If you are programming on a computer and not punching holes in paper you are just a little bitch script kiddy.

https://en.m.wikipedia.org/wiki/Computer_programming_in_the_punched_card_era

My philosophy is that anyone familiar with the code shouldn’t need the comments to understand it, and anyone who isn’t should be able to with the comments.

This means that I have semi long names like “this_variable_varies” would be about the max name size I would use. But I also use plenty of small functions and I use comments to explain big chunks of code with just a simple goal and method description.

I sometimes find my code a bit overly verbose, but I know I would rather see that then some code that I need to de-obfuscate mentally.

You can train an LLM to generate 1:1 copies

Check out my new site TheAIBay, you search for content and an LLM that was trained on reproducing it gives it to you, a small hash check is used to validate accuracy. It is now legal.

I got xiaomi, opened it up and disconnected the Bluetooth / wifi card. Connect it to a linux device and now it is a shitter version of a dumb tv. It’s crazy how smart tvs really really suck at being dumb. But it does work once you get used to some annoying quirks.

Tip: connect a cheap air mouse/keyboard to it as a remote

Please babe! Just one more parameter, then it will be AGI!

Vibe warfare sounds like a terrible idea

RAG is basically like telling an LLM “look here for more info before you answer” so it can check out local documents to give an answer that is more relevant to you.

You just search “open web ui rag” and find plenty kf explanations and tutorials

Have you tried RAG? I believe that they are actually pretty good for searching and compiling content from RAG.

So in theory you could have it connect to all of you local documents and use it for quick questions. Or maybe connected to your signal/whatsapp/sms chat history to ask questions about past conversations

Yeah, personally I just looked for second hand high vram gpus and waited. I got 2 titan Xp (12gb vram) for only $180 each.

I know there is all of that AI hate, which i’m all for. But taking models to run locally does not benefit the AI companies. If anything this is the way to make something that is actually good out of that hot mess.

My personal best is $4K because of one unchecked option.

You could use an llm with an mcp to the local filesystem and hope it can do it for you

And it will always end with a more modern and streamlined infrastructure that they never update again and then two years later there is an article that shows how they all went bankrupt.

That is actually missing an important issue, hallucinations.

Copying from SO means you are copying from a human who might be stupid or lie but rarely spews out plausible sounding hot garbage (not never though) and because of other users voting and reputation etc etc, you actually do endup with a decently reliable source.

With an LLM you could get something made up based on nothing related to the real world. The LLM might find your question to be outside of it’s knowledge but instead of realizing it it would just make up what it thinks sounds convincing.

It would be like if you asked me how that animal that is half horse and half donkey is called and instead of saying “shit i’m blanking” I would say “Oh, that is called a Drog” and I couldn’t even tell you that I just made up that word because I will now be convinced that this is factual. Btw it’s “mule”

So there is a real difference until we solve hallucinations, which right now doesn’t seem solvable but at best reduced to insignificance (maybe)

Nice

You remind me of Joey from friends when he gets a female roomate.

We could also force all people to wear ankle bracelets and be monitored by the government, AI could predict any ongoing or potential future crimes and we will have a crime free society where everyone is safe.

I think this should show internals of the car and have them all have a finished car but agile is like 50% zipties and AI is just a bunch of wheels and screws in a car shaped contianer.

Makes sense if you look at abliterated models. Once abliterated and retrained they seem to improve. Imo we are adding too much human bias by trying to guide the LLM. Censored models are good and need to be used in some situations, but shouldn’t the base be just data and only then finetune to desired output?