Daily Tech News - 2026-08-16

Done.

Featured Articles

Ubuntu is growing faster on Windows 11 than on native Linux PCs, says Canonical

Canonical’s VP of Engineering says Ubuntu’s user base inside Windows Subsystem for Linux is growing significantly faster than its native desktop user base, and he expects WSL to overtake native Ubuntu...

Keywords: linux wsl, wsl ubuntu, introducing wsl, ubuntu wsl, windows wsl, platform windows, wsl developers, built wsl, versus linux, pc wsl
Source: windowslatest.com

Show HN: Laptop is the last place your secrets are still in plaintext

Just-in-time credentials for your dev machine. Documentation · Quickstart · Supported tools · Command reference · Security Status: macOS-only (Apple Silicon), and still in development. Your secrets li...

Keywords: tap jitpass, jitpass tap, install jitpass, jitpass homebrew, credentials jit, jitpass, jitpass jit, download jitpass_darwin_arm64, credential_process jit, upgrade jitpass
Source: github.com

Models Are Getting Dumber on Purpose

Models Are Getting Dumber on Purpose Reasoning scores keep climbing while per-token compute keeps dropping. GLM-5.2 scores 99.2% on AIME 2026 with about 40 billion parameters active per token. Qwen3.5...

Keywords: benchmark factual, rate benchmarks, knowledge benchmark, factual recall, benchmarks conclude, benchmarks, benchmarks ask, code benchmarks, benchmark means, benchmark
Source: w4g1.dev

Patterns and problems in emerging multi-agent systems

Subscribe to the Frontier Red Team newsletter Get updates on our latest red-teaming research and findings. Models are improving and AI agents are taking on more tasks in shared codebases, markets, and...

Keywords: agent strategizes, collaborate agents, ai agents, coordination agents, agents broadly, agents prevalent, efficiently agents, agents future, agentic coordination, agents excelled
Source: anthropic.com

Qwen 3.8 27B is excellent, but it defaults to wildly overthinking things

Qwen 3.8 27B is excellent, but it defaults to wildly overthinking things 16th August 2026 Friday’s big release was Qwen 3.8 27B, an Apache 2 licensed 27B parameter vision-capable LLM from Alibaba’s Qw...

Keywords: model qwen3, qwen supports, qwen strongest, larger qwen, default qwen3, boost qwen, qwen3, qwen models, qwen3 comes, qwen3 27b
Source: simonwillison.net

I've built a free, open-source local and remote memory system for agentw and CL

Give your coding agent a memory in one command — no signup, just a folder Your agent re-solves yesterday's problem because nothing wrote it down. The fix isn't a platform you adopt — it's one command,...

Keywords: agent memory, amnesia usual, folder agent, agent writes, 5432 agent, invisible memory, amnesia, run read, wakes amnesia, cleanup task
Source: lorekit.io

Four minutes after midnight, Codex said the page was live

At localhost, Dominik Kundel walked through how coding with Codex has changed at OpenAI over the past year: autocomplete, then pair programming, and now what he calls agentic delegation. Code has stop...

Keywords: agent context, agentic delegation, agent needs, agent write, agent explaining, agents exactly, asking agent, agents working, tell agent, agent
Source: render.com

RAM, GPU, and Storage for Agentic AI: How Much You Actually Need

Updated August 16, 2026: Initial publication. Model sizes and memory figures verified against official model releases as of this date; this table is refreshed quarterly because model sizes churn. Ever...

Keywords: memory requirements, memory budget, model memory, sizes memory, memory recommendation, models 16gb, model sizes, 78gb q4, costs memory, 2026 memory
Source: storagereview.com

Show HN: I built a native app for coding agents with Rust and GPUI

One native app for all your coding agents. Waku drives the agent CLIs you already have — sessions, transcripts, tool activity, and checkpoints in one fast graphite window, entirely on your machine. Na...

