llama.cpp is at the core of almost all offline, open weights models. The server it creates is Open AI API compatible. Oobabooga Textgen WebUI is more user GUI oriented but based on llama.cpp. Oobabooga has the setup for loading models with a split workload between the CPU and GPU which makes larger gguf quantized models possible to run. Llama.cpp, has this feature, Oobabooga implements it. The model loading settings and softmax sampling settings take some trial and error to dial in well. It helps if you have a way of monitoring GPU memory usage in real time. Like I use a script that appends my terminal window title bar with GPU memory usage until inference time.

Ollama is another common project people use for offline open weights models, and it also runs on top of llama.cpp. It is a lot easier to get started in some instances and several projects use Ollama as a baseline for “Hello World!” type stuff. It has pretty good model loading and softmax settings without any fuss, but it does this at the expense of only running on GPU or CPU but never both in a split workload. This may seem great at first, but if you never experience running much larger quantized models in the 30B-140B range, you are unlikely to have success or a positive experience overall. The much smaller models in the 4B-14B range are all that are likely to run fast enough on your hardware AND completely load in your GPU memory if you only have 8GB-24GB. Most of the newer models are actually Mixture of Experts architectures. This means it is like loading ~7 models initially, but then only inferencing two of them at any one time. All you need is the system memory or the Deepspeed package (uses disk drive for excess space required) to load these larger models. Larger quantized models are much much smarter and more capable. You also need llama.cpp if you want to use function calling for agentic behaviors. Look into the agentic API and pull history in this area of llama.cpp before selecting what models to test in depth.

Huggingface is the goto website for sharing and sourcing models. That is heavily integrated with GitHub, so it is probably as toxic long term, but I do not know of a real FOSS alternative for that one. Hosting models is massive I/O for a server.

tendentious milieu, or mircomilieu

True there is a long path ahead, but O’Neill cylinders solve most issues using 1970’s materials science. We are unlikely to survive the next providential scale volcanic traps type eruption like the Columbia Basin, Siberian, or Deccan Traps. We must reach the final technology of fully understood biology and an engineering corpus first. Then we unlock interstellar travel. We are already on a rock-ish ship circling the galaxy. The real challenge is when to get off the ship.

I am stuck in bed around 95% of my day. I am dealing with most of the issues you described.

PLA will be better for hardware store and hobby junk. You cannot use automotive class finishes and expect them to last. Generally stick to one brand. Most paints are formulated for steel. ABS is the closest to steel in thermal properties. The expansion is the most important attribute. PLA has a different thermal profile so catalysed 2-part paints will not work very well long term. Rattle can enamel is junk by comparison, but it never fully cures like automotive paints. That property helps it stay in place longer in general. There are special adhesion promoters like bulldog for automotive stuff, but the thermal properties will still be an issue.

Pro automotive paint is 99.9% sanding and prep work. It is far more intense and rigorous than people realize. Perfection happens in the prep work. The actual paint is just a way of showing off that perfection. Mastering automotive paint is actually all about defeating yourself. Perfection is not subject to your emotions or expectations. It is right when it is perfect.

You want the highest pressure spray cans as possible. Also, if you do not used all of the can at once, flip it upside down and clear the nozzle by letting the siphon into the empty void and spraying. If you have a compressor that does not shoot out a bunch of oil or water, a cheap Harbor Freight pink gun with the nozzle of the can beside the spray gun will work wonders by atomizing the spray far more effectively.

Decompresses my spine and makes me normal again?

Gravity prisons suck. Let’s plot our escape when the guards are not looking!

No country on Earth is like this. No polity is unified without opposition. Additionally, making this statement here, when it is obviously false, in a place where at least half the people are from the US, and likely all of those are left leaning and just as much victims if not more so, seems like rather tepid malevolence.

I think we should consider adding political philosophy to the cultural zeitgeist. We are unlikely to find Plato’s lofty philosopher king in the form of low quality reality TV or musical artist celebrity.

You are solely responsible for vetting the software that you choose to run.

