4 ms·

What does “our computer” mean when it is not owned by you, but issued to you to perform a task with by your employer? Does that also apply to the operator at a

What does “our computer” mean when it is not owned by you, but issued to you to perform a task with by your employer? Does that also apply to the operator at a switchboard in a nuclear missile launch facility?

z3phyr 2y ago
Does the switchboard in a nuclear missile launch facility run Crowdstrike? I picture it as a high quality analog circuit board that does 1 thing and 1 thing only. No way to run anything else. Globally networked personal computers were kind of cultural revolution against the setting you describe. Everyone had their own private compute and compute time and everyone could share their own opinion. Computers became our personal extensions. This is what IBM, Atari, Commodore, Be, Microsoft and Apple (and later desktop Linux) sold. Now given this ideology, can a company own my limbs? If not, they can't own my computers.
derefr 2y ago
> What does “our computer” mean when it is not owned by you, but issued to you to perform a task with by your employer? Well, presuming that: 1. the employee is issued a computer, that they have possession of even if not ownership (i.e. they bring the computer home with them, etc.) 2. and the employee is required to perform creative/intellectual labor activities on this computer — implying that they do things like connecting their online accounts to this computer; installing software on this computer (whether themselves or by asking IT to do it); doing general web-browsing on this computer; etc. 3. and where the extent of their job duties, blurs the line between "work" and "not work" (most salaried intellectual-labor jobs are like this) such that the employee basically "lives in" this computer, even when not at work... 4. ...to the point that the employee could reasonably conclude that it'd be silly for them to maintain a separate "personal" computer — and so would potentially sell any such devices (if they owned any), leaving them dependent on this employer-issued computer for all their computing needs... ...then I would argue that, by the same chain of reasoning as in the GP post, employers should not be legally permitted to “issue” employees such devices. Instead, the employer should either purchase such equipment for the employee, giving it to them permanently as a taxable benefit; or they should require that the employee purchase it themselves, and recompense them for doing so. Cyberpunk analogy: imagine you are a brain in a vat. Should your employer be able to purchase an arbitrary android body for you; make you use it while at work; and stuff it full of monitoring and DRM? No, that'd be awful. Same analogy, but with the veil stripped off: imagine you are paraplegic. Should your employer be allowed to issue you an arbitrary specific wheelchair, and require you to use it at work, and then monitor everything you do with it / limit what you can do with it because it’s “theirs”? No, that’d be ridiculous. And humanity already knows that — employers already can't do that, in any country with even a shred of awareness about accessibility devices. The employer — or very much more likely, the employer's insurance provider — just buys the person the chair. And then it's the employee's chair. And yes, by exactly the same logic, this also means that issuing an employee a company car should be illegal — at least in cases where the employee lives in a non-walkable area, and doesn't already have another car (that they could afford to keep + maintain + insure); and/or where their commute is long enough that they'd do most non-employment-related car-requiring things around work and thus using their company car. Just buy them a car. (Or, if you're worried they might run away with it, then lease-to-own them a car — i.e. where their "equity in the car" is in the form of options that vest over time, right along-side any equity they have in the company itself.) > Does that also apply to the operator at a switchboard… Actually, no! Because an operator of a switchboard is not a “user” of the computer that powers the switchboard, in the same sense that a regular person sitting at a workstation is a "user" of the workstation. The system in this case is a “kiosk computer”, and the operator is performing a prescribed domain-specific function through a limited UX they’re locked into by said system. The operator of a nuclear power plant is akin to a customer ordering food from a fast-food kiosk — just providing slightly more mission-critical inputs. (Or, for a maybe better analogy: they're akin to a transit security officer using one of those scanner kiosk-handhelds to check people's tickets.) If the "computer" the nuclear-plant operator was operating, exposed a purely electromechanical UX rather than a digital one — switches and knobs and LEDs rather than screens and keyboards[1] — then nothing about the operator's workflow would change. Which means that the operator isn't truly computing with the computer; they're just interacting with an interface that happens to be a computer. [1] ...which, in fact, "modern" nuclear plants are. The UX for a nuclear power plant control-center has not changed much since the 1960s; the sort of "just make it a touchscreen"-ification that has infected e.g. automotive has thankfully not made its way into these