Practical technology knowledge
Tutorials, guides, comparisons and analysis on security, Linux, networking, development, cloud and the tooling that powers modern technology — written by practitioners, for practitioners.
Free tools
Most run entirely in your browser, so nothing you paste — a token, a password, production data — is ever sent to us.
Dedicated server listings compete on core count and price. The things that actually determine whether a machine suits your workload — disk condition, network commit, IPMI access, what happens when a drive dies — are usually absent from the listing and answerable in a few minutes if you ask.
A drive advertised at 7,000 MB/s rarely feels seven times faster than one at 550. The gap between the benchmark and the experience comes down to queue depth, random access and a cache that runs out — and knowing which matters changes what you should buy.
Virtualising a home server buys snapshots, isolation and painless rebuilds. It costs you direct hardware access, a layer of complexity on every problem, and a restore path you probably have not tried. Here is how to decide.
Advice about RAM is usually a number with no reasoning attached. Your own machine can tell you the answer in a few minutes — and on Linux, macOS and Windows alike, the figure most people read is the wrong one.
Every VPS host oversubscribes; the question is by how much. Steal time, run-queue depth and fsync latency expose it in minutes, and knowing the numbers turns a vague "feels slow" into something you can act on.
Putting smart devices on their own VLAN is straightforward until the television stops appearing on your phone. The reason is mDNS, which does not cross subnets by design — and the fix is a reflector with rules tight enough to be worth having.
A GPU with twice the compute can generate tokens no faster than one with half of it. Generation reads the entire model from memory for every token produced, so bandwidth sets the ceiling — and you can calculate yours before you buy anything.
Quantisation makes models fit, and the format you choose decides which hardware can run them, how much quality survives, and whether the file you downloaded will load at all. A practical comparison of the formats you will actually meet.
Most small web tools upload whatever you paste to a server you know nothing about, and most of that paste is a production token. The browser can do all of it locally now — and a CSP lets you prove it rather than claim it.
Lists of fifty questions with model answers optimise for the wrong thing. Interviewers are working out how you think, where your knowledge stops, and whether you will say so when it does.
