Every note, newest first.
7 notes
SecurityAI
I checked the GitHub Universe 2026 security launches in the official changelog and docs: npm package firewall, Dependabot reachability, AI Scan, and more. Plans, costs, and limits.
9 min
Infrastructure
Unlike monitoring, observability lets you ask new questions later. Logs, metrics, and traces differ in use and cost, and a trace ID links them, as I checked with OpenTelemetry.
7 min
AutomationAI
Lightpanda, a headless browser built in Zig, hit 1.0. With no rendering, official tests show 9x Chrome's speed at 1/16 the memory, but screenshots and layout don't carry over.
8 min
AI
US wealth manager Arca has AI agents run client cases that take up to 10 days. From a Latent Space interview: per-client records, rule-based stops, change logs, and nightly fixes.
9 min
AI
Agentic Workflows, Copilot app Projects, and Copilot Dreaming, checked against GitHub docs: release stage, plans, how AI credits are charged, and what changes without approval.
7 min
AI
A deck that condenses 137 checklist items into 16 principles drew attention. I use it as a yardstick for yomiyasu, natural-japanese, meiseki, and textlint's AI rules.
10 min
AI
Israel's 42AI left stealth with $62.5 million for a world model that predicts human behavior. How it differs from Genie 3, Cosmos 3, Marble, and V-JEPA 2, with uses and limits.
10 min
48 notes
AI
Using Claude Code via Google Cloud Agent Platform in an organization: budgets and limits, collecting usage, and measuring impact, checked against Anthropic's article and the docs.
9 min
AI
Claude Platform's engineering and product leads described separating an agent's brain from its hands on Latent Space. Checked against the docs: credentials, latency, Rust.
8 min
Security
The attacker behind the Onamae.com Desktop Cloud breach demanded money from GMO and leaked claimed user data. A timeline from the August intrusion to GMO's November 2 press event.
12 min
AI
Apollo chief economist Torsten Slok says AI has brought no financial payoff to large US companies outside tech. This post separates what his five charts show from what they don't.
6 min
AI
OpenAI published ChatGPT work-use data for about 100 countries in 165 work activities. It counts messages and excludes business plans, unlike Japanese surveys of people and firms.
7 min
Microsoft 365
Timing for Work IQ's Human in the loop, Agent Policy Governance and Real Time Transcripts, compared with today's tenant-wide policy, Copilot Studio approvals and Teams APIs.
7 min
AI
Claude Duet, a Claude agent in research preview since October 21, 2026, reaches out first and keeps working while you're away. How it works, approvals, usage limits, gaps and dots.
8 min
Security
Exploitation of the PHP flaws php2shell was seen right after the October 21 fixes. How to confirm the fix landed, gaps updating leaves, Cloudflare's WAF, and signs of intrusion.
6 min
AI
Anthropic's Applied AI team folded specialist subagents back into enterprise agents' main loop. Three reasons to keep one, where the work moves, and what the migration kit does.
10 min
Microsoft 365
Azure's sovereign AI framing and October 20 update: Foundry Local and Microsoft 365 Local status, why Copilot in-country processing in Japan moved to 2028, and offline conditions.
8 min
EventsAI
Findy's free AI Engineering Summit Tokyo 2026 winter edition runs December 14 and 15 in Hamamatsucho. Who speaks on which day, what is streamed, and how it differs from June.
6 min
AI
Everything Claude Code ships 311 skills. I compared installing it all as a plugin with installing only what you need: per-turn tokens, when rules load, and what --skills installs.
6 min
AI
I read the 123 entries in knowledge-work-plugins, a role-based plugin catalog built for Cowork, then installed some in Claude Code to check always-on tokens and MCP connections.
6 min
AI
REA, an MCP server giving reverse engineering tools to AI agents, drew GitHub stars fast in October. Star counts, repo dates, what Ghidra and others do, and a static Electron test.
7 min
AI
A viral post asked why a beer shop sign read as AI in a second. I map cues like broken text, mismatched styles, and color to research, and to what C2PA and SynthID can verify.
8 min
Security
PHP fixed two flaws allowing unauthenticated code execution on October 21, 2026. How to tell if a site runs PHP and is affected, risk by setup, update order and mitigation limits.
10 min
ServiceNow
ServiceNow's "Four Zeros," posted October 19, 2026, is not a product but a guide to ordering ITSM, ITOM, and CMDB. The four goals, the phases and gates, and open questions.
9 min
AI
OR-License tests whether AI agents can turn client documents into feasible plans on 31 tasks from paid optimization projects. No model matched experts; infeasible plans cost most.
6 min
ServiceNowAI
ServiceNow AI Control Tower puts every AI in one inventory to approve, monitor, stop and track cost (GA May 2025). Five pillars, roles, licenses, Agent 365 and unknowns.
9 min
ServiceNowSecurity
ServiceNow closed its about $7.75 billion Armis deal in April 2026. How Armis finds devices agentlessly, why ServiceNow bought it, six Autonomous Security product lines, unknowns.
9 min
AI
Since v2.1.292, Claude Code can set effort each time it launches a subagent. I measured time and estimated cost with search subagents at low and with the docs' "time matters" line.
5 min
Infrastructure
Keeping VDI would have made the server renewal cost over 4x the last one, so Toride City moved 350 LGWAN desktops to a Remote PC Array. What changed, and what it takes on.
7 min
AI
HKUDS's CLI-Anything generates agent-ready CLIs for GIMP, LibreOffice, and more. I read its code: how generation works, the CLI-Hub registry, and what to check before installing.
6 min
AI
Claude writes about 90% of Claude Code. I look at the team's 7 review layers, risk-based approvals, and fixes for endless AI review loops, then at doing the same with Code Review.
