kcb

The igpu is a tiny RDNA2 GPU. Pretty certain the 9950x cpu will outperform it in compute tasks.

Gemma 4 wouldn't really be a competitor to this. Gemma has the dense 31b model, but this has like 25x the total number of params.

Why? What unsuccessful game does Blizzard have?

Ability to run any OS is a pretty nice benefit versus the spark.

LLMs are still easily replaceable. If the SOTA frontier model provides meaningful impact for your critical business function, then worse case you flip the switch to the next most capable model.

You really think China doesn't have massive amounts of capital expenditure related to AI? They're actively bootstrapping an entire chip industry.... https://www.reuters.com/world/china/china-prepares-295-billi...

No reason to use fiber on short runs like that. DAC cables are cheap and better in pretty much every way over short distances. You're probably thinking of RJ-45 NICs and SFP modules which are known to run pretty hot.

Nvidia's Nemotron 3 Super also shipped with MTP.

Google and Nvidia...

Yes, all the closed providers are probably doing this already. As well as open models like Gemma and Nemotron.

The product here is the research and licensing the tech.

That's not it. They will keep the assets closed and under license. There will be no free version of the full game.

Not sure why we should encourage using open source as a vehicle to market and get free work building your fundamentals. Just to reap the profits yourself later.

It was. Not anymore though.

I think their plan is to make assets closed licensed. It's all pretty lame. Now any open source code contribution is packaged up and sold with someone else getting the check.

It's Apple, so it's some revolutionary big brained play, and not just yet another llm sdk.

With 2 Blackwells, would make sense to run NVFP4 quants

In 2014 Tesla cars were endemic, you could see them everywhere

Tesla sells like 8x the number of cars in Norway today than they did in 2014.

I think Vulkan Video is just another api to access those hardware decoders. It's not going to bring support for codecs to hardware without the support.

The same chip that's been available in the DGX Spark for like 8 months now...why are we pretending like its the next big thing.

A real example https://github.com/OHF-Voice/micro-wake-word

This will likely have worse single threaded performance than recent Qualcomm CPUs.

So in the end, the government still needs a department that assesses properties to determine if the owner has undervalued it.

https://cdn.wccftech.com/wp-content/uploads/2026/04/NVIDIA-R...

SRAM is still high bandwidth memory...it's literally the most space inefficient and expensive type of memory.

Any system mandated by the government will have a backdoor to deanonymize users. Nothing would convince me otherwise.

The point is metric shit-tons of SRAM is still large amounts of expensive memory.

There's no magic bullet for inference on cheap accelerators. Any accelerator will still require large amounts of high bandwidth memory.

Elon Musk has been failing any minute now since like what? 2015

This reads a lot like "the way I choose to live is the best and everyone else is sad." Anyone in a dense suburb is getting all the fresh food they want from a choice of 6 different grocery stores. And it's silly to complain about suburbs being crowded in comparison to cities.

Nonsense, Apple has on package memory and the primary reason for that is overall packaging and layout not performance