401–410 of 437 posts
I'm waiting for powerful on device LLM models, since that not worth it
Have you tried Qwen 3.6 or Gemma 4? They're not frontier level but certainly have their uses.
Earlier quoted context omitted.
Right, the export controls are only forcing Chinese AI to innovate, build their own fabs, and make training and inference more efficient. The end game of this will be NVIDIA chips won’t be wanted because you can get a $50 chinese chip running a ternary model that is competitive with claude in English and is much better in Mandarin.
The US government has failed to learn from its own history. 60 years ago the US government had forbidden the export of fast computers to France, with the hope that this sanction will prevent the French from developing thermonuclear bombs. The result was that the French state (which at that time was lead by de Gaulle, not much less autocratically than China) subsidized some of their computer manufacturers, which previ…
When De Gaulle did ask the french via a poll if they wanted him to leave, they said yes, and he left. He is also the guy who did setup the balance between the various political powers, which has been kind of working... until now (currently the government can hardly get laws from the parliament, because few people representatives are on gov side, and they won't die or disappear if they disagree, bugger!). The fact that the president must leave after 10 years is kind of recent though.
France has always been a very strong US ally, in an honest relationship, namely without agreeing or being on board with everything. And France never had the intention to nuke the US... unlike some other country we talk about all the time in news (that said, France is not far from the US on their list...). And compared to the rest of the world, don't forget the 'western world' (which is not 'western' only anymore...) has very, very close core values. A good way to think about it: a big dysfunctional family.
On the software side, aka the 'silicon master control' side: currently, the french are just Big Tech slaves. To be more current, Holland president and Valls prime minister did install a document (2015/2016) which has been "law" since which literaly "pushes" (hard) administration online services to be hardcore dependent on Big Tech (mostly the whatng cartel) without any reasonable technical way out (unless noscript/basic HTML web sites, are brought back in the security infrastructure, like they were a few years back). This document is out of reach of even the parliament, namely only the president and prime minister have control over it, in other words, to interact with this document you need the same level of power required to decide to increase the number of atomic bombs(huh). The following president and prime ministers did nothing and kept increasing french administration dependency, I guess they were/are as guilty OR BRAIN WASHED than Hollande and Valls.
Open source does not matter anymore (look at how big tech controls open source software via often-non-pertinent complexity and size), _LEAN_ open source does, and that includes the SDK (aka the computer languages: if you need a giga huge and complex compiler, you already lost).
On the hardware side, state-of-the-art chip is an international effort with an insane supply chain. This is mostly 'driven'/hogged by US chip designers. State of the art, foundries are currently in TW (the US is working at getting some back), EUV is from EU-ish (the EUV light is from the US), and many, many more high-hech tools are from the US/JP/TW/etc.
What I am wondering: did Holland and Valls "give" France to Big Tech... or "sold" it, if you see what I mean, because it is very easy to setup public money channels using 'Big Tech' which look "clean", aka hidden behind a technologi-blablublo smoke screen, since most people are scared of tech and/or don't understand the fine details.
It is all about simple file formats and network protocols, good enough to do the job and stable in time. A good compromise is to use a strongely and dynamically defined subset of Big Tech stuff, which you know can be locally implemented with reasonable effort (by citizens, small companies, state administrations, etc). That will foster alternatives (good I guess). That's why I am talking about web sites, and not web apps (noscript/basic HTML), and we could talk about a strongely defined subset of PDF.
Ofc, the devil hides in the details, this is a very coarse overview: you have to basically decide in a fine-grained case by case, mistakes will be made and will have to painfully be fixed. You cannot get it all in one shot, it is module per module, back and forth, and probably slowly.
Earlier quoted context omitted.
This thread is almost 1:1 identical to when Apple released their own silicon. This has the potential to be a worthy competitor for the Windows ecosystem, precisely because of NVidia’s moat as the grandparent pointed out. Microsoft pulls in their weight as well, so this seems like it has a decent chance of getting industry support.