Keywords: agents waku, telemetry waku, waku cloud, waku, app coding, coding agents, cloud agents, chat log, agent works, timeline agent
Source: waku.sh

Oracle APEX 26.1: When the Governed Runtime Becomes the Product

The Question CIOs Should Be Asking Isn’t Whether AI Can Write Apps It can. That argument is over. Your developers are already using coding agents, and so are people on your team who don’t call themsel...

Keywords: oracle ai, make ai, platform oracle, oracle realistically, isn ai, applications oracle, ai application, execute ai, oracle making, apex ai
Source: moorinsightsstrategy.com

Joy & Curiosity #95

Joy & Curiosity #95 Interesting & joyful things from the previous week In the last two weeks I’ve shipped: a new experimental provider backend for our orbs, an in-product bug reporting feature (not re...

Keywords: orbs backend, backend orbs, warnings orbs, orbs jellyware, orbs working, fixed bugs, logs orbs, orbs product, works orb, bug reporting
Source: registerspill.thorstenball.com

The complete AI SEO playbook: from zero to 4.6M impressions in 3 months

The complete playbook for building an AI-powered content engine that actually ranks — from zero to 4.6M impressions in 3 months. This is the methodology, the toolkit, and the hard-won lessons from bui...

Keywords: ai build, content engine, building ai, ai agent, ai agents, guide ai, ai seo, ai template, automated quality, ai overview
Source: github.com

Show HN: Grafana agent observability for Hermes Agent

Grafana Agent Observability plugin for Hermes Agent. Records LLM calls and tool executions as generations and emits OTel traces + metrics. Paste this into Hermes (or any Claude / Codex / Cursor / simi...

Keywords: hermes plugins, hermes plugin, hermes agent, plugin hermes, hermes logs, install hermes, grafana agent, hermes client, hermes install, hermes config
Source: github.com

Waku combines coding-agent CLIs in a native local desktop app

Waku current version 0.1.0 is a native desktop control plane for the coding-agent command-line tools developers already run. Built in Rust with GPUI, Waku collects sessions, transcripts, tool activity...

Keywords: api waku, developer waku, waku operate, gpui waku, waku builder, waku controls, agents waku, infrastructure waku, waku desktop, independently waku
Source: runtimewire.com

Mads Thines launches LoreKit to give coding agents local, inspectable memory

Mads Thines, a Copenhagen-based designer and product engineer, has launched LoreKit, an open-source memory layer that gives AI coding agents a record of earlier mistakes, fixes and project-specific le...

Keywords: memory agent, agent memory, thines users, mads thines, engineering lorekit, memory developers, agents thines, coding agents, thines describes, tools lorekit
Source: runtimewire.com

New Linux Driver Enables 12VHPWR Monitoring For ASUS ROG Astral Graphics Cards

New Linux Driver Enables 12VHPWR Monitoring For ASUS ROG Astral Graphics Cards For those concerned about current imbalance or voltage changes with the 12VHPWR power connectors and having an ASUS ROG A...

Keywords: 12vhpwr monitoring, 12vhpwr sensor, 12vhpwr reporting, monitoring asus, driver 12vhpwr, enables 12vhpwr, monitoring hwmon, 12vhpwr pin, 12vhpwr, changes 12vhpwr
Source: phoronix.com

Big Pickle on SWE Atlas – Codebase QnA

Task Resolve Rate: 50.8% (63/124) — big-pickle , the free stealth model on OpenCode Zen, evaluated on Scale AI's SWE Atlas Codebase QnA benchmark using the mini-swe-agent scaffold. Run on 2026-08-11 w...

Keywords: qna benchmark, qna leaderboard, benchmark scale, atlas benchmark, qna tasks, raw benchmark, tasks scaleapi, scoring benchmark, model benchmark, comparable leaderboard
Source: github.com

MathCode, Mathematical Coding Agent

MathCode is a terminal AI coding assistant with a built-in math formalization engine. Give it a math problem in plain language and it will automatically convert it into a Lean 4 theorem and attempt a...

