Is my old laptop still good enough?
"Good enough" is not a property of a laptop. It is a property of a laptop and a job. A 2017 ultrabook is an excellent web machine and a hopeless video workstation at the same time. The useful question is: good enough for what?
Stop asking about the laptop. Ask about the work.
Benchmarks answer "how fast is it", which is rarely what you want to know. What you want to know is whether the next two years of what you do will feel fine on this machine, and the honest answer is different for each thing you do.
Innards scores a machine across twelve workloads, and each one is limited by a different part. That framing works just as well on paper. Go through the list and mark which ones you actually do:
| Workload | What usually limits an old laptop |
|---|---|
| Everyday use (web, office, media) | A spinning hard drive, or less than 8 GB of memory. Almost any laptop with an SSD passes. |
| Web & app development | Memory (editors, a browser with dev tools and a local server add up) and core count. |
| Docker / container stacks | Memory first, then cores. Each container is a process tree that wants its own share. |
| Home server / self-hosting | Memory and a wired network port. Speed matters less than staying up. |
| Photo editing | Memory and graphics. A 24-megapixel raw file expands to hundreds of megabytes in an editor. |
| Large compiles (C++, Rust, kernels) | Cores. Two cores compile at roughly a quarter of the speed of eight. |
| Video editing | Cores, memory and a dedicated GPU, all at once. |
| Running AI models locally | A dedicated GPU with enough memory, or failing that a lot of RAM and a CPU with AVX2. |
| Modern gaming, ML training / GPU compute | A dedicated GPU. Without one these are simply not on the table. |
The pattern is that the first four rows are about memory, storage and stability, all of which can be cheap to fix, while the last four are about parts a laptop usually cannot gain. That is why an old machine can be "excellent" and "not suitable" in the same report without contradiction.
A real example: a 2017 ThinkPad
The machine Innards was written on is a 2017 ThinkPad X1 Carbon: Intel Core i7-7500U with two cores, 16 GB of soldered LPDDR3, a 2 TB NVMe drive that was fitted later, integrated graphics, and a battery at 45% of its original capacity. Here is how its report grades it:
- Excellent: everyday use (96/100).
- Good: web and app development (76), home server (75). Limits named: too few CPU cores, not enough memory.
- Workable: Docker and container stacks (54).
- Struggles: photo editing (28), running AI models locally (28).
- Not suitable: large compiles (24), modern gaming (23), video editing (21), ML training (4).
Nine years old, and it is an excellent everyday laptop and a decent home server. It is not going to train a model, and no amount of money spent on it will change that, because the two limits that matter for the bottom half of the list, soldered memory and no GPU, are not parts you can swap.
Three questions, in order
1. Is something slowing it down that costs nothing to fix?
Before judging the hardware, make sure you are judging the hardware. The most common things that make an old laptop feel older than it is:
- No swap. When memory runs out with no swap configured, Linux does not slow down gracefully; it kills a program. See what happens when memory runs out.
- A nearly full system drive. SSDs get slower and updates start failing past about 90% used. Free 10–15% and the machine often feels new again.
- Dust and dried thermal paste. A processor that throttles at 95°C is slower than its specification says. A loud fan at idle is the tell.
- Something in the background. A 5-minute load average of 8 on a 4-thread machine is not an old CPU; it is a process saturating it.
2. Which cheap part would change the grade?
Some parts move a grade by a lot and cost less than a new charger:
- A spinning hard drive to an SSD is the single biggest upgrade that exists: everything opens 3–10× faster.
- Memory, if the laptop has slots. Going from 8 to 16 GB turns "struggles" into "good" for web development and containers. Whether it has slots is the subject of should I upgrade RAM or buy a new laptop.
- A battery, if you use it unplugged. Usually a ten-minute job. See when to replace a laptop battery.
3. Is the ceiling a part, or the chassis?
Three findings mean the answer to "can I upgrade it?" is no, and honest advice should say so plainly:
- Soldered memory. Whatever you have now is the ceiling.
- No dedicated GPU and no way to add one. Laptops rarely have a slot; external GPU enclosures exist for some Thunderbolt machines but cost more than a used desktop.
- No AVX2. Processors from before 2013, and many budget Pentium and Celeron chips since, lack an instruction set that a lot of modern software, especially AI tools, expects.
If the workloads you marked are blocked by one of these, the machine is not "bad"; it is done growing. Keep it for what it is excellent at, and let the next machine be chosen for the rest.
Where old laptops go next
A machine that is "excellent" at everyday use and "good" as a home server has years of work left, especially now that Windows 10 support has ended and many of these laptops will not take Windows 11. A modern Linux desktop runs comfortably on an eight-year-old ultrabook with an SSD and 8 GB of memory, keeps getting security updates, and makes a quiet, low-power home server with a built-in battery for a UPS.
The one thing to be honest about is the battery and the thermals: a laptop used as a server should be kept plugged in with its charge capped around 80% if the firmware allows it, and dusted once in a while.