more mission-critical systems yet. (I believe it's all computers under the hood now, but those computers are GPIO-relayed up to panels with lots and lots of analogue controls. Or maybe those panels are USB HID devices these days; I dunno, I'm not a nuclear control-systems engineer.) Anyway, in the general case, you can recognize these "the operator is just interacting with an interface, not computing on a computer" cases because: • The machine has separate system administrators who log onto it frequently — less like a workstation, more like a server. • The machine is never allowed to run anything other than the kiosk app (which might be some kind of custom launcher providing several kiosk apps, but where these are all business-domain specific apps, with none of them being general-purpose "use this device as a computer" apps.) • The machine is set up to use domain login rather than local login, and keeps no local per-user state; or, more often, the machine is configured to auto-login to an "app user" account (in modern Windows, this would be a Mandatory User Profile) — and then the actual user authentication mechanism is built into the kiosk app itself. • Hopefully, the machine is using an embedded version of the OS, which has had all general-purpose software stripped out of it to remove vulnerability surface.
- derefr 2y ago
  Tangent — a question you didn't ask, but I'll pretend you did: > If employers allowed employees to "bring their own devices", and then didn't force said employees to run MDM software on those devices, then how in the world could the employer guarantee the integrity of any line-of-business software the employee must run on the device; impose controls to stop PII + customer-shared data + trade secrets from being leaked outside the domain; and so forth? My answer to that question: it's safe to say that most people in the modern day are fine with the compromise that your device might be 100% yours most of the time; but, when necessary — when you decide it to be so — 99% yours, 1% someone else's. For example, anti-cheat software in online games. The anti-cheat logic in online games, is this little nugget of code that runs on a little sub-computer within your computer (Intel SGX or equivalent.) This sub-computer acts as a "black box" — it's something the root user of the PC can't introspect or tamper with. However: • Whenever you're not playing a game, the anti-cheat software isn't loaded. So most of the time, your computer is entirely yours. • You get to decide when to play an online game, and you are explicitly aware of doing so. • When you are playing an online game, most of your computer — the CPU's "application cores", and 99% of the RAM — is still 100% under your control. The anti-cheat software isn't actually a rootkit (despite what some people say); it can't affect any app that doesn't explicitly hook into it. • In a brute-force sense, you still "control" the little sub-computer as well — in that you can force it to stop running whatever it's running whenever you want. SGX and the like aren't like Intel's Management Engine (which really could be used by a state actor to plant a non-removable "ring -3" rootkit on your PC); instead, SGX is more like a TPM, or an FPGA: it's something that's ultimately controlled by the CPU from ring 0, just with a very circumscribed API that doesn't give the CPU the ability to "get in the way" of a workload once the CPU has deployed that workload to it, other than by shutting that workload off. As much as people like Richard Stallman might freak out at the above design, it really isn't the same thing as your employer having root on your wheelchair. It's more like how someone in a wheelchair knows that if they get on a plane, then they're not allowed to wheel their own wheelchair around on the plane, and a flight attendant will instead be doing that for them. How does that translate to employer MDM software? Well, there's no clear translation currently, because we're currently in a paradigm that favors employer-issued devices. But here's what we could do: • Modern PCs are powerful enough that anything a corporation wants you to do, can be done in a corporation-issued VM that runs on the computer. • The employer could then require the installation of an integrity-verification extension (essentially "anti-cheat for VMs") that ensures that the VM itself, and the hypervisor software that runs it, and the host kernel the hypervisor is running on top of, all haven't been tampered with. (If any of them were, then the extension wouldn't be able to sign a remote-attestation packet, and the employer's server in turn wouldn't return a decryption key for the VM, so the VM wouldn't start.) • The employer could feel free to MDM the VM guest kernel — but they likely wouldn't need to, as they could instead just lock it down in much-more-severe ways (the sorts of approaches you use to lock down a server! or a kiosk computer!) that would make a general-purpose PC next-to-useless, but which would be fine in the context of a VM running only line-of-business software. (Remember, all your general-purpose "personal computer" software would be running outside the VM. Web browsing? Outside the VM. The VM is just for interacting with Intranet apps, reading secure email, etc.) (Why yes, I am describing https://en.wikipedia.org/wiki/Multilevel_security https://en.wikipedia.org/wiki/Multilevel_security.)

