Neat! I’ve known that Regional Prompter is powerful, but it’s too much of a pain for me to bother using. Hopefully this makes it easier.

I guess this is neat, but I don’t understand what the point is. Here’s what it says this program offers:

The program provides a digital studio environment, access to advanced AI tools and technologies, partnership with experts in the field, and opportunities for collaboration within our established artist community. It also includes a $1000 grant and promotion to our community of millions.

$1000 is nothing nowadays; I’m not sure what that would even do. Buy most of a 4090?

I already have a “digital studio environment” set up on my computer. I like it the way it is.

Partnership with experts and community collaboration are already pretty easy with social media. The generative AI community is generally helpful.

I guess this program might help people who aren’t already using generative AI get into it?

I’m all for blocking intrusive ads, but so far CivitAI’s are pretty normal. Why not support one of the biggest sites that’s helping this community thrive?

Especially with AI now becoming more mainstream and getting new developments every week or month. Didn’t check the right blog/newspost? The workflow you’re using is now outdated and slow.

Saint Oniisan is about as Christian as you can get.

This looks like the opposite of friendly to me. Is it supposed to be targeted towards cloud computing or web apps? I don’t really understand what its ideal use case is.

Every job will have some sort of crunch time. Even just staying in a programming position, the definition of “crunch time” will vary wildly. I’m lucky enough that “crunch time” just means that I set aside all my other tasks until I fix whatever is on fire, but I still get to go home on time unless I really want the overtime pay.

I don’t envy positions with forced 80-hour workweek crunch times. That’s a sign of bad management.

Because I am addicted to solving puzzles.

That makes sense. I really like that the documentation is right at the top; many times all I want to do is find the right page in the official docs. You might want to look at how results are prioritized though: right now when I search for something simple like “how to center a div”, that result from Mozilla’s docs is included but it’s hidden as the second or third result. I would expect the page that’s explicitly about centering a div to be the top result, followed by the docs page for the element itself and maybe pages for flex or grid or something. That’s a really simple example, so maybe it’s not the target of this project, but I would still hope that simple topics are covered just as well as complex ones.

EDIT: I was a bit mistaken: “how to center a div” does bring up the Mozilla documentation for centering an element, but “center a div” brings up a page about accessibility as the top result.

It’s a good start. I’m curious why you didn’t include a section for social media like StackOverflow or Reddit. If I go to Google with a question, it’s usually for an edge case not covered by the documentation. Maybe add them as a section at the bottom to indicate that they might be less relevant?

Also, this might just be a web developer thing, but why include blogs? Almost all coding blogs I’ve seen are SEO cancer that just copy from the documentation or each other. Are there actually useful blogs out there that I’ve just been missing?

Looks like this may be a known issue for some users.

Bluey.

It wasn’t that hard if you kept feeding it quarters. It took a lot of trial and error, but having infinite lives means it was eventually beatable.

I don’t really understand the science behind it, but in my experience I’ve had much more success using basic models for training.

Also, I’ve found that LoRAs are generally much easier and faster to train than embeddings. Is there a reason you’re going for an embedding over a LoRA?

Embeddings should generally be trained on base models to improve compatibility with models derived from the base. For SD 1.5, that means using either regular SD 1.5 or the NovelAI leak. You can sometimes get away with using more “basic” models that don’t have many merges, but that can be tough to gauge.

I got this mostly working, but it was not easy. Not only does Obsidian have a few peculiarities that make it less compatible with standard Markdown, but Word also does a few funny things.

Here’s the config.yaml I used for Pandoc:

from: docx
to: markdown-smart-simple_tables-multiline_tables-grid_tables+pipe_tables+yaml_metadata_block-superscript-subscript-bracketed_spans-native_spans-link_attributes-raw_html+rebase_relative_paths+four_space_rule
extract-media: "./"
wrap: preserve
markdown-headings: atx
tab-stop: 2
shift-heading-level-by: 1
standalone: true
template: obsidian.md
filters:
  - compact-list.lua
  - remove-single-characters.py
  - remove-extra-linebreaks.py
metadata:
  tags: "tags/go/here"

The three filters:

• Removed extra linebreaks added between bulleted lists to make them more compact.
• Removed lines with only a single character in them. Usually an invisible character like nbsp, which made Pandoc’s linter not remove them automatically.
• Removes linebreaks enclosed in Strong tags. This is an artifact from Word where a line is bolded but has no content: technically the line break is bolded.

I then ran the resulting file through a RegExp replacement to change the superscript carats into HTML sup tags.

Even after all this, I still have to go through with an Obsidian plugin to convert the standard Markdown links and embeds into [[Wikilink]] style, since Obsidian will only do one or the other throughout your whole vault.

I’m getting a 500 internal server error when accessing your site.

It’s funny you would reply about that: I actually did escalate it again and I’m working on getting a process implemented. It’s like pulling teeth, but I’m determined to get this fixed. Luckily my manager is finally with me on this, so I’m making some real progress for once.

I’ve thought about it many times but can’t find a good way to implement it. I don’t have access to the company’s GitHub or any shareable network locations. Don’t want to upload to my personal GitHub either since there is proprietary information in some of them. Right now I have them shared in a OneNote notebook that I manually update as I revise the scripts.

I’ve just given up at this point. I have my scripts and I’ll share them if I’m helping someone with an issue, but it was such a fight to even get them rejected that I don’t want to bother with that again on top of the rest of my work. If nobody in this chain that I’ve already gone through seems to care, and if developing these scripts doesn’t change my eligibility for a promotion (which I’ve been directly told it doesn’t), I don’t see the point in pursuing it any more.