deleted by creator

they have their own ai

They don’t make it:

https://proton.me/blog/lumo-ai

Lumo is based upon open-source language models and operates from Proton’s European datacenters. This gives you much greater transparency into the way Lumo works than any other major AI assistant. Unlike Apple Intelligence and others, Lumo is not a partnership with OpenAI or other American or Chinese AI companies, and your queries are never sent to any third parties. Our research and development will remain focused on open models, which we will continuously improve and augment.

Per discussions online, it seems to be a mix of small Mistral and Olmo models, and GLM for the max tier.

This’d be like arguing a company hosting database software to serve a product is a “database software company.” They are a hosting some tool made by someone else, with a sticker.

they also have the most ridiculous brand fans.

Perhaps. I’m a bit scared of how much critical mass they have, TBH. But this does not make them an AI company.

I would be careful with any “neutering” tools then, depending on your work, because some can mess with Teams, CoPilot, or whatever else you need. Check the options before you run the tools.

Windows Store apps like Teams work fine for me, though, as do IDEs.

Honestly I don’t know which projects are up to date, so I’m afraid to recommend what I vaguely remember.

Search GitHub/Gitlab for “windows debloat” or similar concepts, and sort the results by date updated (e.g. the most recent commit). This is precisely what I’d do with a fresh Windows install, anyway.

On top of that, I go through running services one by one and disable the ones I don’t need.

I disable Defender active protection (except for the firewall/anti tamper) as well, as it absolutely tanks the performance of many applications. As one example I measured, it multiplies the startup time for modded Rimworld, and dramatically slows game ticks. It seems to slow Cyberpunk 2077 as well, but not as reproducibly.

Most “debloaters” won’t do this, as it’s generally not a good idea. But I treat my Windows partition as disposable: nothing sensitive stored, no sensitive data accessible on drives nor networks, data on a separate partition. If it was infected, I could wipe it and reinstall without a care.

Maybe it’s because I neutered my W11 and only dual boot to launch the occasional game or Windows specific driver, but… I haven’t had significant issues with W11? Not that I can remember over the past two years.

No broken updates, no weird copilot things, no excessive background usage, nothing. The only maintenance I do is manually updating graphics drivers.

It wastes none of my time.

The one issue I had was the LG monitor debacle, that tried to install Mcafee, but that was LG’s fault (and I’m never, ever buying an LG display again, as long as I live. I hope you’re happy, LG marketing).

I find this remarkable, because to me it means, Windows doesn’t have to be junk, yet they deliberately configure it that way, and make it extremely difficult to undo. I used a slew of tools neuter it, but I’d have no idea where to start without them.

And to be clear, I use Linux 90% of the time, so I’m not advocating for Windows or anything.

Ah, sad. I asked because some of the Xiaomi phones themselves (AFAIK) now use a LOFIC sensor.

…If this camera module doesn’t, thats kind of perplexing. It would make the phone camera on some phones better than this dedicated attachment, in many shooting situations.

Aside note, but I love that crazy camera module on the phone.

Is it LOFIC sensor, too? It doesn’t seem to specify.

Admittedly, no, but this is slowly improving. Stepfun published their SFT dataset, smaller labs are publishing their datasets for task specific tunes. I believe there was another Chinese lab that published bulk pretraining data, but I can’t find it in my history at the moment.

And, notably, these comparatively tiny labs generally aren’t scraping the internet so abusively like OpenAI/Meta. They don’t need as much. Going by statements in their papers, they tend to use existing archives of web data, buy commercial data, or (more recently) generate a lot synthetic data.

If you run Nemotron, from open datasets.

See the dataset table here. There are sections for pretraining and for specific tasks/topics, like STEM, code and such: https://huggingface.co/nvidia

Chinese models are starting to open their pretraining datasets too, albeit slowly.

this new model, couldn’t universities rebuild it with different contexts and study it in ways you can’t reproduce in other models? like, is k2 horizons a good scientific foundation on which to study machine learning?

Not on a university budget, I don’t think.

But there is all sorts of interesting stuff that can be done with a truly open model other than replicating it.

On top of what others said, Nvidia Nemotron is truly open, with an open dataset.

There have been others (like Olmo), but Nemotron was unique in that it was reasonably useful/cutting edge, while other open models before were basically research artifacts.

Yeah. Kinda feels like OpenClaw and a few ‘trends’ before that.

…The heck happened?

I’ve been under a rock for a bit.

Explain to me: why is there so much hype (and controversy) around some Hyperland Arch Linux config, within like a month? Is this some astroturfing thing?

If you are just looking to anonymize YouTube, I’d browse with Cromite,download longer videos with yt-dlp, and watch them in mpv. Or use something that automates that.

You get better video quality that way, anyway.

Just FYI, it might not be sophisticated enough to beat YouTube ads. Not sure about that.

Yep, built in, with customizable lists.

Nope.

I think many Chrome extensions don’t work either, because of disabled APIs.

FYI extensions are another thing that can be used to fingerprint you. There’s a discussion on this somewhere on the Cromite GitHub, hence they don’t recommend extensions.

I mean Google can identify you with stuff like your resolution, cpu and gpu, measured execution speed, extensions, APIs, software versions. Browsers expose all that.

Check this site out. Its a great explainer: https://amiunique.org/

Now, if you run amiunique in Cromite, the results are interesting.

It may identify some stuff, but if you actually look at the readouts, you can see Cromite is spoofing data. It intenionally lies about your hardware and configurations to throw off fingerprinting, amongst other things.

This means it can actually mess up some sites that rely on the data: thats what I mean by “draconian.” Tons of features are disabled or stripped out. Cromite is a very extreme browser to daily drive, because its purpose built to defeat fingerprinting at any cost.

Use: https://github.com/uazo/cromite

My guess is YouTube is fingerprinting your browser, and no browser does antifingerprinting specifically as obsessively as Cromite.

Just read its feature list. Its quite draconian, heh: https://github.com/uazo/cromite/blob/master/docs/FEATURES.md

Me too.

Though I’m also sympathetic to copygirl because, in this rare instance, the video is the primary source, not the article.