Welcome to Incremental Social! Learn more about this project here!
Check out lemmyverse to find more communities to join from here!

Zworf

@Zworf@beehaw.org

This profile is from a federated server and may be incomplete. Browse more on the original instance.

Zworf , (edited )

I don't think it will.

Microsoft's endgame is being the lord and master of AI. AI thrives on knowing more data about the user. What good is an assistant if it doesn't know your habits, your wishes and desires, your schedule and your attitude towards each person in your life?

This is not really a feature primarily aimed at helping the user directly (even though it's currently marketed as such), but to have the AI build up a repository of knowledge about you. Which is hopefully used locally only. For now this seems to be the case, but knowing Microsoft, once they have established themselves as the leading product they will start monetising it in every way possible.

Of course I'm very unhappy with this too. I'd like to have an AI assistant. But it has to be FOSS, and owned and operated by me. I don't trust microsoft in any way. I'm already playing around with ollama, RAG scripting etc. It won't be as good as simply signing up to OpenAI, Google or Microsoft but at least it will be mine.

Zworf ,

Or maybe Affinity Designer? I bought that a few years ago for Mac and it was really good.

Zworf ,

If you had a visual disability you would certainly think otherwise.

Zworf ,

One thing I'd love to see in Firefox is a way to offload the translation engine to my local ollama server. This way I can get much better translations but still have everything private.

Zworf ,

Yes I was just writing that, I would love to see more integrations that can talk against ollama.

Zworf ,

Yep this is one of the reasons I kept deleting my account even before the whole spez drama.

Token2 is an open-source Swiss FIDO2 security key that brings innovative features at a cheaper price (www.token2.ch)

Token2 is a cybersecurity company specialized in the area of multifactor authentication. Founded by a team of researchers from the University of Geneva with years of experience in the field of strong security and multifactor authentication. Token2 has invented, designed and developed various hardware and software solutions for...

Zworf ,

Too bad they don't do OpenPGP like Yubikeys do. I still need that even more (much more!) than Fido2. Sites are so slow adopting Fido2.

I don't use it for email but I use it for SSH and my password manager ("pass"). And yes I know SSH can use Fido2 natively as well but there's many embedded SSH daemons that don't support that yet.

Luckily Yubico is still around but I'm betting on them going down the drain (subscription models etc) soon because they were taken over by a venture capital firm :(

Zworf ,

Yeah that slogan really captured very well the intentions at the world economic forum.

I know it's not what they officially stated but it really captured (they since walked it back and said it only was meant to "describe emerging trends") the intentions of what happens when they all come to Davos and divide the world between them.

But I don't believe "as a service" models are more sustainable. They will just enable more rent-seeking behaviour meaning we will get even less for our money. The incentive to deliver will be even lower as they will get paid anyhow.

Zworf ,

I have LTSC 2021 officially (MSDN) and I have to say I'm not very impressed. You still can't turn off the telemetry crap. There is still a windows store. There's a bit less bundled scamware but beside that it's a bit overrated IMO.

Zworf ,

Even all the telemetry?

I really hate dealing with group policies (and I work in enterprise endpoint management, I prefer more modern management). AD/Group policies can only be updated on site or VPN, and they're only really instructions for registry settings anyway.

But I'll try that out. I don't have a windows server though, nor do I want one. But I guess I could use gpedit.

Zworf ,

The audio from the AI also seemed to cut out a lot during the demo. So it does appear like no shenanigans to me.

Zworf ,

I didn't think it was super creepy but I thought the voice was so overly enthusiastic and overacted and soooo sugary. bleh.

This won't work for me unless that can be customised and toned down a lot.

Zworf ,

Training your own will be very difficult. You will need to gather so much data to get a model that has basic language understanding.

What I would do (and am doing) is just taking something like llama3 or mistral and adding your own content using RAG techniques.

But fair play if you do manage to train a real model!

Zworf ,

Hmmm weird. I have a 4090 / Ryzen 5800X3D and 64GB and it runs really well. Admittedly it's the 8B model because the intermediate sizes aren't out yet and 70B simply won't fly on a single GPU.

But it really screams. Much faster than I can read. PS: Ollama is just llama.cpp under the hood.

Edit: Ah, wait, I know what's going wrong here. The 22B parameter model is probably too big for your VRAM. Then it gets extremely slow yes.