Yeah but like, Apple put Rosetta 2 out and it was damn good. Vendors didn’t have to do shit to support the platform, they just got better performance if they did (like factorio). There is something of a difference between “all your stuff will still work, at comparable performance” and arm windows which (as evidenced by all the vendor promises) you can’t really currently say with prism. I would describe prism as “surp…
For decades, Windows made it too easy for games and even some application to install drivers. Windows games use drivers for anti-cheat (and historically for copy protection too). Neither Apple Rosetta nor Microsoft Prism can translate/emulate drivers, but since drivers have been much more prevalent on Windows, now Windows has a much biggest compatibility problem.
Earlier quoted context omitted.
> just like their GPUs are So with proprietary blobs that give you more trouble that they're worth?
I've never had a single problem with my Nvidia GPUs on Linux.
But once we're talking about laptops, hybrid graphics etc. it quickly shows that this is not a platform Nvidia cares about.
Earlier quoted context omitted.
You misspelled llama.cpp
I’ve read all the stuff about how llama.cpp is much faster and better than ollama, and i believe it - but good god llama.cpp isn’t user friendly. You’d think in an era where “code is free” there would be an easier story around running local ai than compiling llama.cpp by hand and then spending hours researching flags - only for it to crash from an oom error every ten prompts or so.
Earlier quoted context omitted.
The ConnectX 7 2x200 Gbps networking card in the DGX Spark alone is worth $700
To be fair the connectx-7 in the spark can't even push 2x200 Gbps since it is connected via 4 pcie lanes.
https://www.nvidia.com/content/dam/en-zz/Solutions/networkin....
Lots of comments are expressing skepticism about compatibility but it's pretty cool how Nvidia has the clout to convince a bunch of game publishers and creative apps to release Arm versions. Popular games like League of Legends as well as stuff like Adobe Photoshop and Premiere are getting native Arm ports. > Over 100 Windows software providers such as Adobe, Blackmagic Design, Blender, CapCut, ComfyUI and OTOY, and…
Only the ones which explicitly list something like the Riot Games mention are really related to the device/Nvidia. The thing which really pushes this along is user adoption/market share, not big names. This device will help that, especially in the gaming space, but it's easy to get over eager as it being from Nvidia means everyone else who has been waiting will just now jump on board too because of that.
Some competition for Apple in this space and competition for Intel and AMD is great. But I really do question how well Windows on Arm is really going to work out long term. For Apple it worked because they were able to force the issue. If you wanted a new Mac it was going to be Arm and we all knew eventually (this year or is it next year?) Intel support would drop. Over time we have seen M series exclusive features.…
For Apple it worked because they waited until they had a really, really good ARM ISA CPU (combined with arguably sandbagging their x86 offering for a few years prior but I digress). Qualcomm is also working on a really good ARM ISA CPU with their acquisition of NuVia and subsequent Oryon architecture. Meanwhile this is just using off-the-shelf ARM CPUs in a MediaTek SoC with blackwell bolted to the side of it. ARM's…
MediaTek is involved in the SoC but both the CPU & GPU from Nvidia are bolted on to it. I.e. it's not a standard MediaTek CPU with an Nvidia GPU added.
Earlier quoted context omitted.
At this point Valve look more capable of running a platform business than Microsoft do. Microsoft have spent the whole Nadella era in "oooo cloud" inspired wonder and actively screwed up everything else.
And yet where are the native games for the platform? Note how game developers rather spend their working hours in Windows with all its issues, even if they happen to have Linux servers for running their MMOs.
My Steam library is full of old win32 games that run better on Linux than they do on Windows 11. There are some native games appearing because of the Steam Deck, but the fact is they aren't necessary.
Valve aren't simply better at running a platform business, they've thoroughly subverted Microsoft's old one and have done a better job at running that than Microsoft themselves.
Look at the absolute state of the XBox business: all native games, tens of billions spent on something, and yet it's just a trainwreck from top to bottom.