wmf

19 ms·

661.
by wmf 7mo ago
IBM wasted plenty of effort on Itanic but at least they were smart enough not to cancel any of their architectures.
662.
by wmf 7mo ago
Probably Intel and AMD aren't willing to do this deal but Arm is.
663.
by wmf 7mo ago
Maybe a better link: https://www.arcee.ai/blog/trinity-large-thinking
664.
by wmf 7mo ago
They could have had file permissions. They could have had a package manager instead of third-party installers. Also note that Microsoft Office has a long history of not following Windows rules. Microsoft didn't even set a good example.
665.
by wmf 7mo ago
Maybe other runtimes should copy dynamic workers.
666.
by wmf 7mo ago
I wish we would admit that you can't have it all. You can't have a product that is open source with neutral foundation governance and also have that same product be de facto proprietary. People have been pushing this bait-and-swit
667.
by wmf 7mo ago
These guys do interesting work but note that it's all on an old obscure Atom processor. I guess they can't hack any Core processors.
668.
by wmf 7mo ago
Yeah, hopefully they release >100B models.
669.
by wmf 7mo ago
Why would they not use their own AI?
670.
by wmf 7mo ago
Why not announce the funding after the milestones have been met?
671.
1-Bit Bonsai: The First Commercially Viable 1-Bit LLMs(prismml.com)
9 points by wmf 7mo ago|1 comments
672.
by wmf 7mo ago
The more things change, the more they stay the same. https://en.wikipedia.org/wiki/Whitey_on_the_Moon
673.
by wmf 7mo ago
Related: NASA's Orion Space Capsule Is Flaming Garbage by Casey Handmer https://news.ycombinator.com/item?id=45794242 Is Orion’s heat shield really safe? New NASA chief conducts final review on eve of flight. https:&#
674.
by wmf 7mo ago
They have argon gas for the ion thrusters that adjust the orbits.
675.
A second Starlink satellite exploded in orbit(twitter.com)
49 points by wmf 7mo ago|22 comments
676.
by wmf 7mo ago
Yes, it is supersampling but historically almost no one runs that way.
677.
by wmf 7mo ago
It doesn't have to be but it's really designed to run at exactly 2x scale.
678.
by wmf 7mo ago
This is not a normal retina configuration. This is a highly unusual configuration where the framebuffer is much larger than the screen resolution and gets scaled down. Obviously it sucks if it used to work and now it doesn't but almost
679.
by wmf 7mo ago
Yeah, it sounds like this Maxio controller is quite decent and it's cheap because they don't have name recognition yet. The Acer FA200 appears to be a non-counterfeit SSD using the same controller. https://www.tomshardw
680.
by wmf 7mo ago
Russia images all US bases whenever they want. That's what spy satellites are for! It's also a very small leap to assume they are selling intel to Iran. Everyone is aware of this and it's taken into account by planners on all
681.
by wmf 7mo ago
focus on chips that nobody else wants to make That's what happened here. Meta wants a Neoverse V3 CPU but no one will make it for them. So Arm has to make it.
682.
by wmf 7mo ago
Context: Early in the firmware boot process the memory controller isn't configured yet so the firmware uses the cache as RAM. In this mode cache lines are never evicted since there's no memory to evict them to.
683.
by wmf 7mo ago
https://www.cl.cam.ac.uk/research/security/capsicum/
684.
by wmf 7mo ago
People are concerned because every government official uses their personal email for work.
685.
by wmf 7mo ago
Just email him?
686.
by wmf 7mo ago
Today a local AI box (Strix Halo, DGX Spark, Mac Studio) is $2,500+. Even if it comes down, almost no one will pay upfront when they could pay $20/month instead. Small models can run on a normal PC but similar or better quality models
687.
by wmf 7mo ago
I'm not sure what you're asking. The link I posted is for a PCIe card.
688.
by wmf 7mo ago
The migration path is Thunderbolt PCIe enclosures (basically eGPU enclosures but you don't have to use a GPU).
689.
by wmf 7mo ago
I don't disagree with any of that but I don't think it's relevant to the on vs. in hair-splitting either.
690.
by wmf 7mo ago
https://newsroom.intel.com/artificial-intelligence/intel-to-... Probably 160 GB for $4,000.

More ›