For LLMs, I’ve had really good results running Llama 3 in the Open Web UI docker container on a Nvidia Titan X (12GB VRAM).

For image generation tho, I agree more VRAM is better, but the algorithms still struggle with large image dimensions, ao you wind up needing to start small and iterarively upscale, which afaik works ok on weaker GPUs, but will gake problems. (I’ve been using the Automatic 1111 mode of the Stable Diffusion Web UI docker project.)

I’m on thumbs so I don’t have the links to the git repos atm, but you basically clone them and run the docker compose files. The readmes are pretty good!

deleted by creator

I loved my course on patterns. It was tough, but I now regularly feel like I can apply mastery of this tricky subject to my software projects. The course used a variety of techniques:

• Read the seminal Design Patterns book by Gamma et al., for an overview of the concepts.
• Every week, we’d incorporate three patterns into a preexisting XML processor project. My final one had like 25 patterns, which was challenging to keep working amidst refactoring. (You don’t have to do them cumulatively, but I enjoyed it.)
• We’d have to ask pattern-specific questions of our classmates in forum threads; and occasionally we’d be assigned to answer some.
• We each wrote up our own pattern. (I designed one based on my experiences handling data exchange between web apps and clients.)

Together, this taught us

• How the patterns could concretely look in practice.
• Pros, cons, and other considerations for each.
• Similaraties, differences, and nuances. (We’d joke that everything was the Template pattern if you squinted.)
• The impact of modifications to the patterns.
• How to recognize, create, hone, collaborate on, and share patterns.

I appreciate this approach because patterns are an inherently fuzzy subject.

It’s more like languages evolved to incorporate the most common idioms and patterns of their ancestors. ASM abstracted common binary sequences. C abstracted common ASM control structures and call stacks. Java leaned hard on object orientation to enable compositional and inheritence-based patterns widely used in C and early OO languages. Python baselines a lot of those patterns, and makes things like the Null Object pattern unnecessary.

By that logic, I should take it up with the delivery guy; both he and the reseller simply passed-through a sealed product.

Ah yes, I must’ve used ChatGPT to generate the photos of it being sold in a sealed box. And the ebay account listing. And the receipt.

I could just make up a receipt from an authorized reseller. What kinda proof is good enough? Do these items degrade in a sealed box? If so, why track the warranty from resale date instead of manufacturing date? If not, photo evidence of a sealed box on sale should be sufficient imo.

The reality is, this sort of resale is common, is hardly more risky than with authorized resellers, and deserves greater consumer protections.

Crazy to think. But I’ll be damned if it isn’t still cool.

Bidet gang arise!

Love this idea. I definitely treat most content lists as an inbox; if I’ve interacted with it, archive it somewhere unobtrusive in case I need to refer to it later.

I’m just waiting for my data export request to come through, then gonna hit the shred button.

What, nine thousand?!

Jerboa’s been working for me. I wonder if it’s a background battery / processing permission issue.

Can we call communities “lemlets?”