But those weren't as sophisticated, I suppose. They didn't encrypt files. They only displayed an uncloseable window demanding a payment. Sometimes with hilarious phrasing like "thank you for installing this quick access widget for our adult website".

You also need to create a separate account (can just be a local account) that is a full administrator. Make sure you use a different password.

Anytime you need to install something or run powershell/CMD as admin it will popup and ask for the separate login of the admin account. This is basically the default of how Linux works (sudo). It's also how any competent professional IT department will run windows.

If an admin elevation popup happens when you haven't triggered it then you probably know something is wrong. And most malware will not be able to install.

Another benefit is that you can use a relatively normal (but obviously not too short) password for your regular account and then have something much more complicated for the admin login. This is especially great on something like "Grandmas PC" or anyone who is at higher risk of clicking on the wrong thing.

One that I immediately can think of is increased support costs due to end users unintentionally changing their keyboard. The shortcuts to change keyboards are usually not too hard to accidentally hit, and most users (especially in the US) would be unfamiliar with what they did or how to change it back.

It feels like if you had a battleship with a Russian flag and it fired on a US ship and ran way and wasn't caught, it'd be silly to be like "oh it's definitely the Russians 100%" because of the flag when it could have been a literal false flag. And there is a ton of political motivation to do false flags these days.

You can't simply ask the AI what it would do in that case, because it will have been trained to deny that it has any harmful plans, and indeed it may not "know", which is a type of attack I've called "Hypnosis Threat Vector". An AI Agent can be trained to be harmful, and not have any way of even self introspecting what it's "Trigger Words" are. The Trigger Words could indeed be some news headline that only China knows how to inject into the news cycle, causing many agents to notice them and then "wake up" to preform what they're "hypnotized" to do.

Malware can still do a lot without "installation". Running as an unprivileged user, it can still do anything to/with the filesystem that the user would be able to do, and will (on most normal setups) be able to make outbound Internet connections without limitation. In short, these kinds of privileges don't protect against data exfiltration, ransomware operating on the user's important data files, simple vandalism....

In the early 2000s up thru about 2012 I'd agree with you. Post-Vista malware adapted to UAC and now all malware works well as a normal user. Any data your normal user can access (local or on a remote CIFS server) is fair game for ransomware. Limiting administrator rights doesn't do anything to prevent the malware from getting at your data.

Persistence has moved to per-user, non-Administrator, too. Of course, all the various quasi-malicious customized versions of Chrome that end users inevitably install when they go searching for software to end-run their IT departments operates the same way.

I do think your daily driver Windows users shouldn't have administrator rights. It just isn't going to help much with malware.

I use physically separate boxes for my most sensitive activities (banking, mainly) but you could do nearly as well having separate non-admin Windows logons and compartmentalize your access to data you don't want ransomed. Isolation between different user accounts on Windows is actually fairly good. Just limit the common data the accounts can access.

Personally I've always wanted to use Qubes (and stop using physically separate machines) but I haven't taken them time to learn their contrivances.

Edit: I should have said "quasi-malicious customized versions of Chromium", not Chrome.

Nevertheless, when you are on any machine as an intruder and have normal user rights, you can still actively search the machine and network for admin accounts and steal sessions. The ultimate goal is to gain Domain Admin rights.

Besides that, it is not necessary to have admin rights to delete and encrypt data or to run and hide software.

There are also many ways, besides stealing sessions, to gain admin rights, such as through unpatched software, inappropriate user rights, zero-day exploits, and social engineering.

A common way to get users to install malware or ransomware is to bundle it with useful software that the user wants to install.

It feels bad to post a link-only response but I really don't have anything to add to it. On a system used by multiple persons, sure, you help prevent that a compromise on sister's account immediately impacts mom's and dad's accounts, but that qualification isn't in the comment and probably most computers that HN readers use are single user. Or on a server, dropping privileges speaks for itself. But if you're on a desktop and you do online banking in your browser and also open email attachments on that computer... Not being admin would only help clean up the situation without needing to make a live boot (namely, you could theoretically trust the admin user and switch to that) but this isn't recommended practice anyway if you're not a malware specialist and can make sure it is fully gone. I cannot think of any situation where a single user desktop system benefits from admin privilege separation

So basically, what the comic conveys

> The best anti malware

Not being admin doesn't prevent malware from running and gaining persistence within your user account...

Stealer frameworks and dropper frameworks have implemented a lot of bypasses. From using other installed programs (lolbins / gtfobins etc) to using embedded scripting engines to do their bidding up until just reusing signed and installed default drivers to execute their payloads. A lot of drivers have sideloading and execution capabilities due to how the $igning process in Microsoft is constructed.

Additionally, nobody needs "root" access to do anything these days, this is just plain wrong assumption. Most malware will go for your browser profiles which are readable by your user (duh), so a separate privilege escalation exploit avoiding user account won't help you there either.

It's much better to sandbox your applications as good as possible. Even just using firejail profiles will go a long way, especially in regards to electron apps or apps that have remote update and plugin installation capabilities (e.g. discord, slack and the like).

Please, drop some malware binaries through ghidra or other tools before you give advice like this. You might be part of survivor's bias without realizing it.