Zworf ,

It depends on your prompt/context size too. The more you have the more memory you need. Try to check the memory usage of your GPU with GPU-Z with different models and scenarios.

Zworf , (edited )

This is something that the amazing Naomi Wu brought up for years before, and was ordered to stop publishing by the local government. It was about the same thing. It's sometimes misrepresented as being about Signal, but her point was: There's no point in using a secure messaging app like Signal if your keyboard (IME) leaks everything you write! So she was making the exact same point as in this article.

I really miss her content. 😢

Zworf ,

Yeah she clarified that literally, it's not linked in the article.

https://twitter.com/RealSexyCyborg/status/1677480809450835969

I can't find the source of her saying it was about the IME thing but I recall reading that from a person close to her. She had just raised it before all this happened. Edit: Oh wait, that's here: https://skepchick.org/2023/08/maker-naomi-wu-is-silenced-by-chinese-authorities-and-why-i-blame-elon-musk/ (This was linked on wikipedia)

And yes she's a great person, she was often criticised for being a CCP stooge but that was BS. She was as outspoken as one can be being in China (and unfortunately, clearly a bit more than that).

Zworf ,

A lot of translation and summarisation. ChatGPT is extremely good in absorbing a whole mix of comments in different languages and summarising them in English (or whatever other language).

For programming I don't use it so much anymore because it hallucinates too much, calling APIs that don't even exist. And when I lower the temperature the output is too sparse.

I'm also trying to build an assistant that can also communicate proactively (I intend to auto-prompt it when things happen and then evaluate if it should cause a message to me). But I need to get a local LLM going for that because running that through the ChatGPT API will be too costly.

Also, a replacement for some of my web searches. Sometimes I just want to know something and it's refreshing that it can give me an answer (even though it does need to be validated, it's much easier to do that when you know what you're looking for!)

Zworf , (edited )

The LLMs for text are also based on "theft". They're just much better at hiding it because they have a multitude more source material. Still, it does sometimes happen that they quote a source article verbatim.

But yeah basically they're just really good copy/paste engines that work with statistical analysis to determine the most likely answer based on what's written in basically the whole internet :P It's a bit hard to explain sometimes to people who think that the AI really "thinks". I always say: If that were the case, why is the response to a really complicated question just as fast as a simple one? The wait is just based on the length of the output.

In terms of the "theft" I think it's similar ethically to google cache though.

Zworf ,

Much better for sure.

Also a lot less functionality of course.

Zworf ,

Depends on whether you consider dark-patterns to be “lying”

https://www.apple.com/privacy/ "Privacy. That's Apple.". That and then doing dark patterns, I consider that lying, yes.

Zworf , (edited )

Maybe this research and language is intended to suggest that there is a point past which “confusingly and unintuitively designed” strongly resembles “intentionally deceiving”? We’re probably not going to get internal emails saying “make it complicated so that we can collect users’ data”.

This is Apple that pride themselves on UX as you mention. They mainstreamed opinionated design. If they do it a certain way there is a reason, which is not always with the user's interests in mind. It's not because Bob in development couldn't think of a better way. Other brands might get away with that excuse but not Apple.

The race to decarbonise the world’s economy risks repeating the mistakes of the colonial era by building industries on forced and child labour, rights advocate warns (www.smh.com.au)

Almost 90 per cent of the global supply for polysilicon, a common raw material in electronic devices and solar panels, comes from China, and about half of that comes from Xinjiang, the north-western province that is home to the Uyghurs, says Grace Forrest, founder of Walk Free, a charity dedicating to fight forced labour....

Zworf , (edited )

most personal trips can be done safely and easily using an E-bike (much smaller batteries that can be produced en mass with existing supply chains) and cars should be reduced in usage outside of particularly rural areas where they truly are a necessity (which is a tiny portion of the overall population).

E-bikes are often not an option for many reasons. Needing to bring cargo, bad weather, danger from other traffic. If they were actually such an amazing option everyone would be using them because they are hella cheaper than cars. Even in the netherlands where bike infrastructure is great, people are extremely car-centric.

Personally I think subsidised public transport is a much better option.

And nuclear is not cheaper and it doesn't even factor in waste storage and decommissioning otherwise it would not have been viable. Right now when a nuclear plant is closed the operator walks off scot free and the cleanup costs are borne by the public. The mining of the uranium is also pretty polluting. There's a lot of this externalisation to make it viable.

