one of those blogs
A year ago I would just read about 397B league of models. Today I can run it on my laptop. The combination of llama.cpp s importance matrix (imatrix) with Unsloth s per-model adaptive layer quantization is what makes it all possible Qwen3.5-397B-A17B is an 807GB beast that, even at Q4, would need more than 200GB of GPU [ ] The post Qwen 3.5 397B on a MacBook @ 29 tokens / s first appeared on . kam…
The last Llama (Scout/Maverick) was released a year ago. Since then US based releases have been super rare: Granite 3.3, GPT-OSS 20B 120B, Nemotron 3 Nano / Super and now Gemma 4. Meanwhile Chinese open model output is multiples of high quality Qwens, DeepSeeks, Kimis, MiniMaxes, GLMs, MiMos, Seeds, etc.. Gemma 4 is like [ ] The post US Public Schools meet Qwen 3.5 122B A10B first appeared on .…
Simple wins. Always. Because it is very composable, fun to work with, and organically evolves. LLMs are great at language and text, but they don t know how to do anything, so they rely on something running (e.g. a server) and listening to their output. If the output is formatted in a certain way, say: sum(41, [ ] The post Mad Skills to Learn The Universe first appeared on . kam .
Hazelcast has a neat feature called Near Cache. Whenever clients talk to Hazelcast servers each get/put is a network call, and depending on how far the cluster is these calls may get pretty costly. The idea of Near Cache is to bring data closer to the caller, and keep it in sync with the source. [ ] The post Hazelcast: Keep your cluster close, but cache closer first appeared on . kam .
This week we learned of a recent alien hacking into Earth. We have a suspect but unsure about the true source. Could be The Borg, could be Klingons, could be Cardassians. One thing is certain: we need better security and flexibility for our space missions. We ll start with the first line of defense: science based [ ] The post Hubble Space Mission Securely Configured first appeared on . kam .
Most of the tools I push to github are created out of something that I needed at the moment but could not find a good alternative for. cprop was one of such libraries. It sat there on github all alone for quite some time, and was used only by several people on my team, until [ ] The post cprop: internal tools worth opening first appeared on . kam .
The first programming language I learned was Basic. It was sometime ago, and my first computing beast at the time was ZX Spectrum. I was in school, then another school, then college, then another college, etc. All that time application configuration was not really a thing I cared about. Who needs to create programs for [ ] The post Why Configuration Makes You Happier first appeared on . kam .
DI framework makes sense for OOP In Java (or most OOP languages): Objects need to be created In most of the cases they are stateful Dependencies (state) often need to be injected Order of the creation needs to be determined/given for the injection to work Hence an IoC framework such as Spring makes perfect sense [ ] The post No Ceremony first appeared on . kam .
Dogma From The Top The best thing about the Clojure community is a large number of independent thinkers. The second best thing that allows this community to exist is the Clojure language itself, that have the hammock philosophy and great tools to enable thinkers to think and create. Often I notice myself following a 90 [ ] The post The Way Nature Creates Things first appeared on . kam .
Development and Structure Factors 12 Factor clearly states that configuration should live in environment variables: The twelve-factor app stores config in environment variables (often shortened to env vars or env). Env vars are easy to change between deploys without changing any code; unlike config files, there is little chance of them being checked into the [ ] The post Config You Know Plus Some…