The Google Open Source Software Vulnerability Reward Program team is increasingly concerned about the low quality of some AI-generated bug submissions, with many including hallucinations about how a vulnerability can be triggered or reporting bugs with little security impact.
The leap from a "functional" network to a professional-grade infrastructure is the difference between a dirt path and a multi-lane highway. As we integrate more high-bandwidth technology-from 8K streaming to AI-driven home security-the "consumer-grade" hardware typically provided by service providers is reaching its breaking point.
There were specialists monitoring dashboards, tuning AI behavior, debugging API failures, and iterating on knowledge workflows. One team member who had started their career handling customer questions over chat and email (resetting passwords, explaining features, troubleshooting one-off issues, and escalating bugs) was now writing Python scripts to automate routing. Another was building quality-scoring models for the company's AI agent. This seemed markedly different from the hyperbole I'd been hearing about customer support roles going away in large part due to AI.
This week, meet a reader we'll Regomize as "Fred" who told us that in the 1990s he worked in tech support for a government agency. "The work was pretty routine, but being in that particular environment was never dull... at least not for me," he told The Register. To prove his point, Fred told us that one day a user he said we should refer to as "Emily" complained that her word processor had lost the ability to print on the resident HP LaserJet.
It was the time of Novell networks, RG58 cables, and bulky tower PCs. It was also a time before the telemarketer's IT department employed specialists. Carter and his two colleagues - boss Mike and part-time student Stefan - therefore handled tasks ranging from programming to support, and everything in between.
In our own environment, over 90% of targeted Level 1 volume is handled autonomously, with resolution rates above 99% for those categories and materially faster than human-only workflows. It accomplishes this by having the Autonomous Workforce operate on top of the live configuration management database (CMDB), active workflows, policy engines, approval chains, and real transaction history - all updated in real time every time a ticket closes, a workflow executes, or a policy changes.
But are things getting worse? According to Register readers, and the company's own release health dashboard, the answer has to be yes. It isn't just you. The frequency of emergency out-of-band releases for the company's operating systems has been rapidly increasing to the point where, for every Patch Tuesday update, there'll likely be at least one out-of-band patch to fix whatever got broken.
"A floor manager responsible for production asked me to fix his PC, which was so slow he could literally make a coffee in the time between double-clicking an icon and having the program open," Parker told On Call. The manager's PC was only a year old and ran Windows XP, a combo that at the time of this tale should have made for decent performance.
This week, meet a reader we'll Regomize as "Curt" who once worked as IT security manager at a company where the helpdesk manager routinely ignored company policy by not logging out of his PC. The machine sat there ready for use, instead of reverting to a password-protected screensaver that could only be dispelled by pressing Ctrl-Alt-Del to spawn a login dialog.
Sudo, for those not familiar with Unix systems, is a command-line utility that allows authorized users to run specific commands as another user, typically the superuser, under tightly controlled policy rules. It is a foundational component of Unix and Linux systems: without tools like sudo, administrators would be forced to rely more heavily on direct root logins or broader privilege escalation mechanisms, increasing both operational risk and attack surface.
"BeyondTrust Remote Support (RS) and certain older versions of Privileged Remote Access (PRA) contain a critical pre-authentication remote code execution vulnerability," the company said in an advisory released February 6, 2026. "By sending specially crafted requests, an unauthenticated remote attacker may be able to execute operating system commands in the context of the site user." The vulnerability, categorized as an operating system command injection, has been assigned the CVE identifier CVE-2026-1731.
The issue focuses on how Windows handles these directories for specific user sessions. Because the kernel creates a DOS device object directory on demand, rather than at login, it cannot check whether the user is an admin during the creation process. Unlike UAC, Administrator Protection uses a hidden shadow admin account whose token handle can be returned by the system when calling the NtQueryInformationToken API function.