z3phyr 2y ago

Does the switchboard in a nuclear missile launch facility run Crowdstrike? I picture it as a high quality analog circuit board that does 1 thing and 1 thing only. No way to run anything else. Globally networked personal computers were kind of cultural revolution against the setting you describe. Everyone had their own private compute and compute time and everyone could share their own opinion. Computers became our personal extensions. This is what IBM, Atari, Commodore, Be, Microsoft and Apple (and later desktop Linux) sold. Now given this ideology, can a company own my limbs? If not, they can't own my computers.

derefr 2y ago

> What does “our computer” mean when it is not owned by you, but issued to you to perform a task with by your employer? Well, presuming that: 1. the employee is issued a computer, that they have possession of even if not ownership (i.e. they bring the computer home with them, etc.) 2. and the employee is required to perform creative/intellectual labor activities on this computer — implying that they do things like connecting their online accounts to this computer; installing software on this computer (whether themselves or by asking IT to do it); doing general web-browsing on this computer; etc. 3. and where the extent of their job duties, blurs the line between "work" and "not work" (most salaried intellectual-labor jobs are like this) such that the employee basically "lives in" this computer, even when not at work... 4. ...to the point that the employee could reasonably conclude that it'd be silly for them to maintain a separate "personal" computer — and so would potentially sell any such devices (if they owned any), leaving them dependent on this employer-issued computer for all their computing needs... ...then I would argue that, by the same chain of reasoning as in the GP post, employers should not be legally permitted to “issue” employees such devices. Instead, the employer should either purchase such equipment for the employee, giving it to them permanently as a taxable benefit; or they should require that the employee purchase it themselves, and recompense them for doing so. Cyberpunk analogy: imagine you are a brain in a vat. Should your employer be able to purchase an arbitrary android body for you; make you use it while at work; and stuff it full of monitoring and DRM? No, that'd be awful. Same analogy, but with the veil stripped off: imagine you are paraplegic. Should your employer be allowed to issue you an arbitrary specific wheelchair, and require you to use it at work, and then monitor everything you do with it / limit what you can do with it because it’s “theirs”? No, that’d be ridiculous. And humanity already knows that — employers already can't do that, in any country with even a shred of awareness about accessibility devices. The employer — or very much more likely, the employer's insurance provider — just buys the person the chair. And then it's the employee's chair. And yes, by exactly the same logic, this also means that issuing an employee a company car should be illegal — at least in cases where the employee lives in a non-walkable area, and doesn't already have another car (that they could afford to keep + maintain + insure); and/or where their commute is long enough that they'd do most non-employment-related car-requiring things around work and thus using their company car. Just buy them a car. (Or, if you're worried they might run away with it, then lease-to-own them a car — i.e. where their "equity in the car" is in the form of options that vest over time, right along-side any equity they have in the company itself.) > Does that also apply to the operator at a switchboard… Actually, no! Because an operator of a switchboard is not a “user” of the computer that powers the switchboard, in the same sense that a regular person sitting at a workstation is a "user" of the workstation. The system in this case is a “kiosk computer”, and the operator is performing a prescribed domain-specific function through a limited UX they’re locked into by said system. The operator of a nuclear power plant is akin to a customer ordering food from a fast-food kiosk — just providing slightly more mission-critical inputs. (Or, for a maybe better analogy: they're akin to a transit security officer using one of those scanner kiosk-handhelds to check people's tickets.) If the "computer" the nuclear-plant operator was operating, exposed a purely electromechanical UX rather than a digital one — switches and knobs and LEDs rather than screens and keyboards[1] — then nothing about the operator's workflow would change. Which means that the operator isn't truly computing with the computer; they're just interacting with an interface that happens to be a computer. [1] ...which, in fact, "modern" nuclear plants are. The UX for a nuclear power plant control-center has not changed much since the 1960s; the sort of "just make it a touchscreen"-ification that has infected e.g. automotive has thankfully not made its way into these more mission-critical systems yet. (I believe it's all computers under the hood now, but those computers are GPIO-relayed up to panels with lots and lots of analogue controls. Or maybe those panels are USB HID devices these days; I dunno, I'm not a nuclear control-systems engineer.) Anyway, in the general case, you can recognize these "the operator is just interacting with an interface, not computing on a computer" cases because: • The machine has separate system administrators who log onto it frequently — less like a workstation, more like a server. • The machine is never allowed to run anything other than the kiosk app (which might be some kind of custom launcher providing several kiosk apps, but where these are all business-domain specific apps, with none of them being general-purpose "use this device as a computer" apps.) • The machine is set up to use domain login rather than local login, and keeps no local per-user state; or, more often, the machine is configured to auto-login to an "app user" account (in modern Windows, this would be a Mandatory User Profile) — and then the actual user authentication mechanism is built into the kiosk app itself. • Hopefully, the machine is using an embedded version of the OS, which has had all general-purpose software stripped out of it to remove vulnerability surface.

