Ghost The 3rd Return from Being Offline ... and we're back !!! Today just so happened to be a standard month of setting-up shop in the new digs. Most of the boxes have been unpacked except for the holiday stuff. I don't think we knew how much decorations we'd amassed over the years
Proxmox VE Featured Installing Proxmox VE 8.x on ARM64 (NanoPi R5s) This little journey started the night before Thanksgiving as we were packing for a long-ago planned trip to Japan. As is my usual wont, I found myself puttering around restlessly with the HomeLab trying to eek out a little bit more from the little bits of hardware I had accumulated
large action model Large Action Models vs the Rabbit R1 (and others) The Rabbit R1 was introduced during CES 2024 and it was probably one of the promised products that helped make the the term Large Action Models (LAM) more widely known compared to mainstream Large Language Models (LLM) and, to a degree, the future iteration which is Large Agentic Models (LAM)
AI PC Meeting with some Copilot+ PCs left me unimpressed. Why? Its literally been a week since I walked into a Costco and took notice of the electronics area. This is the normal store layout: beautiful 8k and 4k OLED displays first followed by some quick-selling items (that day it was golf shirts), turn the slight corner and you have the
UniFi UniFi Cloud Key Gen2 Plus Protect Hardware Upgrade Tips Some two-(2) years ago now, I posted a a video about a bunch of UniFi/Ubiquiti hardware that I'd suggested for purchase to better manage my research team's various network use remotely with the idea that there was no guarantee I would consistently be able
Ollama Running Local LLMs on Older Hardware? As part of this local LLM series, I wanted to know the lower limits of being able to self-host and potentially share demos or development frameworks with my team. Given the acceptable baseline discussion in my previous post, we'll use that to audition the hardware we come across.
SBC Comparing the AI Capabilities of Single-Board Computers (SBCs) During the start of the pandemic in 2020 and we were scrambling to do work-from-home set-ups, I recorded a short video on how to make an OPNsense/pfSense router using any spare desktop lying around and as many USB dongles you could get your hands on to extend your networking.
GPT-4o Summary: GPT-4o vs Local Language Models Evaluation Results This entry was intended to be Part-2 of my most recent report but realized on review that it is probably the natural follow-up to the other writings on the current state of LLMs (we've not even gone down the road of Small Language Models, yet!) in the first
GPT-4o Featured Measuring Azure OpenAI GPT-4o vs Local LLM Models Microsoft Build 2024 just concluded last week and one of the highlights was the catch-up release of OpenAI GPT-4o alongside a new hardware line-up of Copilot+ PCs running Windows on ARM with Qualcomm Snapdragon X chips under the hood. I'll address the huge x86 elephants that were in
Ghost Lights at the end of the (Cloudflared) Tunnel We'd had some marble sized hail in the afternoon, followed by a big storm last night which upturned the gazebo chair, and then the power went out which left me a bit stuck at home due to the weight of the garage doors. That being said, this site
Ghost The 2nd Coming of this site (like a Ghost) That's what it has been like getting this blog back online and having spent my whole Memorial Day weekend in front of the screens furiously trying to recover from a Proxmox VE 7.x hypervisor storage failure. It isn't my first rodeo using ZFS but it
Llama 3 Extending Llama-3 8B with 1-million context by Gradient Wow, things are heating up on this latest Meta model. I was looking to turn things up a notch on my local LLM workflow which I finally got working and gotten usable responses using the llama3:latest model from Ollama. For some reason, llama3:text was giving me some loopy
Docker LLM Workflows vs Docker Compose If you've been doing any experiments or projects using Docker Compose and somehow your API calls bring you to the point of frustration with a "Connection error" message, such as: http://localhost:<port_number> Then, let me tell you a secret. I promise, it
n8n n8n in Turnkey Linux Node.js For those of us, like me, who are just getting back to full-stack and understanding the changes in the Node.js realm, this error is going to annoy you for sure. I'm simply trying to get a running latest n8n working with this command: npm install -g n8n@
Meta Llama 3 Meta Llama 3 vs Ollama Performance In the middle of me taking on the challenge of a small full-stack project by my lonesome by rapidly developing a parallel alternative to one in production backed by more than a dozen engineers, I started looking to get a bit more embedding performance or alternatives to the "text-embedding-3-large&
Ollama Ollama Performance Comparison Chart I stumbled upon Ollama while looking for ways to run large language models (LLMs) locally for research at work some time last year during the initial explosion of the interest in ChatGPT. Being a long time Linux dabbler all my GPU have almost always been from Team Red (AMD) unless
OpenCTI How to fix OpenCTI Heat Map? If you self-host an OpenCTI instance via Docker (the recommendation), you may run into the situation where the heat-map suddenly does not work. In some instances, it probably had never worked depending on where you got your build instructions. I'm probably on my 5th iteration and it has
ZimaOS ZimaOS on Proxmox VE I'm always on the lookout for compact ideas to beta test in my #homelab before deploying in my office network cluster. While scrolling through Kickstarter and YouTube for Raspberry Pi projects to extend the lifetime of my rPi 3B+ cluster, I came across IceWhale's CasaOS and
Lacus Plumbing the Depths of Lacus Now that we've gotten a fresh new instance of the AIL Project installed, it is time to setup the web crawler back-end that will help map out the deep and dark web using another project called Lacus – which is a partial snippet of a much larger project called
AIL Project AIL Project 5.2 Adventures in a Linux Container Coming out of the recent Virus Bulletin 2023 Small Talks in London, I was lucky to have been present and see more cybersecurity practitioners take notice of this cool project, especially from those who probably had never heard of it before (well, before me, at least!). Wanting to make a
travel Marrakech in the Kingdom of Morocco The origins of this post come from a team presentation we were asked to do. Since I hadn't really made a proper recap of my last travel, I figured why not hit a few stones at once. Some background Where to begin to describe this amazing place? The
AIL Project Analysis Information Leak (AIL) Project - Proxmox VE Installation in LXC First attempt. For compatibility reasons I went with using Ubuntu 18.04 LTS (Bionic Beaver) as this project appears to be about 4-years old now. The problem with this, however, is that tlsh has a functional build dependency to use CREATE_LINK which was only introduced in CMake 3.14.
Mikrotik Exploring a used MikroTik CCR1036-8G-2S+ Tile-GX was a chip product from the Tilera Corporation which eventually shutdown in 2014 when it was purchased by EZchip Semiconductor which then merged will Mellanox Technologies in 2016 and went on to be purchased by Nvidia in 2019. Linux dropped official kernel support for Tile in 2018, but the
tracking Deep and Dark Web Monitoring When it comes to little boogeyman terms online we often hear about the deep web or the dark web. These terms are often interchanged, and that's alright when trying to simplify things. However, one most note that over-simplification is akin to saying that a thunderstorm and a deluge
Pi-hole Pi-Hole and AdGuard Blocking Performance These are the stats of running an ad-blocker as your primary and secondary DNS. Based on the resolved domains it appears that the IoT devices have been calling back home a bunch of times. We have a lot of Google devices, Ubiquiti wireless access points in a mesh, and the