there are two predominant architectures right now (right or wrong), amd64 and arm64. Why the F would amd invest in risc when their gpus are well above intel in specs and explain the biz market approach for risc...
All this money saved should be "unsaved" until a decent alternative is made. Everything that the government should spent here, should be invested into pan-european organization to develop a new office suite.
Libre office in my opinion is one of the reasons Microsoft is so dominant. Unfortunately, libre office, even though useful, is one of the worst desktop applications to use.
Everyone I proposed this to tried it and said that its horrible and they don't want to use it. And I agree with them: because libre office is so sh*t, u use Google docs.
> Libre office in my opinion is one of the reasons Microsoft is so dominant.
Microsoft Office was already dominant long before LibreOffice started. Hell, MSO was already dominant when StarOffice was renamed OpenOffice.org, long before LibreOffice was a thought in anyone's mind.
> Unfortunately, libre office, even though useful, is one of the worst desktop applications to use.
You only feel this way because you're used to MS Office. Ask anyone who's more well versed in Google Workspace and they'll tell you that MS is difficult to use.
Yes Microsoft Office was already dominant. Why didn't libre office affect this dominance being 'so good'? It's also free, so it should be interesting to companies, right? Wrong. I recently set up an SME who when saw it ran away and immediately bought a Microsoft license, which means they would rather pay and be spied on than use libre office.
Believe it or not, people like nice things. Microsoft Office looks nice. Libre office looks like a car accident. It's shallow, I know, but this is the response I get every time.
Everyone I know hates it. It is a small sample, true, but it says something. So I think, the fact Microsoft is so strong is to be blamed on the alternatives, or in this case only one alternative.
Turn it around. If it would be good, people would jump on it. Especially small companies.
Google workspace is a web app, no? So it being better than desktop is comparing different things. I use it and I like it, especially since I don't need advanced Features Microsoft offers
We have no good alternative. We have an alternative.
> You only feel this way because you're used to MS Office.
You are kind of proving the point. You point to a Google product as a good alternative, not to the OSS product. I really want to like LibreOffice and it is a good product for what it is, but it is far from being a great product.
Libreoffice is fine, better than we deserve. If they want it to be even better, maybe instead of throwing a few hundred million at MS, they could throw a few hundred million at Libreoffice. It's old as hell and kept up by charity.
The idea that Libreoffice is so bad that giving up your freedom to Google or Microsoft is unavoidable just shows your actual level of objection to being slaves to US companies is close to zero. You'll only be pried away from your dependence on the latest popular versions of US products kicking, screaming, and complaining the entire time. You wouldn't be satisfied with anything but a clone, and you'd complain that the clone lacked the most obscure features of the real thing.
And it's not just you, but a typical sort of aimless ridicule of FOSS product from people who feel guilty about not using them when their professed politics say they should. You'll talk a big game about independence, but your fictional pan-European office suite is far worse than Libreoffice, seeing as it doesn't exist. Couldn't be more feature-light.
I disagree completely. If libre office would be fine, it would be popular. And, as said, most of the people I know avoid it like a plague.
Also, im not trying to ridicule FOSS as a whole. if anything I'm a financial supporter for several projects and organizations. It's not a lot of money, but it's every month. So, no, it's not this.
The proposal would be to fund something like collabora, build on top of libre office or do something greenfield, but all this money that was supposed to go to Microsoft should be redirected.
Decision that governments did are not based on the fact that libre office is good. It's based on 1) political reasons called digital sovereignty and 2) price. Maybe 3) being pissed at trump. They didn't do it because LO was good.
Some american products have no good alternatives, yet. Some do, like windows can be replaced with gnome, but mobile phones cannot. Probably you are not typing or reading this from a European os on your phone, yet alternatives exist. Just not good ones.
Microsoft Office clone is not what I want, but what I would accept. Let's say UI should be very similar to Microsoft Office or even better Google office. That's it. Make it in a desktop suite and we are all good.
Every single time I’ve installed anything with snap I’ve regretted it. It causes all sorts of annoying issues, including e.g. preventing integration between apps because of its supposed security model. This is typical of consumer operating systems, like Windows, but not what should be encouraged on Linux.
On top of that, Canonical is pushing snap very hard. Try to uninstall the Firefox snap - e.g. because it doesn’t integrate well with password managers - and install it using apt from the Mozilla repo. Ubuntu will later just silently replace it back with the snap version.
I’m about to switch away from Ubuntu as a result of this.
The base model is a pure next token predictor. It just continues whatever prompt you give it — if you ask it a question, it might just keep elaborating the question. To turn these models into something that can actually chat (and more recently, that can do things like tool calls) they do a second phase of training, including reinforcement learning, which teaches the model to maximize some kind of reward signal meant to represent good answers of various kinds. This reward signal applies at the level of the whole response (or possibly parts of the response) so it is not predicting the most likely next token. I don’t know in an absolute sense how much this ends up changing the base model weights, and it’s surprisingly hard to find discussions of this, I guess because the state of the art is quite secret. But it’s clear that RL is important for getting the models to become useful.
There are other posttraining techniques that are not strictly speaking RL (again, not an expert) but it sounds to me like they are still not teaching straightforward next token prediction in the way people mean when they say LLMs can’t do X because they’re merely predicting the most likely next token based on the training corpus.
Thanks for the explanation and the link. I learned something today.
I'm definitely not an expert, but to me, RL and other techniques looks like a guide or a constraint on the still 'next token prediction' concept. What I do not get is - is this all about training? Or is this about inference.
In any case, this is still an eye opener and I need to study this a bit more.
When talking inference, models from huggingface are composed of what then? Because they can do angentic stuff, no?
2 out of 3 points should not be in the list at this point of time.
There is no browser around servo that is usable and they start with accessibility?
The effort should go first to have it work as an everyday browser and get as much market share as possible. Then think of accessibility and WebView Integration.
This is a unique opportunity to have a browser engine not depending on Google, Microsoft or some of AI companies wanting to get into this field. By watering it down, you will just lose the market to the ladybird which is focused.
1. Accessibility is easier to build in at the start rather than trying to tack it on at the end.
2. Building a browser around an existing webview is fairly easy compared to all the other parts that go into the finished product.
3. Making the webview integration better means more than just a single browser are using it (which would increase the user share that you deemed as important).
Having an MBA for strategic business decisions probably is more relevant than having an engineer aside from some highly technical niche sector. You need an open culture to discuss strategy and involve different perspectives (No, it doesn't matter if your MBA of choice is black or not).
Most MBAs know too that wanking to quarterlies doesn't make a good long term strategy. Most companies just forget to do any long term strategy as long as such basic controlling data points look decent.
Do that and you will be ass raped by stronger powers in record time... if we don't start a war in eu amongst ourselves.
The eu was established as an economic union trying to prevent war in Europe and it has been very successful in that. Going the other way would probably cause havoc. Look at Hungary and Slovakia even with war on our doorstep.
I am of the opinion that the EU needs to federalize. It needs to look as one country from the outside, while having independent countries on the inside. The goal here is not to be overwhelmed by the economic and military might of china. Build up resiliency and self sufficiency and keep democracy alive, which seems to be on the decline everywhere.
The question is how not to end up in a totalitarian regime, which chat control is about. It doesn't and it never will be about the children.
But, wouldn't it make more sense for amd to go into risc-v at this point of time?
reply