AIToday

Today's AI, for Today's work.

Start free60,000+ monthly readers

Latest AI News

Earlier

AI chair runs GTD weekly-review 13 times, gaps and all

The author used a Claude Code skill with a weekly-review command to chair GTD reviews, completing 13 between A…

Zenn AI/MLOct 2, 2026

NVIDIA's DIN Deploy ships C++ local AI samples

NVIDIA published DIN Deploy, open-source C++ samples that combine ONNX Runtime with the TensorRT RTX execution…

Top Companies AIOct 1, 2026

IBM Bob agentic dev platform goes self-hosted for enterprises

IBM rolled out a self-hosted option for IBM Bob, its agentic software development platform, supporting on-prem…

SiliconANGLE AIOct 1, 2026

DevOps engineer: Hermes Agent build cost him the thrill

A DevOps engineer says he finished his "PersonalAI Health Agent" project in 2 days with Hermes Agent and GPT 5…

Lobsters AIOct 1, 2026

Kirill Brilliantov: elaborate agent harnesses add nothing

In a paper, Kirill Brilliantov, Alejandro Hernández-Cano, and Emmanuel Abbé report that under an equal time bu…

Apple Machine LearningOct 1, 2026

Apple's RLTL;DR lifts Qwen 3.5 9B from 0% to 13% Pass@1

Apple researchers introduced RLTL;DR

Apple Machine LearningOct 1, 2026Also reported by Top Companies AI

Tokyo dev validates typesafe/jev job-listing checks at 97.8% accuracy

A developer ran TypeSafe's Jev, a non-generative judgment model on Cloudflare Workers AI, over 30 fictional ca…

Zenn AI/MLOct 1, 2026

Three live mobile RPGs move to Unreal Engine 5.8 without a rewrite

The developer kept the old Cocos2d-x 3.17.2 C++ and Lua running untouched, and built a compatibility runtime o…

Zenn AI/MLOct 1, 2026

LogicStudio: logic IDE splits loose matching from strict deduction

Kisaburo Yamada sketched the LogicStudio logic IDE, its LLint checker, and a .wm world-model format, splitting…

Zenn AI/MLOct 1, 2026

Adversarial AI code review: useful on 99 回, but 9× the tokens

A team ran its own Claude Code adversarial-review skill for 99 reviews of 85 diffs, using 8 parallel lanes plu…

Zenn AI/MLOct 1, 2026

Claude Code 2.1.278: @AGENTS.md import works, stand-alone AGENTS.md fails

The author tested Claude Code 2.1.170 and 2.1.278 in 12 tool-disabled runs on Windows; @AGENTS.md import retur…

Zenn AI/MLOct 1, 2026

requirements-interview skill turns talk into SPEC.md and PLAN.md

The Zenn write-up describes requirements-interview, a Claude-only (claude.ai / Claude Code) AI skill that take…

Zenn AI/MLOct 1, 2026

Claude Code scripts: 50% one-off, not 69%

Recounting 301 Claude Code session logs on 2026-10-01, the one-off script rate was 50%, versus 69% in the firs…

Zenn AI/MLOct 1, 2026

7 subagents, day one: qa-reviewer fails a draft, 2 roles idle

In a repository running 7 subagents, qa-reviewer — limited to read-only tools — rejected a draft because a fir…

Zenn AI/MLOct 1, 2026

AMD unveils AMD Ross AI agent for embedded design

AMD announced AMD Ross, an agentic AI assistant for embedded design and development that runs across the AMD E…

AI Watch (Impress)Oct 1, 2026

Google DeepMind Launches Gemini 4 Argon, Limited to Fairwind Program

Google DeepMind launched Gemini 4 Argon for coding, enterprise knowledge work and cyber defense, claiming firs…

Latent SpaceOct 1, 2026

Panasonic puts 13.3-inch NC7 Let's Note on sale Nov 18

Panasonic Connect added the 13.3-inch NC7 to its Let's Note lineup for individual buyers, selling it only on i…