Keywords: mathcode git, mathcode launcher, mathcode terminal, local mathcode, mathcode, mathcode prove, org mathcode, use mathcode, mathcode setup, login mathcode
Source: math-ai-org.github.io

How I think about reducing AI costs

How I think about reducing AI costs AI inference costs are something I've been writing about for a while. It's clear that for many companies this is becoming a huge problem: AI spend per employee per...

Keywords: ai costs, costs ai, spending ai, ai spend, audit costs, reducing cost, reducing ai, inference costs, understanding costs, cost optimization
Source: martinalderson.com

ProofRun – a local verification receipt for AI coding agents

ProofRun doesn't judge whether your code is correct. It proves — cryptographically, not by asking nicely — which checks actually ran against the exact code you have right now. An AI coding agent says...

Keywords: pass proofrun, calls proofrun, checkable proofrun, proofrun run, runs proofrun, proofrun command, proofrun status, proofrun init, proofrun binary, proofrun v1
Source: github.com

A True Telnet BBS on a Casio Calculator

Hosting a Bulletin Board Service on My Casio VX-4. About a eight weeks ago I had virtually no interest in calculators. Blank, dull, tedious-looking objects that remind me of long, arduous rainy aftern...

Keywords: casio calculator, calculators capabilities, calculator emulator, software casio, virtually calculators, available calculator, calculators imported, make calculator, calculator moniker, popular calculator
Source: ei3lh.eu

Untitled

Everybody knows that chatbots make mistakes. They have come a long way since Eliza decades ago, but there is still plenty of room for improvement. I’ve been having a lot of conversations with chatbots...

Keywords: chatbots rely, chatbots fail, reason chatbots, problem chatbots, chatbots turn, chatbots produce, chatbots say, chatbots make, chatbots lately, chatbots professionally
Source: vijil.ai

The Case Against Formal Verification, 50 Years Later

The Case Against Formal Verification, 50 Years Later Engineers are getting excited about software verification! This may come as a surprise, since verification has long been considered useful only in...

Keywords: software verification, verification engineering, verification rarely, software correctness, formal verification, verification software, verification formal, automatic verification, make verification, correctness assurance
Source: ivan-gavran.github.io

You can just choose how many bugs you want now

There’s a bizarre aspect of AI coding that I’ve been trying to put my finger on, and I think it’s this: you can basically just decide how many bugs you want your software to have now. We discovered th...

Keywords: fixing bugs, decide bugs, bugs correctness, bugs complex, finding bugs, bugs technique, security bugs, bugs pr, bugs nearly, overall bugginess
Source: nolanlawson.com

Nvidia discloses $21B stake in SpaceX at end of second quarter

Nvidia disclosed on Friday that its stake in Elon Musk's SpaceX was worth about $21 billion at the end of the second quarter. The chipmaker said in a filing with the Securities and Exchange Commission...

Keywords: spacex stock, investor spacex, spacex worth, nvidia shares, nvidia billion, shares spacex, billion nvidia, spacex acquired, investment musk, biggest investor
Source: cnbc.com

XCancel — An Unofficial Twitter/X Mirror

XCancel is an instance of Nitter. Nitter is a free and open source alternative Twitter front-end focused on privacy and performance. The source is available on GitHub at https://github.com/zedeus/nitt...

Keywords: twitter javascript, browse twitter, use twitter, alternative twitter, follow twitter, needing twitter, twitter, twitter users, important twitter, twitter account
Source: xcancel.com

Access to telemetry data: Automotive industry criticizes intelligence reform

Access to telemetry data: Automotive industry criticizes intelligence reform The automotive industry does not want to be made a prolonged arm of the secret services and is demanding improvements to Do...

Keywords: data automotive, vehicle data, intelligence services, intelligence service, federal intelligence, intelligence agencies, german automotive, telemetry data, rights data, data access
Source: heise.de