ralakus

joined 1 year ago
[–] ralakus@lemmy.world 4 points 2 weeks ago* (last edited 2 weeks ago)

If you're using an LLM, you should limit the output via a grammar to something like json, jsonl, or csv so you can load it into scripts and validate that the generated data matches the source data. Though at that point you might as well just parse the raw data and do it yourself. If I were you, I'd honestly use something like pandas/polars or even excel to get it done reliably without people bashing you for using the forbidden technology even if you can 100% confirm that the data is real and not hallucinated.

I also wouldn't use any cloud LLM solution like OpenAI, Gemini, Grok, etc. Since those can change and are really hard to validate and give you little to no control of the model. I'd recommend using a local solution like running an open weight model like Mistral Nemo 2407 Instruct locally using llama.cpp or vLLM since the entire setup will not change unless you manually go in and change something. We use a custom finetuned version of Mixtral 8x7B Instruct at work in a research setting and it works very well for our purposes (translation and summarization) despite what critics think.

Tl;dr Use pandas/polars if you want 100% reliable (Human error not accounted). LLMs require lots of work to get reliable output from

Edit: There's lots of misunderstanding for LLMs. You're not supposed to use the bare LLM for any tasks except extremely basic ones that could be done by hand better. You need to build a system around them for your specific purpose. Using a raw LLM without a Retrieval Augmented Generation (RAG) system and complaining about hallucinations is like using the bare ass Linux kernel and complaining that you can't use it as a desktop OS. Of course an LLM will hallucinate like crazy if you give it no data. If someone told you that you have to write a 3 page paper on the eating habits of 14th century monarchs in Europe and locked you in a room with absolutely nothing except 3 pages of paper and a pencil, you'd probably write something not completely accurate. However, if you got access to the internet and a few databases, you could write something really good and accurate. LLMs are exceptionally good at summarization and translation. You have to give them data to work with first.

[–] ralakus@lemmy.world 10 points 2 weeks ago (1 children)
[–] ralakus@lemmy.world 2 points 2 weeks ago

Women are very nice to see you soon as you scramble around in the dark hoping to pull the way out of work and get a little more of your time to think and there's a lot more non nazis in your heart

I don't know what this says about me lol

[–] ralakus@lemmy.world 4 points 3 weeks ago

In small datasets, the speed difference is minimal; but, once you get to large datasets with hundreds of thousands to millions of entries they do make quite a difference. For example, you're a large bank with millions of clients, and you want to get a list of the people with the most money in an account. Depending on the sorting algorithm used, the processing time could range from seconds to days. That's also only one operation, there's so much other useful information that could be derived from a database like that using sorting.

[–] ralakus@lemmy.world 5 points 3 weeks ago (1 children)

I like how you sent a screenshot of a failed attempt

[–] ralakus@lemmy.world 2 points 4 weeks ago (2 children)

So what you're saying is we destroy the unclean power sources

[–] ralakus@lemmy.world 19 points 4 weeks ago (3 children)

I'm pretty sure it's because the use of absurd amounts of high fructose corn syrup. There's 39g (can't confirm, I got it from Google) of sugar in a 12oz (340ml) can. US soda is pretty much just carbonated high fructose corn syrup water with a bit of flavoring. There's probably other significant differences too since the US has barely the minimum food safety laws.

[–] ralakus@lemmy.world 9 points 1 month ago

Anxious but looking forward to moving out. Things just have been really rough on me mentally for the past few years where I currently am and I'm just really looking forward to at least put some of those memories behind me for a while

[–] ralakus@lemmy.world 8 points 1 month ago* (last edited 1 month ago)

I'd place the blame more on businesses and dumb managers keeping them afloat since they still think Intel is the best bet for their computers or they're stuck with Intel due to existing contracts. After AMD came in and bitch slapped Intel with Zen, a lot of the community switched teams and went over to AMD for CPUs but most businesses haven't yet.

[–] ralakus@lemmy.world 39 points 1 month ago (3 children)

For anyone curious, I couldn't find an exact statistics but hearing aids in the US cost between $2000 to $8000 per pair with the average costs sitting around $5000-$6000 per pair.

Insurance coverage varies per insurance provider and per state. It looks like many people will end up paying the maximum required by law before insurance takes over which is roughly between $1000-$3000 depending on state.

Not only is a single purchase expensive, you usually have to replace them every 3 to 5 years.

[–] ralakus@lemmy.world 5 points 1 month ago (1 children)

I'd place the blame on the Democrats pushing such an unlikable candidate, Hillary Clinton, that they couldn't move enough people to vote to get past the electoral college bias. There's also the fact that most people thought "How bad could Trump be. He looks honest and confident in what he says" and a lot of positive media around that time being about Trump.

This time, we have Kamala Harris who's actually really good at campaigning and getting support. The move the Democrats pulled managed to derail the GOP Propaganda machines (like Faux News) and put the positive focus on Harris instead of Trump.

Regardless of what is happening in the Republican side, we still need to vote since this could be our last chance to vote.

view more: ‹ prev next ›