AI Watch (Impress)Oct 1, 2026

AWS ties Kiro prompt tuning to LLM judge, A/B tests

AWS said on August 25, 2026 that it evaluates Kiro's system prompt by scoring real internal conversations with…

ITmedia AI+Oct 1, 2026

Claude API checks pages against source of truth

A developer built a checker that feeds source-of-truth JSON and a published HTML page to the Claude API, which…

Zenn AI/MLOct 1, 2026

yomiyasu skill fixes AI-slop Japanese at the syntax level

Developer nanaism released 'yomiyasu', a rewriting skill for AI agents such as Claude Code, Codex and Cursor…

Zenn AI/MLOct 1, 2026

『リーダブルコード』 re-read: excess AI code defense hurts readability

Writing up notes on Chapter 1 of 『リーダブルコード』, the author reviewed AI-generated code in his own 'AI Portfolio Ma…

Zenn AI/MLOct 1, 2026

Claude Code Skills: build Japanese office tasks with SKILL.md

After building 12 Claude Code Skills for Japanese business tasks, the author published the recipe and the full…

Zenn AI/MLOct 1, 2026

Zenn guide: SKILL.md turns Claude Code into routine-work handler

A Zenn guide explains building Claude Code Skills via SKILL.md to handle routine work — meeting minutes, invoi…

Zenn AI/MLOct 1, 2026

Local LLM vs Claude Opus 5: same app, same blind spots

The same requirements document was given to gpt-oss:120b (single-shot generation) and Claude Opus 5 (agent), b…

Zenn AI/MLOct 1, 2026

Google unveils Gemini 4 Argon for cyber defenders

Google announced Gemini 4 Argon, rolling out to trusted cyber defenders via its Fairwind Program, at $2 per mi…

Top Companies AISep 30, 2026Also reported by Ars Technica AI and 1 others

JR East unit leans on GitHub Copilot to modernize work, not just systems

JEIS, a wholly owned JR East unit running over 200 systems, is using GitHub Copilot under a policy of giving m…

Top Companies AISep 30, 2026

Canon PPS adopts GitHub Enterprise Cloud for AI-driven development

Canon Production Printing Systems set up its Digital Innovation Promotion Section in January 2026 and brought…

Top Companies AISep 30, 2026

Claude Code site past 100 pages, sales at 0 yen in one month

A developer published a one-month operations record for a site of more than 100 pages built with Claude Code…

Zenn AI/MLSep 30, 2026

Lauren Tan: 2,000 PRs a month via Trust Engineering

Lauren Tan says she shipped about 2,000 pull requests a month to production on the SpaceX AI Grok Bot team

Zenn AI/MLSep 30, 2026

AI-DLC 3 rounds at Kanri cut PBI days by 3割

Kanri's benefits team ran three AI-DLC rounds; round one's three PBIs finished about 30% faster in days than c…

Zenn AI/MLSep 30, 2026

Claude Code AWS cost probe tops GENIEE Tech Blog reads

GENIEE's tech blog team named its three most-read posts from September 2025 to September 2026, led by an SRE t…

Zenn AI/MLSep 30, 2026

Claude Code agents froze at last line on .claude/ permission prompt

In a trial run, an agent set up .NET, built the core library, ran 34 unit tests and configured CI in about fiv…

Zenn AI/MLSep 30, 2026

Spec index cut AI reading by about 97.2% in developer's test

Reading one spec-index file of 803 lines on 2026年9月29日, an external program returned about 798 tokens against…

Zenn AI/MLSep 30, 2026

KnowledgeSense to launch CodeSense within weeks

KnowledgeSense said on September 30 that CodeSense, a Japanese-made AI agent, will ship within a few weeks, an…

AI Watch (Impress)Sep 30, 2026

50.0% of large firms say AI is used company-wide

Among respondents at companies with 1,001+ employees, 50.0% said AI is used company-wide, and 46.0% flagged AI…

ITmedia AI+Sep 30, 2026

Google names 4 winning AI agent design patterns from Challenge

Google published four design patterns from the top entries of its Google for Startups AI Agents Challenge, jud…

ITmedia AI+Sep 30, 2026

Turn today's AI news into today's work.

