bahmboo

aka Rocky (this is a reference to Project Hail Mary). Great descriptions in the book about how the fictional creature can use the arms together or differently and on multiple tasks at the same time.

I'm optimistic this won't be a problem in 10-20 years (I agree that's a long time). We have the technology in matrix headlights that were banned in the US and basically in Canada up until a few years ago. I expect to see kits on Amazon in a few years for retrofits. Since they need a camera and some brains they will likely start at many hundreds of dollars but that should come down.

throwaway account and "virtue signalling" plus a dig at so called subsidies. Oh and "no value judgement" except for the value judgement. You are not helpful here.

For sure. Clearly Nvidia mops the floor with the competition. I'm looking forward to M6/M7 and to see if Apple wants a bigger piece of the pie.

That's not what I'm seeing. It is much faster but not an order of magnitude. Not trying to be pedantic, only setting expectations.

"The Blackwell RTX PRO 6000 provides up to 1,792 GB/s of memory bandwidth, while the 40-core Apple M5 Max tops out at 614 GB/s"

Yes. The memory is just located very close to the cpu with wires "welded" directly to it. This allows the memory to be run as fast as possible but it's still a RAM component.

The cache parts of memory are on the CPU itself but they are on the order of MB not GB.

To be fair it's "only" half the throughput of a 4090 and a third of an RTX 6000. Significant but not an order of magnitude.

Yah, some of us use integer _math_ extensively. It's not just for arrays.

Another recent project that runs a huge model on a 48gb Mac is https://github.com/danveloper/flash-moe - it gets over 5 tokens/sec on an M3 Max compared to this projects very impressive 1 token/sec on an M5 Max. So for anyone wanting to tackle a Mac only version that targets lower spec machines this looks like a good candidate with plenty of room for speedups [edit: because it doesn't use the gpu].

Not hijacking anything as this project is amazing.

Some people are legit and some people are not. Got it. Thanks for the clarification and analysis.

I work with a lot of 3D graphics and geo stuff so I can hit the ceiling with my 48 GB mac. It's not all LLM work. I prioritized more storage than RAM with my budget. Being able to run local llms has greatly helped me understand how they work. For day to day dev I pay for Gemini or Claude.

RAM is a commodity and nvidia will be paying the same prices. The used market will reflect the cost of RAM. nvidia owns the top of the market but many of us don't need that.

I agree. It was very annoying to me to spend the money (and on the nano matte one too) and still have that stupid notch. But it never makes any difference at all which is good news.

Yes the MBP is very quiet when doing easy stuff but run games and heavy duty software and those fans are required to get work done.

How do you know that? What information do you have that would explain your position? We are talking about a specific circumstance and you have brought unsupported generalities to the discussion.

Their recommendation engine has surfaced awesome obscure content I would have never found otherwise so I value it.

Stuff like random recorded conference talks with 3 views. A super enthusiast in Latvia.

It does recommend crap sometimes but on balance I like it.

The are guilty of letting these terrible pilots fly humans over oceans. Sometimes the driver is bad and yet we point at the car and say it should have been designed "better". I have read a lot about this flight over the years and I have my obvious opinions.

potato potatoh

$2600 gets MBP M5 Pro 48gb. 64gb requires a Max which bumps it to $4200 at which point you may as well spend the $800 to go to 128gb.

How many phone numbers do you have memorized? These days I only know a few but I used to know dozens.

I run local models. They are very good now (roughly 30 billion parameters).

There are no sources or references.

Y'all look at as a paradigm shift. The rest of the world just wants to touch the screen when it makes sense.

We trained billions of humans to touch the screen.

Because he is explicitly talking about x86 - maybe you missed that.

This should be the link. Thanks

Yes this jumped out at me too. It isn't remotely true. The opposite is more accurate: I'd wager that at least half the trucks on the road are built more like small planes from the 60s WRT to operational systems.

Hindsight. They should have tried given they didn't know. I remember it unfolding at the time and astonished at the lack of Russian response.

i.e. dumb people don't know they are dumb

Bummer

It doesn't compile for you or doesn't compile at all? Honest question. It's a nice project in the face of it but if it's all AI fever dreams that would be disappointing. I don't have the cycles to try it out right now.