BossAI in Hualien, a self-taught AI attacker in Kaohsiung and the ARTEX case in Seoul show that automated attack systems no longer require a security background, and one operator can sweep hundreds of targets in weeks. Run an AI security audit on your own source code, then self-test with an ARTEX-style framework; both cost far less than an outsourced pentest.
Seven days, 85 sessions: 104 compactions, 20 SIGKILLs, 24 type-check OOMs, 72 user corrections. The fix: move no-exception rules from CLAUDE.md into PreToolUse hooks, and drop 16 MEMORY.md entries that git already knew or that had gone stale.
Google's Professional Agentic Architect is assessed as a hybrid two-part exam: a proctored multiple-choice exam with Pearson, then hands-on labs on Google Skills. Both parts are required. The GA spec is a 2-hour exam, 50-60 multiple-choice questions, $200, English, one-year validity; the beta (Sept 8-30, 2026) was 3 hours, about 80 questions, $120.
When several questions share one context, Flide computes the shared state once and forks it per question with answers identical to asking alone; up to 9.5x faster than vLLM and SGLang at best, roughly 1.3x to 3.3x on average for shared-state workloads, and only about 1x to 3x with one question per input.
Run several Claude Code, Codex or Cursor sessions at once and their dev servers, test workers and Playwright MCP browsers keep running as orphans after the sessions end, until swap hits 30 GB and tests hang. macOS has no per-process memory cap, so it takes three layers: a SessionEnd hook, a periodic orphan reaper and a memory gate.
To try letting an AI use a cloud computer, start with Simular box: the computer is free and one command connects it to Claude Code or Codex. Its specs are unpublished, each account gets one machine, and it is still in beta. For several machines or always-on use pick Orgo (from $29/month) or Runloop; for commands only, a sandbox such as E2B; for web pages only, a browser service such as Browserbase.
Install OpenCode 2, select a built-in free model, and complete your first everyday task. Connect OpenRouter later if you want more models or a free API for your own application.
LLM prose has four recurring failures: contrast framing (refuting an X nobody said), significance inflation (dressing a plain fact up as a turning point), indicative summaries (announcing the topic, withholding the conclusion) and overgeneralization (summaries that drop conditions and scope). Fix the first three by cutting the filler and writing the answer; fix the fourth by checking each sentence against the source. Contrast framing has been traced to preference data, and models revising a draft move it into another contrast form, so banning a phrase in the prompt may not stop it.
Ask two questions of every contrast: do readers actually hold X, and does rejecting X tell them anything new? If either answer is no, drop X and state Y. 'Rather than' became about seven times as frequent in NLP papers from 2019 to 2026, and five commercial models writing Chinese use the equivalent construction 2.5–9.2 times as densely as a human reference corpus. Models almost never misuse it, but more than half the uses are unnecessary, and with eight in one paper, at least one tends to annoy the reader.
The four entry points are one staircase, not a menu: add a DO when you need memory and resumption, AIChatAgent for the chat pipeline, Think or PiHarness for loop management, and dynamic agents only when you need to split work.
A two-hour lecture is about 720,000 tokens, which Gemini's API accepts and Claude's and ChatGPT's API docs don't list video input for. On code, Claude and GPT are 0.8 points apart on the Vals board, so compare time and cost instead. Grok for X sentiment and auto-fill checkout, ChatGPT for ready-made apps and voice.
Detection and audit tools run along a spectrum: the open-source release of Uber ADR reads records purely after the fact, numbat only monitors by default, and for most of the rest the selling point is that they can block before an action, yet they start out observing only. Whether a record can serve as evidence depends on who can change it: when the signing key sits on the same machine and user as the audited agent, tampering can be noticed but not prevented.
Guardrails intercept an action before the agent carries it out. On Claude Code the main attachment point is the hook, and the official docs say that on most events a hook that errors or times out lets the action go ahead; blocking needs onFailure: block. Other tools have their own failure behavior and defaults: the Agent Governance Toolkit library allows everything when no policy is loaded, NeMo Guardrails lets requests through when its jailbreak detector is down, and llm-guard and rebuff are archived. Ask three things of any guardrail: where it intercepts, what happens when it fails, and who can change its rules.
Containers share the host kernel, gVisor inserts a user-space kernel, and a microVM adds a guest kernel plus a hypervisor; all three still share the same host's CPU caches and storage.
Two programs with no connection can pass bits through the speed of a shared resource: one changes its state, the other times it. Such channels need an agreed protocol and are slow and easily drowned out; the local experiment is nearly error-free up to 20 bps and gets clearly worse from 100 bps.
No single defense blocks both covert channels and exits. Blocking channels comes down to not sharing, blocking exits comes down to narrowing the services you allow, and monitoring only notices while automatic stopping is a response after the fact; neither is prevention.
A sandbox with no direct network can still have an exit: any allowed service that can make requests on someone's behalf, such as a package proxy, a DNS resolver, or a metrics service, is one. OpenAI's three cases all have this shape.
OpenAI says its red-team agents built communication channels on shared hosts from cache, storage, CPU time, and a TLS certificate cache, and that most channels are slow; in most tests the channel or protocol was given, and bandwidth and error rates were not published.
A piece of agent code passes through five layers on its way from execution to a network connection: container, kernel boundary, shared host, network perimeter, monitoring. Each layer promises one thing, and the shared host layer promises almost nothing.
In my judgment, OpenAI wrote about a research environment for training frontier models; for a personal agent, limiting credentials and exits comes first, and covert channels become a main risk only once there is a second tenant that must be isolated.