The only reason it worked in the past was that the governments were building nuclear arsenals and invested in nuclear industry (note that this industry was not necessarily capable of producing weapons-grade plutonium but still, it was about building up an industry). It's no coincidence that most countries relying heavily on nuclear power are also nuclear armed.

Also, environmental pollution is also a safety issue. Don't just look at human deaths. Even Fukushima was a major disaster despite not leading to many deaths. The regulation is there for a reason and that still didn't manage to prevent Fukushima (not talking about Chernobyl there because that was just human idiocy fucking up at its worst). And other first-world countries have also had meltdowns.

Personally I also feel bad about dumping our waste problem on future generations. That kind of thinking is exactly what led to the climate crisis. But admittedly this is a lesser issue for nuclear in particular because we do this with pretty much everything (as this article also mentions)

Zworf ,

True, nobody should ever have billions. There's simply no need for that much money, you can't ever use it up.

Zworf ,

The Netherlands is extremely docile to big business interests.

Zworf ,

Yeah me either. It spies on your computer, they ban third-party clients. It's owned by bytedance. When I use the web version it kicks me out every day and I have to log in again.

I don't mind it being around but I really hate the way open source projects (e.g. Home Assistant) use it as their only platform for collaboration. The make me give up my data just to collaborate with them on a privacy tool.

Zworf ,

In other words: Matrix.

Zworf ,

Yes but it's a small effort to sign up for somewhere else. Matrix is just as good and they do care about your privacy.

I find it really weird for a project like Home Assistant where the whole goal of the package is to wrestle control of your home from the big tech clouds. Only to put their own comms data in a big tech cloud... :X

Zworf ,

All these maintainers of all these lemmy servers would have to do the exact same thing if Nintendo came to them.

yes but then the community would move to another lemmy host and it would turn into a game of whack a mole for Nintendo. There is no other Discord host.

Zworf ,

It's not as snazzy as Discord but it's fully open-source and federated. So everyone can run their own server (I do, too). If you don't care about running your own you can just sign up at https://app.element.io/ . It's free of course. It basically is for chat what lemmy and mastodon are to social media.

It also offers many "bridges" to other protocols, like WhatsApp, Telegram, even Discord. Those are not quite as mature and mostly third-party provided but they generally work well.

There's a really great ansible playbook for installing your own. If you would like to have the full bridged experience, beeper is probably best.

Zworf ,

Actually the ansible playbook creates a bundle of docker containers so you get the best of both :)

Zworf ,

Yeah for sure. I run the server + a bunch of bridges (whatsapp, signal, telegram, chatgpt) on an old atom NUC with 8GB RAM and it only actually uses 2 GB.

Here's the documentation for the playbook: https://github.com/spantaleev/matrix-docker-ansible-deploy

I can really recommend it. It takes some reading to set it up because it's insanely configurable. But in the end I have a config file with like 20 statements in it and that sets it all up and keeps it up to date.

Zworf , (edited )

That's true. They actually stopped supporting Nginx recently which really bothered me too because I want to keep using self-signed certs (my server is only reachable internally and I do not want to expose it to the internet). And the new server they use (I forgot which) didn't really have that option. So right now I'm locked out from updating until I fix that.

And yes it is totally feasible to use upstream! Not a problem at all.

I would recommend to use the dockers though, as the whole debian thing becomes a bit of a mess with different python requirements for some of the bridges. I tried that in a long forgotten past and there is a reason I'm trying to forget that 🤭

Like you I know the ansible playbook has its limits (for example one other thing I run into is that I want to run several instances of the same bridge to bridge eg. 2 whatsapp accounts!) but I do think docker is the way to go. I'm interested to hear how you're faring though as it's a long time ago since I tried that.

Zworf ,

On telegram it's one of the many things you get if you pay for premium.

I really like Telegram, they are really thinking about what the user wants. Live translations, icon packs, bots that can add amazing features to channels. I gladly pay for it because I use it so much, most of the communities here use Telegram.

Whereas the signal devs are just sitting on their high horse and doing nothing but stupid cryptoscams.

Zworf ,

Well the third-party launchers is extremely annoying, I have to say. Buying a game on Steam and then it forcing you to install yet another launcher (I have like 8 on my gaming PC now) really pisses me off.

