HN user
evil-olive's comment includes citations for fake information like the "nanometer" feature size of the chips. everyone knows those numbers are meaningless! they have no relation to the chips' ability to survive radiation in any configuration! but evil-olive adds the citation chaff knowing HN will upvote to avoid feeling bad about having to disagree with a strident "absolutely never going to happen" claim.
What's the incidence of autism in country towns without functioning septic systems?
What even is an industry grade machine?
Would you own a computer if the smallest computer you were allowed to buy was a $27,000 Supermicro rack that draws 900 watts all the time?
There are also "natural" paracrystalline viruses that turn isopods blue and then kill them.
https://en.wikipedia.org/wiki/Invertebrate_iridescent_virus_...
The issue author is most likely quite literally a child.
Nvidia didn't ship a 256gb system at sub-500gb/s transfer rate
DGX Spark has 128 GB and only 273 GB/s BW. Are we lucky that NVIDIA did ship something even worse than what you specified? I'm confused.
People have been complaining [1] about how little VRAM NVIDIA ships with their GPUs for decades. Their whole game has been "oh, you want more VRAM? Buy more or pay us 50x for server grade with 10x as much VRAM. The more you buy, the more you save."
Apple did everyone a solid by shipping something way out of that distribution. We now know more than we did before! We know that a 284B parameter model with 13B active params (or 35B with 3B active, or 671B with 37B active) can outperform a 2T model and draw a fraction as much power. How can you think that's a bad thing?
You could point out that Apple didn't invent the idea of MoE. Everyone knows that. But other than Macs, there simply were no machines with >100GB VRAM directly coupled to ~50 TFLOP/s of compute until the DGX Spark last Dec. If you wanted to run a model with more than 32 GB of weights, you had to either pay up for dozens of GPUs idling at hundreds of watts or really pay up for some $50,000 server GPUs idling at... also 100-200W each.
I feel lucky to have a $3k machine on my shelf that can run DS4-Flash with 1M context at 20t/s while drawing ~150W and making very little noise. The best part? It idles at 30W with DS4 loaded, dropping to 6W after a reboot. There isn't a single GPU on the market that can match that in the same shoebox volume.
[1] https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcRlOW0N...
Have you heard of time-domain reflectometry? A $20,000 multimeter could have the "impossible" feature you describe all but built in.
Please show me on the 2016-era $400 Pascal GPU where you can install the 256 GB of VRAM.
It doesn't really read like a Luddite policy.
Rust is already well past 1.0. At best an LLM could discover a vulnerability (and the human using it can file a patch) or can help a human improve ergonomics.
Nope.
That prefill number isn't right. M4 Max hits 200-300: https://github.com/antirez/ds4/blob/main/speed-bench/m4_max_...
At a certain point the level of abstraction / genericization necessary for a big flexible project (like llama.cpp or Linux) blows things up into a huge number of files. Something newer and smaller can move faster.
Straight propaganda. Printers have burned down apartments in the last five years.
https://www.reddit.com/r/anycubic/comments/1j4kfsr/guys_just...
On both the llama.cpp based version and the custom Metal version, the model forgets how to use tools somewhere around the 50,000 token mark.
It is still available for the M5 Max Macbook Pro, but yes, the Mac Studio is now only offered with up to 96 GB.
The world is not China.
Let's be realistic. GM is not interested in burning R&D money to make a good 800V electric crate kit. They only want to burn R&D money to make new profitable trucks.
They are doing this because they have a billion dollars of refurbished batteries sitting in warehouses that they need to dump.
You can buy head unit parts on eBay. It sounds like you have rodent harness damage.
Thanks
It includes a 66 kWh battery. This might be a strategy for offloading all the Bolt EV batteries they had to refurbish.
Doy!
Why would you lie about this?
which is the exact same reason China bans Starlink.
so, to be clear, you're not pissed at Tangled because they're building an open-source, freely available tool that helps people develop in public without demanding anything in return, but you are pissed at them because they took VC money and are using atproto, the latter of which in some way legitimizes crypto / billionaire techbros?
if so, you value ideology over pragmatism to the extent of "cutting off your nose to spite your face"
in your world, literally who isn't a dickhead? yourself, presumably? you must make a living somehow.
catpart is definitely NOT someone with a long history of railing against anything vaguely connected to ATproto: https://news.ycombinator.com/item?id=46670016
catpart has totally 100% provided a citation for the "fuck the users" moment (sike): https://news.ycombinator.com/item?id=46672904
the actual "fuck the users" moment is described here [4,3095771], wherein a group of activists failed to get a person (Jesse Singal) they deemed unpalatable deplatformed from Bluesky, mostly by claiming that this person broke the ToS, and the Bluesky moderators mocked the activists in public.
[4] https://techcrunch.com/2025/10/05/waffles-eat-bluesky/
[3095771] https://en.wikipedia.org/wiki/Jesse_Singal#Subsequent_events
I think they're mad because 'knot' is a furry fetish term.
Ferrules often have insulating shells around the base which are good strain relief.
Ferrules also evenly distribute the clamping force and prevent crushing of the strands.
Ferrules solve the problem of having to align the threads every time you re-post them in a terminal block.
What if they're stupider because of it?
The plan fell off.