Private Devices
Lesson 3 of 4
Private · Lesson 3 · ~8 min
Make it last longer.
Sometimes nothing is broken — the device just stops being supported, or it's too weak for what you need it to do. The reflex is to replace. The smarter move is to change what the device does — switch the operating system, offload heavy work to the cloud, or give the hardware a different job entirely.
This lesson is about extending useful life through software and creativity, not new hardware.
When the OS quits, the hardware doesn't
Manufacturer abandonment ≠ device end-of-life
October 2025 was Windows 10 end-of-support. Hundreds of millions of laptops worldwide were suddenly classed as "unsupported" by Microsoft. Most of them work perfectly. Their CPUs are not slower than they were the day before; their batteries (if replaced) hold full charge; their keyboards still work. What changed is not the device — it's the manufacturer's commercial interest.
This pattern is common: Apple drops macOS support for older Macs, Google ends ChromeOS updates for older Chromebooks, Android manufacturers stop pushing updates after 3–4 years. The hardware keeps working; the official software lifecycle ends.
You have three real options, in order of preference:
1. Switch the operating system. Replace the manufacturer's OS with one that still gets updates. Linux distributions are the textbook example — they run on hardware that Microsoft has long abandoned, and they receive security updates for years. ChromeOS Flex (Google's free conversion image) is another option for kiosk-style use.
2. Offload the heavy work to the cloud. When the local hardware can't keep up with modern apps, use the cloud as a backstop: a browser-based version of Office, cloud-rendered IDEs, or even full remote-desktop access to a more powerful machine elsewhere.
3. Give it a different job. When the device can't run modern user-facing software well, repurpose it into something simpler: a media centre, a NAS, a kids' homework computer, a kitchen-table music player.
This pattern is common: Apple drops macOS support for older Macs, Google ends ChromeOS updates for older Chromebooks, Android manufacturers stop pushing updates after 3–4 years. The hardware keeps working; the official software lifecycle ends.
You have three real options, in order of preference:
1. Switch the operating system. Replace the manufacturer's OS with one that still gets updates. Linux distributions are the textbook example — they run on hardware that Microsoft has long abandoned, and they receive security updates for years. ChromeOS Flex (Google's free conversion image) is another option for kiosk-style use.
2. Offload the heavy work to the cloud. When the local hardware can't keep up with modern apps, use the cloud as a backstop: a browser-based version of Office, cloud-rendered IDEs, or even full remote-desktop access to a more powerful machine elsewhere.
3. Give it a different job. When the device can't run modern user-facing software well, repurpose it into something simpler: a media centre, a NAS, a kids' homework computer, a kitchen-table music player.
Knowledge check
Your 7-year-old laptop is on Windows 10. Microsoft ended support in October 2025. The device runs fine but Windows 11 refuses to install (CPU not on the supported list). What's the most carbon-efficient option?
Correct.
Not quite — review the section above.
Cloud as a backstop, not a default
When local hardware reaches its limit
There is a thoughtful Green-IT argument for cloud — and a thoughtless one. The thoughtful version: when your local hardware genuinely cannot run a workload, the cloud lets you keep the local hardware in service for everything else. You avoid a new device by offloading the one demanding workload.
Examples where this works well:
• AI and large language models. Running a local LLM needs a recent GPU (typically RTX 3060 or better, post-2021 hardware). Older laptops simply cannot do this. Cloud-hosted services (ChatGPT, Claude, Gemini, or self-hosted models on a rented GPU) let you use AI without buying a new laptop just to get an NVIDIA card.
• Heavy video editing or 3D rendering. Occasional rendering can be done on a rented cloud workstation (Paperspace, Vast.ai, AWS) for a few euros per session, much cheaper than buying a workstation outright.
• Modern web apps on old machines. Office 365 web, Google Docs, Figma, and others run fine on a Chromebook or a Linux laptop with 4 GB of RAM, because the heavy computation happens server-side.
The thoughtless version is treating cloud as a default substitute for local computing. Cloud has a carbon footprint too — data centres globally consume ~1.5 % of electricity. The Green-IT logic is: cloud to extend hardware life, not to replace local computing wholesale.
Examples where this works well:
• AI and large language models. Running a local LLM needs a recent GPU (typically RTX 3060 or better, post-2021 hardware). Older laptops simply cannot do this. Cloud-hosted services (ChatGPT, Claude, Gemini, or self-hosted models on a rented GPU) let you use AI without buying a new laptop just to get an NVIDIA card.
• Heavy video editing or 3D rendering. Occasional rendering can be done on a rented cloud workstation (Paperspace, Vast.ai, AWS) for a few euros per session, much cheaper than buying a workstation outright.
• Modern web apps on old machines. Office 365 web, Google Docs, Figma, and others run fine on a Chromebook or a Linux laptop with 4 GB of RAM, because the heavy computation happens server-side.
The thoughtless version is treating cloud as a default substitute for local computing. Cloud has a carbon footprint too — data centres globally consume ~1.5 % of electricity. The Green-IT logic is: cloud to extend hardware life, not to replace local computing wholesale.
Knowledge check
You have an old smartphone you haven't used in two years. The screen works, the battery has degraded to ~50 % capacity but holds charge when plugged in. Which is the best Green-IT use for it?
Correct.
Not quite — review the section above.
A note on Cloud and your data
What the cloud sees, what stays local
Moving workloads to the cloud has a privacy dimension that Green IT shouldn't ignore. When you put files on Google Drive, OneDrive or iCloud, the provider has technical access to them. When you ask ChatGPT a question, the conversation is processed on someone else's hardware.
For most everyday use, this is acceptable — the providers have business and legal incentives not to misuse the data. But it's worth being intentional:
• Sensitive documents (passwords, medical, financial, work-confidential) often belong in client-side encrypted services (Proton Drive, Cryptomator on top of any cloud) or on local-only storage (the NAS pattern above).
• Routine documents (school papers, hobby projects, family photos) are fine in mainstream cloud, with the understanding that you trade some privacy for reliability.
• Local-first alternatives exist — Nextcloud (self-hosted file sync), Joplin (notes), Bitwarden (passwords, with optional self-hosting) — and they often run perfectly well on an old repurposed laptop acting as your home server.
For most everyday use, this is acceptable — the providers have business and legal incentives not to misuse the data. But it's worth being intentional:
• Sensitive documents (passwords, medical, financial, work-confidential) often belong in client-side encrypted services (Proton Drive, Cryptomator on top of any cloud) or on local-only storage (the NAS pattern above).
• Routine documents (school papers, hobby projects, family photos) are fine in mainstream cloud, with the understanding that you trade some privacy for reliability.
• Local-first alternatives exist — Nextcloud (self-hosted file sync), Joplin (notes), Bitwarden (passwords, with optional self-hosting) — and they often run perfectly well on an old repurposed laptop acting as your home server.
Why this matters for the rest of the path
Connecting the dots
Lessons 1–3 covered three of the four R's: don't buy (Reuse), fix what's broken (Repair), extend through software and creativity (Refurbish at the user level).
The last R — Recycle — is the topic of Lesson 4. It is the last resort by design: every device that gets recycled while still working is a missed opportunity. But when recycling is genuinely the right answer, doing it well matters: data wiping, donation to Digital Inclusion before recycling, choosing the right Luxembourg drop-off route.
The last R — Recycle — is the topic of Lesson 4. It is the last resort by design: every device that gets recycled while still working is a missed opportunity. But when recycling is genuinely the right answer, doing it well matters: data wiping, donation to Digital Inclusion before recycling, choosing the right Luxembourg drop-off route.