The day's essentials from 100+ sources, delivered to Email, LINE, or Slack. Free to start.

60,000+ monthly readers
Get Started Free

Codex CLI refresh adds /voice as Codex Cloud lands

OpenAI refreshed Codex CLI with a full-screen interface, /agents to see and jump between running tasks, /fork…

GIGAZINE AISep 30, 2026

OpenAI DevDay 2026: Codex plugin builds Chromatic games

At OpenAI DevDay 2026, OpenAI launched a Codex plugin for ModRetro's Chromatic handheld, letting users develop…

GIGAZINE AISep 30, 2026

OpenAI brings Ultrafast to GPT-6 Astra, 8x faster in Codex

OpenAI said Ultrafast is available today for GPT-6 Astra in Codex, ChatGPT Work, and the API, running up to 8x…

GIGAZINE AISep 30, 2026Also reported by ITmedia AI+

Equals Money opens MCP server, but no payments by AI

Equals Money's James Simcox said the firm opened an MCP server to customers' AI tools, but chose that customer…

SiliconANGLE AISep 30, 2026

CLAUDE.md grows 2 lines only for the 7 rules added

In the roughly two days after switching to a four-step placement test, the author added seven rules and pieces…

Zenn AI/MLSep 30, 2026

cmux's NODE_OPTIONS preload can stop node from starting

In cmux (version 0.64.25), the shell environment includes a cmux preload file in NODE_OPTIONS; once that file…

Zenn AI/MLSep 30, 2026

Claude Code subagents silently inherit opus; echo task hits 101,418 tokens

A freelance engineer measured his Claude Code setup and found 4 of 6 subagent definitions left the model field…

Zenn AI/MLSep 30, 2026

Anthropic's finance agent templates trip engineers on tool_use

Anthropic has published agent templates for financial work such as KYC, pitchbook generation and monthly closi…

Zenn AI/MLSep 30, 2026

Cheapest Claude Code + Codex pair beats one upgrade

A student who started out on Claude's Pro plan ran out of usage quickly, so he subscribed to the cheapest paid…

Zenn AI/MLSep 30, 2026

Claude Code drives Blender to build game character base mesh

A developer built the head and body base for an anime-style game character in Blender, running the tool headle…

Zenn AI/MLSep 30, 2026

Sonnet 5.5 plus Opus 5.5 advisor pitched as easy cost cut

A Zenn author says choosing Sonnet 5.5 for the main session while setting Opus 5.5 as an advisor is a good way…

Zenn AI/MLSep 30, 2026

Google Cloud Developer Plugin for AI Coding Agents announced

Google Cloud announced the Google Cloud Developer Plugin for AI Coding Agents, which gives coding agents such…

PublickeySep 29, 2026

Sonnet 5.5 handles collection and coding, but falls short on review work

A Zenn write-up tested Opus 5.5, Fable 5.1, Sonnet 5.5 and Haiku 4.5 on design-doc review, code review, design…

Zenn AI/MLSep 29, 2026

Zenn author: cheap AI failures reshape prototyping

In a Zenn essay, a developer describes AI-assisted prototyping shifting toward many cheap attempts, combining…

Zenn AI/MLSep 29, 2026

Claude Code bypassPermissions still blocked by classifier

With bypassPermissions and 100 allow rules set, a user's Claude Code still rejected some operations

Zenn AI/MLSep 29, 2026

Claude Code: the five layers that silently stall unattended runs

A months-long field note on running Claude Code unattended overnight logs the failure modes where the tool kee…

Zenn AI/MLSep 29, 2026

AI code runs clean but breaks: 101 DB calls, 7-day tokens

A Zenn write-up lays out five AI-written code patterns that never throw errors: I/O inside a loop (101 DB call…

Zenn AI/MLSep 29, 2026

Claude Code subagents hand off via files, not disallowedTools: Agent nesting

Per Claude Code's official docs, subagents can launch their own subagents via the Agent tool by default, neste…

Zenn AI/MLSep 29, 2026

AI COBOL-to-Java conversion called 'more harm than good'

