looks like I made eurofash mad again 🤣
@lemmy.ml
Dream might be a good starting point actually. It's been last updated a year ago, but it does have a solid core for decentralized streaming on top of bittorrent and kademlia DHT. It handles upload, search, and stream videos without central servers.
That's a really good idea given all the problems with youtube and that they can just take down videos or channels without notice. One thing to explore would be having discovery without necessitating tracker sites. Stuff like dht-spider might be worth a look.
It might also be interesting to take an existing GUI like Freetube and try swapping out the backend with a torrent based one.
Sure, I'm not arguing there isn't a gap currently. I just don't see the gap as being consequential in any way that matters already.
It seems like Tau architecture might close the performance gap even without EUV though. Similarly Huawei managed to get clever with optical connections in their Ascend clusters to actually outperform NVIDIA for AI training. https://www.scmp.com/...
So, the gap really isn't that big in practice. Also worth noting that software side plays just as much role here as well. Android, iOS, Windows, MacOS, and even Linux all have a lot of legacy decisions baked into them for backwards compatibility needed to support existing software. Huawei building a fresh stack on top of HarmonyOS allows them to make a much leaner stack that's not saddled with all the prior baggage. And that can make overall user experience a lot better even on slower hardware. Modern software is incredibly bloated, and addressing the bloat is a low hanging fruit that can be plucked right now without the need for EUV machines. The benefits will stack with faster chips as well just the same way Tau will stack with EUV.
In a way, decoupling from Western tech stack could actually provide a lot of benefits because it opens up the way for doing things differently without having to worry about the way existing legacy stack works.
Right, and on top of it Huawei is coming up with new ways to arrange transistors with stuff like Tau folding architecture, which combined wtih EUV might actually allow Chinese chip makers to push far ahead of traditional chip designs.
I see no issue with locally run models like Z-Image Turbo. They take very little resources to run, so there's no real issue with energy use here. They produce good results when prompted properly.
The whole theft of artist work applies to commercial models where companies are profiting of other people's work. However, I simply don't see the argument when it comes to open models that anybody can use freely, and especially in cases where you're not producing images for profit.
They started destroying bridges across the Dnieper in the south. Where supplies come from Romania.
Amazed it took them this long to finally start cutting supply lines from Europe.
oh and here's how relations between China and the west are actually like https://www.youtube.com/watch?v=kVOS6RVr6H0
China does not yet have the technology to overtake Taiwan.
China absolutely does have the technology to overtake Taiwan, and if you look at the history of how fast Chinese technology develops, it should be obvious that it's not going to take long. Look at what happened with solar panels, EVs, batteries, phones, etc. In every case, once China ramped up research and production, they leapfrogged the rest of the world within years. China sees chip production as a national security issue, they will be pouring state level resources into it.
Meanwhile here's what Stanford has to say about AI https://hai.stanford.edu/...


Not only has China basically closed the gap already, but they're doing it at a tiny fraction of the cost.
It's basically a bank bailout in disguise. So many of these properties are financed by multiple banks as the developer struggles to stay afloat. Can't afford your $200m load, take out another one + $50m. Too big to fail at this point.
I actually think substack format would be a better fit. It's already a blogging platform at its core with good discoverability. That's exactly what sites like substack provide and why they're popular for blogging. Reddit/Lemmy is more of a news aggregator where people post links and discuss them.
Reddit does have the advantage of being a single site where search is easy to do in a shared db. Since Mastodon is federated, discoverability is going to be inherently worse as requests have to propagate through the network. But I think that for blogs it's less of an issue since you tend to follow people for their writing.
In my opinion, Lemmy is already a great replacement for Reddit. So, it makes sense for Mastodon to focus on its core functionality which is blogging, while Lemmy can fill the Reddit niche in the fediverse.
thanks for using Leebra!
go to feed...