.js 16.4.1 and 15.5.28 patch two critical unauthenticated RCE bugs and a Server Actions denial of service.
Vercel patched unauthenticated Next.js remote code execution in SVG image optimization, with no active exploitation reported.
Datadog tests whether prompt injection can steer TypeSafe’s Jev and bypass LangChain tool-risk gating.
MarkTechPost ranks nine 2026 AI gateways, placing self-hosted Bifrost first and noting Portkey's Palo Alto deal.
Security Affairs round 599 collects weekly headlines on breaches, arrests, ransomware, and exploited flaws.
Nudge Security markets an agent that inventories and scores risky SaaS OAuth grants teams cannot review manually.
Google paused some bug bounty submissions after low-quality AI-generated vulnerability reports overwhelmed triage.
Paulos Yibelo claims a KVM guest-to-host root escape zero-day found via Vercel’s Firecracker sandbox bounty.
Researcher claims a KVM guest-to-host escape zero-day; Vercel confirmed it, but no details are public.
Vercel confirmed researcher Paulos Yibelo's KVM zero-day guest-to-host root escape and paid a $50,000 bounty.
TypeSafe's Jev and rivals return calibrated choices and scores instead of text, at much lower cost and latency.
A fake client shared a Dropbox git repo whose post-checkout hook fetched and executed a remote binary.
Photon raised $4.5 million seed to let developers run agents over iMessage, WhatsApp, and other channels.
Unauthenticated attackers can execute code on Next.js 16.2.0–16.3.5 servers via crafted SVG in ImageResponse.
OpenAI turns ChatGPT into an app discovery and launch platform with Dots agents, extensions, Sign in with ChatGPT, and an enterprise marketplace.
About 670 lookalike Jev domains appeared days after launch, with fake shops marking up API access.
TypeSafe AI launched Jev, a closed decision model that returns typed, calibrated choices for agent loops.
Attackers impersonate US HR and payroll vendors with fake desktop apps that silently install ConnectWise ScreenConnect.
Google released Gemini 3.8 Flash TTS and Flash-Lite TTS for designed, cloned, and multilingual synthetic voices.
Vercel patched critical Next.js ImageResponse flaw CVE-2026-94545 (CVSS 9.5) allowing server code execution via crafted SVG input; no attacks reported yet.
TypeSafe AI launched Jev, a transformer-based 'System One' model API returning typed, calibrated decisions with probabilities, claiming 193x faster and 445x cheaper than LLMs.
TypeSafe AI, founded by ex-OpenAI RLHF inventor Diogo Almeida, released Jev, a non-LLM transformer outputting calibrated probabilities for cheap software automation.
OpenAI released its Agents API in public beta, exposing the managed Codex harness with hosted or self-hosted sandboxes, MCP tools, and subagents.
Top AI open source projects like Vercel, Astro, Flue, and tldraw are restricting external PRs and using agent-based software factories to triage, fix, and review contributions.
Vercel patches two critical Next.js unauthenticated RCE flaws: a libheif AVIF heap overflow (CVSS 9.5) and a Windows path traversal (CVE-2026-75604).
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.
Today 673 readers · 822 views · 621 reads · all time 159K views