derefr 2y ago
Tangent — a question you didn't ask, but I'll pretend you did: > If employers allowed employees to "bring their own devices", and then didn't force said employees to run MDM software on those devices, then how in the world could the employer guarantee the integrity of any line-of-business software the employee must run on the device; impose controls to stop PII + customer-shared data + trade secrets from being leaked outside the domain; and so forth? My answer to that question: it's safe to say that most people in the modern day are fine with the compromise that your device might be 100% yours most of the time; but, when necessary — when you decide it to be so — 99% yours, 1% someone else's. For example, anti-cheat software in online games. The anti-cheat logic in online games, is this little nugget of code that runs on a little sub-computer within your computer (Intel SGX or equivalent.) This sub-computer acts as a "black box" — it's something the root user of the PC can't introspect or tamper with. However: • Whenever you're not playing a game, the anti-cheat software isn't loaded. So most of the time, your computer is entirely yours. • You get to decide when to play an online game, and you are explicitly aware of doing so. • When you are playing an online game, most of your computer — the CPU's "application cores", and 99% of the RAM — is still 100% under your control. The anti-cheat software isn't actually a rootkit (despite what some people say); it can't affect any app that doesn't explicitly hook into it. • In a brute-force sense, you still "control" the little sub-computer as well — in that you can force it to stop running whatever it's running whenever you want. SGX and the like aren't like Intel's Management Engine (which really could be used by a state actor to plant a non-removable "ring -3" rootkit on your PC); instead, SGX is more like a TPM, or an FPGA: it's something that's ultimately controlled by the CPU from ring 0, just with a very circumscribed API that doesn't give the CPU the ability to "get in the way" of a workload once the CPU has deployed that workload to it, other than by shutting that workload off. As much as people like Richard Stallman might freak out at the above design, it really isn't the same thing as your employer having root on your wheelchair. It's more like how someone in a wheelchair knows that if they get on a plane, then they're not allowed to wheel their own wheelchair around on the plane, and a flight attendant will instead be doing that for them. How does that translate to employer MDM software? Well, there's no clear translation currently, because we're currently in a paradigm that favors employer-issued devices. But here's what we could do: • Modern PCs are powerful enough that anything a corporation wants you to do, can be done in a corporation-issued VM that runs on the computer. • The employer could then require the installation of an integrity-verification extension (essentially "anti-cheat for VMs") that ensures that the VM itself, and the hypervisor software that runs it, and the host kernel the hypervisor is running on top of, all haven't been tampered with. (If any of them were, then the extension wouldn't be able to sign a remote-attestation packet, and the employer's server in turn wouldn't return a decryption key for the VM, so the VM wouldn't start.) • The employer could feel free to MDM the VM guest kernel — but they likely wouldn't need to, as they could instead just lock it down in much-more-severe ways (the sorts of approaches you use to lock down a server! or a kiosk computer!) that would make a general-purpose PC next-to-useless, but which would be fine in the context of a VM running only line-of-business software. (Remember, all your general-purpose "personal computer" software would be running outside the VM. Web browsing? Outside the VM. The VM is just for interacting with Intranet apps, reading secure email, etc.) (Why yes, I am describing https://en.wikipedia.org/wiki/Multilevel_security https://en.wikipedia.org/wiki/Multilevel_security.)