From Dev.to Community
Tech Articles
Curated developer articles, tutorials, and guides — auto-updated hourly
Rickesh T NAug 24, 2026 • 4 min read
A 4B model on a 6GB laptop beat Claude Opus on our 440K-token corpus. The fix was giving the model less to do.
Private AI has a hardware story nobody measures honestly. The pitch is that your data never leaves.....
#privateai#llm#inference#localllm
1 1
Nathan C.2d ago • 8 min read
I A/B tested my own system prompt: 24 generations, one clear win, one rule that did nothing | Flash Onyx 2.2
Five new sections in Flash Onyx 2.2, measured against the build from before them, then six rounds of...
#ai#ollama#localllm#opensource
0 0
Nathan C.3d ago • 4 min read
Flash Onyx 2.2: teaching a local model law and game feel
Starting work on the next Flash Onyx. Two new domains in the system prompt, and why I cut both of th...
#ai#ollama#localllm#opensource
1 0
Syed Masood Shah5d ago • 2 min read
A year of meetings, and my minutes are plain text files on my own disk
Why I keep every meeting's minutes as local markdown I can grep — no vendor database, no lock-in.
#productivity#ai#localllm#windows
0 0
jan paul6d ago • 13 min read
I built a forgetting curve for an agent with one user
What I learned, mostly: I overbuilt the clever part, and then kept it anyway. In Greek myth the....
#localllm#aimemory#rust
0 0
Syed Masood Shah14h ago • 2 min read
My whole AI home lab is just a folder of docker-compose files
The most underrated thing about my local LLM setup isn't the model. It's that the whole thing is som...
#selfhosted#docker#openwebui#localllm
0 0
Syed Masood Shah3d ago • 2 min read
My meeting minutes go out as files now, not a chat bubble nobody scrolls back to
A working IT admin on why meeting minutes belong on disk as Markdown and PDF, not trapped inside a w...
#productivity#ai#localllm#windows
0 0
Syed Masood Shah3d ago • 2 min read
When our internet died, my local AI didn't notice
Our office line went down for three hours. The only AI that kept working was the one on my own desk.
#selfhosted#localllm#ai#homelab
0 0
Syed Masood Shah3d ago • 3 min read
My live dubber waits a sentence before it talks. That's not a bug, it's the honest one.
Every real-time dubbing demo makes it look effortless. Mine has a beat of silence baked in, and here...
#ai#localllm#whisper#python
0 0
Syed Masood Shah14h ago • 3 min read
My dub tool always picked the wrong voice. Here's how pitch solved it.
One speaker, one voice — until a scene has both. How I made a local dubber switch between male and f...
#ai#localllm#whisper#python
0 0
![Verify GGUF Model Hashes Supply Chain [2026]: 10 Steps](https://media2.dev.to/dynamic/image/width=1200,height=627,fit=cover,gravity=auto,format=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fpk2r9f1vz0npa8r3wnyg.png)
Kunal1d ago • 7 min read
Verify GGUF Model Hashes Supply Chain [2026]: 10 Steps
A team-ready workflow to verify GGUF integrity: compute SHA256, require signed manifests, handle mir...
#supplychain#localllm#gguf#llmsecurity
0 0
Patrick Hughes12h ago • 4 min read
Ollama Load Time Can Hide a Fast Local LLM
My RTX 5090 run spent 43.48 seconds on model load and 0.17 seconds on output. Split wall time before...
#localai#localllm#ollama#rtx5090
0 0
Patrick Hughes5d ago • 5 min read
49W Average Hid a 338W Burst on Gemma 26B
Gemma 4 26B Q4_K_M averaged 49 W on a long RTX 5090 run and peaked at 338 W. Keep both watt numbers ...
#localllm#gguf#ollama#rtx5090
0 0
Patrick Hughes12h ago • 5 min read
Split Local LLM Prose and JSON Jobs by Model
My local AI stack routes prose to Gemma and bounded JSON to Qwen. The split keeps model choice tied ...
#localai#localllm#ollama#modelrouting
0 0
Patrick Hughes5d ago • 5 min read
Preflight Local AI Before You Benchmark a Model
My local AI test stopped before model load because the runtime could not build. This preflight keeps...
#localai#localllm#inference#rtx5090
0 0
Patrick Hughes3d ago • 5 min read
Your Benchmark Row Never Saved the Driver Version
I audited seven RTX 5090 benchmark files on 2026-08-27. Not one records the GPU driver version. Here...
#localai#localllm#benchmarking#rtx5090
0 0
Patrick Hughes3d ago • 5 min read
Your Local LLM CSV Needs a Schema Version
I audited seven RTX 5090 benchmark CSVs and found four header widths, three power schemas, and no sc...
#localai#localllm#benchmarking#rtx5090
0 0
Patrick Hughes6d ago • 5 min read
How to Calculate Local LLM Energy per Token
My RTX 5090 test shows how watts and output rate become joules per token, and why the faster of two ...
#localllm#modelevaluation#ollama#rtx5090
0 0
Patrick Hughes2d ago • 5 min read
A 32 GB GPU Still Needs Host RAM Headroom
My RTX 5090 had 32 GB of VRAM ready, but my local policy stopped Ollama at 7.02 GiB of free host RAM...
#localai#localllm#ollama#rtx5090
0 0
Norvik Tech3d ago • 3 min read
Local LLM vs Cloud API on Cost…
Originally published at norvik.tech Introduction Explore the cost implications of using...
#webdev#localllm#cloudapi#costanalysis
0 0
Mustafa ERBAY4d ago • 12 min read
vLLM vs llama.cpp: Which Local Inference Engine, and When
Everyone who asks me "I'm going to run a model locally, vLLM or llama.cpp?" is really asking one...
#vllm#llamacpp#inference#localllm
2 0