In a column for ITmedia, author Yoshihiro Matsuda argued that mechanically converting COBOL business logic int…

ITmedia AI+Sep 29, 2026

LSTM guide predicts AAPL prices for option trading

A Qiita guide demonstrates an LSTM model with two 50-unit layers and 0.2 Dropout, trained for 50 epochs on a 6…

Qiita 機械学習Sep 29, 2026

panda studio runs a store with Claude Code, sales up 約3.5倍

panda studio, a one-person digital goods store run with Claude Code, saw daily sales rise about 3.5x from July…

Zenn AI/MLSep 29, 2026

Claude Code subagent hit 約87万トークン; split by task

Across about 80 Claude Code sessions in one week in September 2026, one implementation subagent reached roughl…

Zenn AI/MLSep 29, 2026

Claude Code, Cursor, Copilot: teams run 2–3, not one

A September 2026 guide organizes AI coding tools into three types — terminal/agent tools like Claude Code, AI-…

Zenn AI/MLSep 29, 2026

Claude Code job on GitHub Actions stayed green for 2.5 days

A GitHub Actions job calling Claude Code on a schedule returned "You've hit your weekly limit" and exited in a…

Zenn AI/MLSep 29, 2026

5 Claude Code sessions, 1 git tree: 8 accidents in 3 months

Running Claude Code in one shared working tree produced 8 accidents and near-misses between June and September…

Zenn AI/MLSep 29, 2026

Lean 4 test: AI proof code fails as-is

Testing three short problems with Lean 4, the author found AI math explanations were correct while the propose…

Zenn AI/MLSep 29, 2026

Claude Code's Boris Cherny: Black-box AI code OK only for prototypes

Claude Code's creator Boris Cherny answered a developer's question on September 11, 2026, saying throwaway pro…

ITmedia AI+Sep 28, 2026

ITR analyst: Japan firms 'almost absent' from harness engineering

ITR principal analyst Hiroaki Koumoto said Japanese firms' efforts in harness engineering are 'almost nonexist…

ITmedia AI+Sep 28, 2026

Momentic launches Mo, an AI agent that tests apps without scripts

Momentic Inc. launched Mo, an AI agent that opens an app from a URL and prompt, spins up a swarm of agents to…

SiliconANGLE AISep 28, 2026

Meta launches Meta Enterprise Platform to sell AI

Meta is launching the Meta Enterprise Platform, a new business unit selling the Muse agent, Meta Business Agen…

THE DECODERSep 28, 2026

The AI Conference at Pier 48 draws 120+ speakers

The AI Conference announced its full agenda, with 120+ speakers and an anticipated 5,000 attendees at Pier 48…

Yahoo Finance AISep 28, 2026

Wuhan court awards 20,000 RMB in first AI-cost copyright ruling

A Wuhan court awarded 20,000 RMB (about $2,900) to a company whose AI-produced one-hour short drama was copied…

THE DECODERSep 28, 2026

AI results can shift; log prompt_version and test data

For teams using AI in business workflows, outputs like text length, bullet counts, or JSON formatting can chan…

Zenn AI/MLSep 28, 2026

Claude Code tasks sit idle 3 days as prior runs stay stuck

Two Claude Code scheduled tasks on 9:10 and 10:01 morning runs produced no start rows, no errors and no notifi…

Zenn AI/MLSep 28, 2026

Claude Code harness built via CLAUDE.md, rules and Hooks in 4 steps

A Zenn write-up describes organizing Claude Code in an existing project, starting with a CLAUDE.md and layerin…

Zenn AI/MLSep 28, 2026

spec-lane v0.11.0 blocks YAML truncation that fooled progress gates

In spec-lane, a plain scalar success condition — some text followed by '#' and a comment — lost the commented…

Zenn AI/MLSep 28, 2026

Today's AI, for Today's work.

Get curated AI news from 100+ sources delivered daily to your inbox. Free to use.

60,000+ monthly readers

Get Started Free

Free · 30 seconds with Google · unsubscribe anytime