I do not review or care about the tools a person uses to create their projects. I appreciate the disclaimer when the person discloses their aptitude and confidence in their code.

Free software and Unix culture is a culture of hackers. Stallman’s very degree is in AI. Emacs is mostly a thing because lisp was adapted early on for AI development many decades ago.

Junk code is nothing new. X11 is notoriously bad, yet you likely have parts of it running on your hardware. Proprietary code is far far worse than anything a hacker posts as open source, yet you are running proprietary blobs on whatever device you are looking at now. Even if you are like myself with a libreboot machine, Leah readily admits that you need to run the core duo microcode if you want it to run right, and are not using that hardware for your primary device. The culture of antiAI is dogmatic nonsense. It is a tool, not a religion. It can be used harmfully or helpfully. I can’t fix stupid in anyone except myself. I do not fault anyone for what they run, the projects they share, or the background they come from. I encourage everyone to be positive and help their fellow hackers. I value participation and enthusiasm. Dogma and negativity are toxic.

I am ultra liberal. You have a right to all information, a right to skepticism, a right to error, and a right to protest in non violent forms aka the right to offend others. You do not have a right to infringe the rights of others.

This anti AI populism infringes the rights to all information and right to error if any administrative actions are taken. Your right to protest and skepticism is duly noted. If these become toxic in any ways that alter the dissemination of information, or toxic/harmful to the individual sharing information, I will remove the offending comments. If the person continues, I will escalate. I am only the janitor here. I clean up the messes. I do not matter, but neither does anyone else here. It is a community, and only the community matters. Garbage software is bog standard. Crusade against things that matter like proprietary software leveraged hardware theft and SaaS.

Thanks for sharing.

It is not a laser beam.

No. The primary way of blocking radio is by raising the noise floor across the band. The type of radio is irrelevant. The protocol is irrelevant. It is all only the electromagnetic spectrum from infrared light, to visible spectrum light, to radio light, through to xray or gamma ray light. How we divide that up into protocols, bands, and names is totally irrelevant. When transmitting radio light, we are all restricted in how much power we are allowed to send. All receiver circuits are listening for meaningful information above the noise floor. Bands are allocated to try to create spaces for certain types of communications. This controls the noise floor. Then electrical engineers design the hardware you buy to operate within this specification. If that noise floor is raised, the physical hardware is unable to retrieve information and effectively makes it useless. If you are a radio wizard and build your own transmitter that has more power, you just created a giant beacon that anyone will track easily to your location. Transmitting always reveals your exact location. In military operations, you constantly hear about some entity going radio silent. This is why. If you are a soldier, you may not carry a cell phone at all when on the job because it is constantly revealing your location. The only way to avoid this is with actually hard wire connections. You are able to use lasers for line of sight communications, but in practice, you will be limited by the optical lens focusing complexity and atmospheric distortion even from the ground with point to point regional communication. If anyone crosses the beam it will still be detected and is likely to leak some light depending on conditions and design.

Ultimately, your only real option is the sneaker net which is damn near useless in US suburbia hell. Don’t forget that the freeway system was not created for the citizenry. It is only about military mobility. That is why the Germans made the autobahn and why the USA and others had to copy the idea. Your only defense is in the democratic political space.

That is one of the original Ole Hanson homes. That far row are the ones on the sea cliff. Sucks because the train is just below them, and the cliffs are sedimentary marine deposition that has been uplifted on geological time scales. When it rains here in SoCal, and you hear about houses falling in landslides, this is one of those areas. None of these places are single family homes. That Ole Hanson one has six units IIRC. Only the ones on the cliff are super expensive.

Sure

I have never used or cared about this W11. It has never seen the internet. I only keep it around for my keyboard’s RGB controller app if I ever need it. So I have no clue if this is everything or whatnot, but that is a screenshot of my access to the windows file system from within the file manager of Fedora. That is a dual boot partition. Fedora is particularly good at coexisting with a dual boot partition.

Use offline open weights.