💾 Archived View for tim.clifford.lol › technology-usage-manifesto.gmi captured on 2023-06-14 at 13:58:09. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2023-01-29)
-=-=-=-=-=-=-
This document lays out a set of rules that I will follow in order to protect myself from the pitfalls of modern technology. Note that if any rules conflict, specific takes precedence over general.
The consequences of aims often have implicit assumptions and always have limitations. I've tried to set them all out explicitly here.
Assumptions:
Limitations:
Data Classification:
Secure: to my knowledge, there is no way another party can gain access to this data. (Of course, that doesn't mean it's impossible, it almost certainly isn't)
Obscure: There are ways that this data could be accessed, but it is not likely that it can be mined, due to some combination of obscurity, anonymisation, low signal-to-noise ratio, low volume, and/or other technical constraints. Note that modern neural networks have _incredible_ correlation power, and you should not be surprised if this data eventually becomes public. Different sources of obscure data should ideally be segregated such that correlations cannot easily be made between them.
Public: The data is easily accessible by any party.
Assumptions:
Limitations:
Comments:
Assumptions:
Limitations:
Programs which only communicate with machines I control are not subject to these conditions.
1.0) I will not consume media over a network through a means where different usages can be correlated with each other. (Aim 3)
1.0) (comment) So I can't authenticate, I can't be personally identifiable, and I must prevent fingerprinting (i.e. (1.2))
1.1) I will not access a remote server that requires authentication, unless I accept that (a) my account may be breached or deleted at any time without warning, and (b) any data I provide will be public knowledge. (Aims 1 and 2)
1.1) (comment) I may use a cloud storage provider I do not trust, provided that the data is both mirrored to a local drive/NAS and either public or protected with end-to-end encryption. File metadata must be hidden as well as file contents.
1.2) If (1.1) applies but none of the data transferred can be linked to my identity, the data shall be considered "obscure" rather than public.
1.3) Unless all transferred data is public, I will only access remote servers through Tor or similar, to prevent fingerprinting. (Aim 1)
1.3) (comment) Further anti-fingerprinting measures should also be taken when possible, e.g. modifying user agent strings to the most common option. (Aim 1)
1.4) I will not grant a remote server access to any sensing peripherals (e.g. microphone or camera), unless it is verifiably protected by client-side end-to-end encryption. (Aims 1 and 3)
1.5) If free (libre) or trivial code is downloaded from a remote server and then immediately executed (e.g. JavaScript), only _the source itself_ should be whitelisted, not the remote server, to prevent a bait-and-switch scenario. (Aims 1 and 2)
In this section, an "untrusted sandbox" (US) refers to a sandbox used for running untrusted software, and can be either a container on a hardened kernel, a VM, or a dedicated untrusted machine. Note that a dedicated machine is far more secure than the other two options.
2.0) I will not depend on any proprietary software. (Aim 2)
2.0) (comment) I may use proprietary software for ephemeral applications, or when a suitable alternative is easy to come by and the switching cost is manageable on short notice.
2.1) Proprietary software must be run in a US. (Aim 1)
2.2) USes can have either read-only access to private data, or a network connection, but never both at the same time. (Aim 1)
2.3) A US that has previously had network access can later have access to private data, but a US that has previously had access to private data can never have network access. (Aim 1)
2.4) USes should be reverted to a clean state after every use. (Aim 1)
3.0) I will use proprietary firmware only if no FOSS firmware exists, and if having no firmware would render the device useless (Aims 1, 2, and 4)
3.1) I will not purchase a device that requires proprietary firmware unless there is no other suitable device in that class (Aims 1, 2, and 4)
3.2) When transitioning from a device with proprietary firmware to a device with FOSS firmware, I will assume all data on the original device is compromised and malicious. (Aim 1)
4.0) I will not be pressured into using software without first taking the time to consider this manifesto.
4.1) If I am pressured to contravene this manifesto against my will, and I cannot resist, I will seek solutions that contravene the aims in descending order. If Aim 1 must be contravened, it should be done in a way that compromises the smallest amount of the least sensitive data possible.
4.2) When transferring data to a person who does not comply with this manifesto or a similar one, but nonetheless through a method which, on my end, does not contravene this manifesto, I will assume the data becomes "obscure" (unless it is already public). (Aims 1 and 4)
4.3) I will not use a proprietary messaging service, unless it can be adequately bridged to a FOSS client. (Aims 1 and 2)
4.4) I will not use a messaging service unless it has verifiable FOSS end-to-end encryption, except when the messages are public (this includes email unless PGP is used) (Aims 1 and 4)
4.4) (comment) i.e. components of the communication channel can be proprietary/untrustable, as long as the E2E encryption is FOSS. Ideally the entire communication channel would be decentralised in order to protect metadata (e.g. Matrix), but this is not a requirement.
5.0) I may use computers I do not trust, but I must assume any data relating to my usage of the system is public - unless my identity can be concealed, in which case the data will be considered "obscure". (Aim 1)
6.0) An organisation I work for is responsible for the work I do, but not my personal data. I need not apply this manifesto to workplace computing, except where it applies to personal data.
6.0) (comment) A clear segregation should be made between personal computing and workplace computing.
6.1) I will, where possible, encourage organisations I work for to apply similar goals to their internal practices.
┌────────────────────────────────┬────────────────────────────────────────────────┐ │ Software │ Replacement │ ╞════════════════════════════════╪════════════════════════════════════════════════╡ │ ✅ Audio Consumption (Spotify) │ Torrents? CD ripping? mp3 purchasing? │ ├────────────────────────────────┼────────────────────────────────────────────────┤ │ ✅ YouTube │ --Selfhosted-- Invidious or CloudTube │ ├────────────────────────────────┼────────────────────────────────────────────────┤ │ Traditional TV/movies │ Torrents? Blu-ray ripping? │ ├────────────────────────────────┼────────────────────────────────────────────────┤ │ ✅ Cloud Storage (backup) │ Borg + Backblaze and local storage │ ├────────────────────────────────┼────────────────────────────────────────────────┤ │ ✅ Cloud Storage (continuous) │ VPS/equivalent + E2E encryption + SFTP │ ├────────────────────────────────┼────────────────────────────────────────────────┤ │ Gaming │ Public metadata + containerised Steam? Hmm. │ ├────────────────────────────────┼────────────────────────────────────────────────┤ │ ✅ MessageEase keyboard │ --Build my own-- use qwerty and be sad │ ├────────────────────────────────┼────────────────────────────────────────────────┤ │ DnD beyond │ Find alternative? Build my own? Use paper? │ ├────────────────────────────────┼────────────────────────────────────────────────┤ │ ✅ Google Photos │ Just use SFTP-backed folders lol │ ├────────────────────────────────┼────────────────────────────────────────────────┤ │ │ Just don't use? Fediverse stuff? Probably │ │ Social media │ empty account and use FOSS frontend selfhosted │ │ │ with authentication of my account (is this │ │ │ possible? probably...) │ └────────────────────────────────┴────────────────────────────────────────────────┘