9 min
AI
I read the 315 plugins in claude-plugins-official, Claude Code's default catalog, and how Anthropic reviews them. Hooks run outside the sandbox; token estimates skip hook costs.
6 min
ServiceNowSecurity
ServiceNow AI Passport (announced October 15, 2026) gives each AI agent a signed credential checked at runtime. How it differs from AI Gateway, NVIDIA's role, and open questions.
7 min
AI
Felix Rieseberg, who leads Claude.ai work at Anthropic, splits AI ambition into speed, altitude, depth and range, and explains the last three with a QBR example and cautions.
9 min
EventsAI
Online Claude Code hackathon for Japan residents, Nov 2 12:00 to Nov 10 14:00 JST. Applications closed Oct 26. 250 selected seats; up to six teams share $50,000 in API credits.
7 min
AI
A walk through the 158 skills in google/skills, the Agent Skills library Google publishes on GitHub: what they can do, how to install them, and what to check first.
6 min
AI
A September 29, 2026 paper from Meta, UW, MIT and others lets a model edit its own context as a file. How it works, the results, and ways to try it, from Pi to Claude Code.
11 min
ServiceNow
ServiceNow Impact app v12 shipped October 8, 2026, and Platform Health scores change on the next scan. How to compare old scores, where sources disagree, and other changes.
9 min
AI
AgentGarten renders code-defined worlds as video, and its agents improved in four rounds using text notes. Why that can't be compared directly with 2019's hide-and-seek.
6 min
AI
AI coding agents add code, but completed work shows no significant gain. A Harvard paper and an Undo survey show the review and comprehension load; this post then weighs metrics.
6 min
AI
From Nov 12, 2026, Anthropic's usage policy bans sustained, needless abuse of its models. What counts, what does not, and what happens when Claude ends a chat.
8 min
AI
OpenAI posted 722 math papers from an unreleased model to GitHub and withdrew 3. Only 42% of main results are in Lean; some Lean code proves less than claimed.
7 min
Microsoft 365
Fujitsu expanded Microsoft 365 Copilot to about 100,000 licenses and reports a usage rate above 90%. What the case suggests about driving generative AI adoption at a company.
8 min
ServiceNow
How to upgrade to ServiceNow Brazil: upgrade paths from Australia, Zurich or Yokohama, the official 7 stages, testing, the 10-day rollback limit and year-end timing.
7 min
ServiceNow
Now Assist bills in assists, not tokens: 1 for a simple skill, 25 to 150 for agentic workflows. Where to check usage by skill, and settings that cap overuse.
9 min
Security
About 1.6 million license images and other ID documents leaked from Times Car. Can an image alone open an account or loan? Misuse risks and countermeasures, from published sources.
7 min
AI
Microsoft's AX definition and its SharePoint Framework measurements: how to gauge usability for AI agents, and where teams using Claude Code or GitHub Copilot can start.
8 min
AI
To comply with the EU AI Act, OpenAI will embed an invisible watermark (textGrain) in EU ChatGPT and Codex text. How it works, where detection falls short, and uses for companies.
6 min
SecurityAI
A Florida woman was arrested over an attack threat she wrote to Claude as a diary. I check how it was reported, plus Anthropic's terms, privacy policy and retention.
6 min
AI
Cowork's features (folder work, built-in browser, scheduled runs, plugins, projects) and official use cases, plus the September 2026 merge with Chat, pricing and safety.
12 min
AI
Airbnb's CTO described putting AI into internal work first, then reshaping the customer experience. Checked against earnings material, six approaches other companies can apply.
9 min
AI
PoS from the paper 'Beyond Memory' gives long-running agents an explicit belief state, not a summarized history. It beat the methods it was compared with, at a higher token cost.
10 min
ServiceNow
ServiceNow announced Flow, an AI service desk, on October 1, 2026: how it handles requests, how it differs from ITSM, pricing, and what readers in Japan should check.
7 min
AI
OpenAI's dots, announced September 29, 2026, are ChatGPT agents with their own cloud computer that keep working between conversations: differences, permissions, rules and plans.
8 min
Microsoft 365
The Copilot Studio skill catalog on the Microsoft 365 roadmap: what is known, how skills are built today, and what to settle before sharing skills across an organization.
6 min
9 notes
Security
29 data breaches disclosed in Japan in September 2026, and the 10 with a known cause: six exploited flaws, mostly in VPNs, admin logins and file uploads.
9 min
AISecurity
Researchers tested nine conversational AIs on web and Android in May 2026. All embedded ad and analytics vendors, and some sent conversation URLs and titles to third parties.
7 min
Microsoft 365
Microsoft's new Copilot billing model by case: what the monthly license covers, what uses Copilot Credits, and what is metered without a license, plus a pre-deployment inventory.
8 min
Microsoft 365
Copilot in SharePoint reaches general availability: how it differs from earlier SharePoint, the license versus Copilot Credits line, where it helps, and what to review first.
9 min
SecurityAI
An independent researcher reports OpenAI agents scanned a UN statistics site over 16,000 times in 2026, using double URL encoding and URL scanner relays to get past limits.
9 min
AI
How to check token usage and estimated cost with Claude Code's /usage, and how to choose among effort levels (low to max), based on Anthropic's official documentation.
5 min
Microsoft 365AI
Microsoft's Copilot reorganization (Home / Code / Autopilot) by role, pricing and rollout schedule, with governance questions for large enterprises.
6 min
ServiceNow
ServiceNow Brazil release dates: Early Availability from September 24, 2026, GA planned for November 2026. Key changes such as ServiceNow Otto, Lux and HR Service Delivery (HRSD).
12 min
Career
This blog is my place for output: looking into and trying out services, technologies and products in the news, plus my takeaways from meetups and tech books.
1 min
