Exactly. The system is what it does.

My biggest takeaway here is that choosing the context length and (to a lesser extent) the temperature carefully is important for reducing hallucinations. I expected model families to vary widely between themselves but not for context length to have such a massive impact tbh.

It seems from this like reducing context length in applications where it isn’t essential for the model to hold very large amounts of context simultaneously would be best practice no?

And diesel prices (which underpin essentially all industry and production) have gone even more brrrrrrrrrr. Which means just about everything else will be shooting up in price soon if the strait remains closed…

I’ve heard Calibre Web has a better UI and experience. Can anyone weigh in on whether that is true? I’m mostly looking for a better way to manage my book library on KOReader (Kindle) than just putting things into 3 directories (reading, to read, read) and using syncthing to copy them to the equivalent folders on my device.

As someone who has used both Evernote and Standard Notes, I can say that the latter is obviously better for privacy but for those who are really looking for a FOSS option and want something more powerful than a simple notes store (albeit with a bit of a learning curve) you should definitely check out Logseq.

Yes, because making locally hosted LLMs actually useful means you don’t need to utilize cloud-based and often proprietary models like ChatGPT or Gemini which Hoover up all of your data.

This is very cool. Will dig into it a bit more later but do you have any data on how much it reduces hallucinations or mistakes? I’m sure that’s not easy to come by but figured I would ask. And would this prevent you from still using the built-in web search in OWUI to augment the context if desired?

I would check out Open WebUI which can be self-hosted via docker etc and configured with any OpenAI compatible endpoint so you can use a service like OpenRouter to run nearly any LLM remotely. Most of the open weights ones like Qwen 3 or Kimi K2 Thinking are great and cost pennies per inquiry and can be configured with Zero Data Retention (ZDR) so your data is not recorded. You could also use something like Ollama to run local LLMs if you want even more privacy and have the hardware (typically a modern Nvidia GPU with at least 16-24 GB of VRAM).

Hopefully these improvements will become available to other Nvidia GPU architectures like Ada and Ampere in the future as well.

I think we can all agree that modifications to these models which remove censorship and propaganda on behalf of one particular country or party is valuable for the sake of accuracy and impartiality, but reading some of the example responses for the new model I honestly find myself wondering if they haven’t gone a bit further than that by replacing some of the old non-responses and positive portrayals of China and the CPC with a highly critical perspective typified by western governments which are hostile to China (in particular the US). Even the name of the model certainly doesn’t make it sound like neutrality and accuracy is their primary aim here.

I used to daily drive Ubuntu some years ago for work/personal use but have been back on Win 10 primarily for the last 4-5 years. I was considering trying to go back due to how much Windows sucks (despite some proprietary software only being available on it) but remembering the trouble I had with some networking/printer drivers and troubleshooting those issues and then seeing this article Is definitely making me reconsider…

Yeah I use voyager pretty much exclusively on my iPhone so maybe I should request a feature like that there? Seems like it would be something that many people would appreciate. Not sure why I end up seeing posts with -10, -15 votes… Those are generally trash haha

The US government’s position on this can be summed up as “massive unaccountable US tech firms having all of your data and manipulating public opinion via their black box algorithms is okay, but Chinese companies doing that is a national security concern”. I call BS. The degree to which China is actually a US adversary is being massively overstated by the US government as they see this as a threat to US geopolitical hegemony and America’s ability to propagandize its own citizens. I have spent some time on RedNote (Xiaohongshu) and all I have seen is friendly cross-cultural exchange and discussion between these supposed ‘adversaries’.

True; I think I used LineageOS or similar back when I was still in Android but if you’re not in the 0.01% who do have a custom Android OS installed it seems like a privacy focused map app is still of limited use potentially.

This looks like it has come a long way, but since this is a privacy community I have to ask: Realistically, whether you are on iOS or Android, isn’t it likely Google or Apple are still tracking your location much of the time directly from the OS?

Or you already use Proton Mail / Drive etc and want to save a bit of money with Proton Unlimited at 9.99 EUR/USD / mo. That’s why I switched from Mullvad. I do agree that overall though Mullvad is the best in terms of trustworthiness (so far as we know anyway)

The guardian article only summarizes a limited amount of the product test information. See here for more products that were tested: https://www.mamavation.com/health/condoms-lubricants-pfas-forever-chemicals.html#Not_Our_Favorite_Condoms

This kind of technocratic approach can sound great and it’s always good to have policy and positions formed by those most knowledgeable in the relevant field. The problem is that the “experts” that will likely have her ear will be the ones that are vetted and approved by the power elite, the wealthiest in this country who will largely recommend what is best to protect their privilege, wealth and position.

This is the kind of news that should actually be posted on all of the ‘uplifting news’ communities, instead of ‘teen works 600 hours for minimum wage to help pay for his mother’s dialysis’ or whatever…