OpenAI put computer-use agents into hospital software on 16 November 2026 and reported reviewers rating 99% of 325 EHR-inbox tasks clinically safe. Give that rate a denominator — about 3,220 messages per clinician-year — and it describes 32 to 90 unsafe dispositions a year, from an evidence base the size of one clinician-month. The value claim has the same missing denominator: a cohort study of 983,154 portal replies found an AI drafter 29.5% faster on the messages it touched and 8.6% slower on the queue as a whole, because a draft creates review work on messages nobody needed to read.
OSWorld vs WindowsAgentArena vs AndroidWorld vs BrowserGym: the verifier is the benchmark
These four get quoted as four measurements of the same ability, and they are four different objects. Two of them do not own a verifier at all: WindowsAgentArena vendors OSWorld's metric library wholesale, and BrowserGym is a harness that imports whatever the wrapped benchmark wrote. The axis that decides whether a score transfers is not the operating system or the task count — it is what the check reads. OSWorld compares a saved file against a gold copy after pressing Ctrl+S itself; AndroidWorld reads device state on a randomised instance no other run shares; BrowserGym's reward float carries a binary page check in one row and a token-bag F1 in the next.
Ask which computer-use tool Claude Opus 5.5 supports and the documentation has two answers: on the Claude API it accepts only `computer_toolset_20260801` and errors on anything earlier; on Amazon Bedrock it accepts the earlier `computer_20251124`. The two differ in shape and in screenshot ceiling — 4,784 visual tokens at 2,576px versus about 1.15 megapixels — so one model string runs two agents with different grounding accuracy and different per-step cost, and no model on the Claude API accepts both.
The model moved onto the laptop; the trust boundary did not
Perplexity Computer now runs its model, planning and file processing entirely on an Apple-silicon Mac, with no inference request leaving the device and no credit charged. That answers data residency and barely touches exfiltration: the agent still reads Mail and Messages on permission, still sits on your files, and still holds an approved cloud route in the same app. Local inference removes the provider's request log — and with it the kill switch, the egress control and the central audit trail.
UI-TARS vs Agent S3 vs cua vs OpenAdapt: you are picking a layer, not an agent
Agent S3 recommends UI-TARS as its grounding model, so two of these four are halves of one stack — and five numbers published under the name “OSWorld” span 63 points. Pick on which of four layers you are missing, and on the one axis that actually separates them: whether a model decides where to click every time the task runs.
Safari MCP vs Chrome DevTools MCP vs Playwright MCP vs extension agents
Tool counts decide nothing here. The axis that determines both whether a browser agent can do the job and how bad a hostile page gets is which session it holds — an isolated automation context, a dedicated profile quietly accumulating logins, or your own signed-in browser. Both browser vendors that shipped an MCP server this year deliberately kept your own session out of it, which is why neither does the agentic-shopping demo everyone expected.
Gemini Spark moved into your Chrome profile, and the handback is on the wrong line
Google’s agent now drives the Chrome you are logged into, with your saved passwords, and hands control back for payments. Payment is the one action with a chargeback window; the mailbox read, the data copied out and the recovery address changed are all on the unattended side of that line.
Browserbase vs Steel vs Hyperbrowser vs Anchor Browser: You Are Choosing Who Holds the Session
All four speak CDP, so the automation code ports in a day and the SDK comparison decides nothing. The real choice is who holds the logged-in profile, the credentials that recreate it and the exit IP whose reputation you inherit — plus the fact that the latency spread between them is entirely control plane.
Atlas Shuts Down on 9 August. Agentic Browsing Just Split Into Three.
OpenAI is retiring the ChatGPT Atlas browser nine months after launch and moving its capabilities into a Chrome extension, an in-app browser and a server-side cloud browser. That is not a retreat from agentic browsing — it is the admission that a browser agent never needed a browser. What it needed was proximity to an authenticated session, and the three replacement surfaces are three different answers to whose session it borrows.
Browser-Use vs Stagehand vs Skyvern vs Playwright MCP: Four Answers to How an LLM Should Drive a Web Page
When there is no API, an agent has to drive the browser itself — and four open-source projects disagree on how it should see the page. browser-use reads the DOM, Skyvern looks at pixels, Stagehand lets you dial between code and AI, and Playwright MCP is not an agent at all but the standard browser-tool layer any model can call. Picking one is really two decisions: Python or TypeScript, and a framework or an MCP server.
Claude Computer Use (post-Vercept) vs Codex Background CU vs Operator vs Gemini: Four Bets on Letting AI Drive the Mouse
72.5% on OSWorld is the new floor, not a milestone — and three labs have made architecturally opposite bets on where the mouse should live. Pick the wrong one and you fight your sandbox forever; pick the right one and the model does in two minutes what your RPA stack does in two weeks.