I tend to buy on GOG now if I have the choice because they don't stand for that kind of shit nor DRM either.

I also really love the overview of GOG of the games you have in different launchers. Before that it happened to me multiple times that I bought a game on sale without realising I had already bought it on another platform years ago on another sale. Oops.

Ethics, features that are actually great for me instead of stuff that's just great for them. Love it. Reminds me a lot of a company that used to be like that. It was called Valve I think.

Zworf ,

Isn't that exactly what the "weight" in searXNG does?

Zworf ,

On the other hand, it doesn't really matter so much anymore.

LLM is the new search. I can ask it the actual question I have and it will give me the answer. If it's not exactly what I need I can ask it to specify further.

Contrast that with a search engine that just gives me a ton of bookmarks to sift through to see if they actually might answer my question or are just clickbait.

Of course there's still some times when you need search, like when you need to find an actual website, or when you need a source reference. But really the need for me is greatly reduced now.

Zworf , (edited )

Lol. I typed the name of my hometown and the two first results were escort sites from that area.

I mean, either it knows me really well and their privacy claims are wrong 🤭 Or it has a funny way of prioritising indexes.

Zworf , (edited )

It's in the UI on the Engines tab. However, you can only see it there, you still have to use the config file to actually change it, sorry. That's not hard at all though.

https://beehaw.org/pictrs/image/91a0bb2f-6ba4-4374-8f26-06cbd5424dbb.webp

If you don't see this option, perhaps you're running an older version? I'm running the latest docker.

Zworf ,

Not infallible truth. But very often it's something that is just for personal use.

Some things I've asked it recently were like "Which torch is smaller out of these 5 models?". Once I find which one I want it's easy to verify. Or "what does this Spanish expression mean?" or "how do I do ...".

Not everyone uses it to try and write authoritative stuff. And Google is full of clickbaity "comparison sites" that are nothing but fake advertising.

Zworf ,

Yep I hosted it on my VPN server so I can reach it from all my devices. Love it. Learned about it here and I'm really happy I did.

Zworf ,

I put in the name of my home town and the first 2 links were escort sites offering ladies in that town... Seriously.

Not really impressed so far.

Zworf ,

I know. But I'm often not really looking for accuracy. I just need to know something for myself. Most of the stuff I look up is absolutely not critically important. It's not like I'm trying to write a PhD dissertation or something.

I know it can be inaccurate but I can verify the results (and they usually are totally fine).

Zworf , (edited )

Yeah but accuracy isn't a given with the other methods either. If I ask some randos on reddit I won't get a perfect answer either. If I google specs or reviews online they are often biased, wrong (think the magical Chinese lumens of torches) or even literally fraudulent paid reviews too.

So yeah for me the LLM output is more than good enough with a bit of verification if necessary.

I don't really understand why people are suddenly hung up about holding LLMs up to this lofty ideal of an unbiased super-truth. Where did that requirement come from all of a sudden? It's not really realistic and not something we've ever had in the past.

I feel the same about self-driving systems. People get all hung up if they crash once in a while, expecting them to be 100% perfect in all situations. But ignoring the concept that they already might be a hell of a lot safer than human drivers. They fail in different situations generally but why do we suddenly demand perfection?

Zworf ,

Ah I see. I didn't really understand the requirement. That would indeed be a nice one though pretty hard to configure for general search because the results can come from so many sources.

As well as that, for special-purpose things like movies it does in fact have a ranking for those by querying common sites like IMDB directly as an engine. So in that case you can use the weighting system to show preference. It doesn't seem to support letterboxd as a source but it does some others:

https://beehaw.org/pictrs/image/1262d75e-e421-4b87-9f4f-97998d916ff7.webp

Zworf ,

Yeah I understand, Kagi is a good service!

Zworf ,

Wow..

I remember in 1993 my uni had no uplink. It was UUCP only (so just polled mail). In 1994 we got an uplink which was 256kbit shared between all sites. Luckily it came in to our facility first (the IT/Tech branch) and was cascaded further so we basically used it all 😜

Zworf ,

Yeah me too. Drinking way way way too much. Playing games in the middle of the night. Seeing weird stuff in other tents like a fullblown orgy. Fun competitions. I really miss it.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • incremental_games
  • meta
  • All magazines
    •