Yeah… In fourth grade I was taught that there is nothing like an outer foe to create inner peace. I never imagined it to be the US to accomplish that, but here we are.

That ability is known as Dave😁

There must be. Recall and info sec is mutually excluding by definition!

I’m just in the beginning, but my plan is to use it to evaluate policy docs. There is so much context to keep up with, so any way to load more context into the analysis will be helpful. Learning how to add excel information in the analysis will also be a big step forward.

I will have to check out Mistral:) So far Qwen2.5 14B has been the best at providing analysis of my test scenario. But i guess an even higher parameter model will have its advantages.

Thank you! Very useful. I am, again, surprised how a better way of asking questions affects the answers almost as much as using a better model.

I need to look into flash attention! And if i understand you correctly a larger model of llama3.1 would be better prepared to handle a larger context window than a smaller llama3.1 model?

Thanks! I actually picked up the concept of context window, and from there how to create a modelfile, through one of the links provided earlier and it has made a huge difference. In your experience, would a small model like llama3.2 with a bigger context window be able to provide the same output as a big modem L, like qwen2.5:14b, with a more limited window? The bigger window obviously allow more data to be taken into account, but how does the model size compare?

Thank you for your detailed answer:) it’s 20 years and 2 kids since I last tried my hand at reading code, but I’m doing my best to catch up😊 Context window is a concept I picked up from your links which has provided me much help!

The problem I keep running into with that approach is that only the last page is actually summarised and some of the texts are… Longer.

Do you know of any nifty resources on how to create RAGs using ollama/webui? (Or even fine-tuning?). I’ve tried to set it up, but the documents provided doesn’t seem to be analysed properly.

I’m trying to get the LLM into reading/summarising a certain type of (wordy) files, and it seems the query prompt is limited to about 6k characters.

Well, that’s been the basis for some other products. AMD and Intel comes to mind😊 They both have IP the other need and historically Intel has been the dominant one, but now the tables have turned somewhat.

Well… Its built on statistics and statistical inference will return to the mean eventually. If all it ever gets to train on is closer and closer to the mean, there will be nothing left to work with. It will all be the average…

An LLM once explained to me that it didn’t know, it simulated an answer. I found that descriptive.

What if we changed it to “Mexican warns of remorseless criminals north of the boarder”?

I know what you mean. The oldest building i pass every day is from dates from 1250, but compared to the parts of Europe rich in architecture that’s not really very old…

“Leftist”, “centrist” and “conservative”? This is not a the open discussion it is made out to be.

Having said that, those “traditional” values mentioned could easily be replaced with conservative values, especially when compared to the very individual, almost liberal, value of freedom to express oneself it is hard to see anything being “leftist” in the premiss of the discussion. Unless you are so far off the scale to the right that everything looks left…

No matter how you turn it, though, AI will be disruptive to life as we know it. The question is how to handle it. No media outlets will be spared until we figure out a way of validating information.

My favourite vision is how an AI robocaller is chatting with an AI answering machine. But honestly, why have a phone at all? All your loved ones calling may just as well be deep fakes trying to scam you for money.

It is clear we need to do something to prepare, but what? As you say, the cat is already out of the bag, so how do we proceed from here? AntiAI-software in the same way we have antivirus software? Even open source self hosting wount be enough, somehow you need to validate who to trust. Even if it would be a start…

I’m not entirely sure how to phrase this, but… Today there are choices…

Well, frankly, Northern Maine isn’t very much north and not enough inland to get the extremes?

Most of Scandinavia has been below -15F for a good part of the new year and being relatively dense (for Scandinavia) in EV coverage I’d say Norway is the best example of EV very cold weather performance.

We’ve had this same “debate” here as well with ice-owners lamenting the perceived loss of range and EV-owners responding “I know, don’t care. Always works, always warm and always topped up”.