Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The impact of competition and DeepSeek on Nvidia

youtubetranscriptoptimizer.com

Post reply on HN
Print view View original

401–410 of 500 posts

Re: The impact of competition and DeepSeek on Nvidia

#401

Post by ozten »

munksbeer wrote: post #347
onlyrealcuzzo wrote: post #331

Earlier quoted context omitted.

No - because this eliminates entirely or shifts the majority of work from GPU to CPU - and Nvidia does not sell CPUs. If the AI market gets 10x bigger, and GPU work gets 50% smaller (which is still 5x larger than today) - but Nvidia is priced on 40% growth for the next ten years (28x larger) - there is a price mismatch. It is theoretically possible for a massive reduction in GPU usage or shift from GPU to CPU to bene…

I can see close to zero possibility that the majority of the work will be shifted to the CPU. Anything a CPU can do can just be done better with specialised GPU hardware.

> Anything a CPU can do can just be done better

Nope. Anything inheriantly serial is better off on the CPU due to caching and it's architecture.

Many things that are highly parallizable are getting GPU enabled. Games and ML are GPU by default, but many things are migrating to CUDA.

You need both for cheap, high performance computing. They are different workloads.

Re: The impact of competition and DeepSeek on Nvidia

#402

Post by Vegenoid »

aurareturn wrote: post #382
yifanl wrote: post #348

Earlier quoted context omitted.

The usage of existing but cheaper nvidia chips to make models of similar quality is the main takeaway. It'll be much harder to convince people to buy the latest and greatest with this out there.

The usage of existing but cheaper nvidia chips to make models of similar quality is the main takeaway. So why not buy a more expensive Nvidia chip to run a better model?

Because if you don't have infinite money, considering whether to buy a thing is about the ratio of price to performance, not just performance. If you can get enough performance for your needs out of a cheaper chip, you buy the cheaper chip.

Re: The impact of competition and DeepSeek on Nvidia

#403

Post by foobarian »

vFunct wrote: post #396
lxgr wrote: post #271

Earlier quoted context omitted.

> You’re basically guaranteeing a virtual switched circuit Which means you need state (and the overhead that goes with it) for each connection within the network . That's horribly inefficient, and precisely the reason packet-switching won. > An internet based on ATM would have been amazing. No, we'd most likely be paying by the socket connection (as somebody has to pay for that state keeping overhead), which sounds h…

The cost for interactive video would be a requirement of 10x bandwidth, basically to cover idle time. Not efficient but not impossible, and definitely wouldn’t change ISP business models. The latency benefit would outweigh the cost. Just absolutely instant video interaction.

It is fascinating to think that before digital circuits phone calls were accomplished by an end-to-end electrical connection between the handsets. What luxury that must have been! If only those ancestors of ours had modems and computers to use those excellent connections for low-latency gaming... :-)

Re: The impact of competition and DeepSeek on Nvidia

#404

Post by ozten »

NVIDIA sells shovels to the gold rush. One miner (Liang Wenfeng), who has previously purchased at least 10,000 A100 shovels... has a "side project" where they figured out how to dig really well with a shovel and shared their secrets.

The gold rush, wether real or a bubble is still there! NVIDA will still sell every shovel they can manufacture, as soon as it is available in inventory.

Fortune 100 companies will still want the biggest toolshed to invent the next paradigm or to be the first to get to AGI.

Re: The impact of competition and DeepSeek on Nvidia

#405

Post by foobarian »

paulddraper wrote: post #283
pjdesno wrote: post #176

The description of DeepSeek reminds me of my experience in networking in the late 80s - early 90s. Back then a really big motivator for Asynchronous Transfer Mode (ATM) and fiber-to-the-home was the promise of video on demand, which was a huge market in comparison to the Internet of the day. Just about all the work in this area ignored the potential of advanced video coding algorithms, and assumed that broadcast TV-q…

I love algorithms as much the next guy, but not really. DCT was developed in 1972 and has a compression ratio of 100:1. H.264 compresses 2000:1. And standard resolution (480p) is ~1/30th the resolution of 4k. --- I.e. Standard resolution with DCT is smaller than 4k with H.264. Even high-definition (720p) with DCT is only twice the bandwidth of 4k H.264. Modern compression has allowed us to add a bunch more pixels, bu…

I'm sure it helped, but yeah, not only e2e bandwidth but also the total network throughput increased by vast orders of magnitude.

Re: The impact of competition and DeepSeek on Nvidia

#406

Post by snailmailstare »

aurareturn wrote: post #385
onlyrealcuzzo wrote: post #331

Earlier quoted context omitted.

No - because this eliminates entirely or shifts the majority of work from GPU to CPU - and Nvidia does not sell CPUs. If the AI market gets 10x bigger, and GPU work gets 50% smaller (which is still 5x larger than today) - but Nvidia is priced on 40% growth for the next ten years (28x larger) - there is a price mismatch. It is theoretically possible for a massive reduction in GPU usage or shift from GPU to CPU to bene…

No - because this eliminates entirely or shifts the majority of work from GPU to CPU - and Nvidia does not sell CPUs. I'm not even sure how to reply to this. GPUs are fundamentally much more efficient for AI inference than CPUs.

I think SIMD is not so much better than SIMT for solved problems as a level in claiming a problem as solved.

Re: The impact of competition and DeepSeek on Nvidia

#407

Post by chrisco255 »

Vegenoid wrote: post #398
munksbeer wrote: post #347

Earlier quoted context omitted.

I can see close to zero possibility that the majority of the work will be shifted to the CPU. Anything a CPU can do can just be done better with specialised GPU hardware.

Then why do we have powerful CPUs instead of a bunch of specialized hardware? It's because the value of a CPU is in its versatility and ubiquity. If a CPU can do a thing good enough, then most programs/computers will do that thing on a CPU instead of having the increased complexity and cost of a GPU, even if a GPU would do it better.

We have both? Modern computing devices like smart phones use SoCs with integrated GPUs. GPUs aren't really specialized hardware, either, they are general purpose hardware useful in many scenarios (built for graphics originally but clearly useful in other domains including AI).

Re: The impact of competition and DeepSeek on Nvidia

#408

Post by 0n0n0m0uz »

Please tell me if I am wrong. I know very little details and heard a few headlines and my hasty conclusion is that this development clearly shows the exponential nature of AI development in terms of how people are able to piggyback from the resources, time and money of the previous iteration. They used the output from chatgpt as the input to their model. Is this true, more or less accurate or off base?

Re: The impact of competition and DeepSeek on Nvidia

#409

Post by nayuki »

WhitneyLand wrote: post #355
paulddraper wrote: post #283

Earlier quoted context omitted.

I love algorithms as much the next guy, but not really. DCT was developed in 1972 and has a compression ratio of 100:1. H.264 compresses 2000:1. And standard resolution (480p) is ~1/30th the resolution of 4k. --- I.e. Standard resolution with DCT is smaller than 4k with H.264. Even high-definition (720p) with DCT is only twice the bandwidth of 4k H.264. Modern compression has allowed us to add a bunch more pixels, bu…

DCT is not an algorithm at all, it’s a mathematical transform. It doesn’t have a compression ratio.

Correct. DCT maps N real numbers to N real numbers. It reorganizes the data to make it more amenable to compression, but DCT itself doesn't do any compression.

The real compression comes from quantization and entropy coding (Huffman coding, arithmetic coding, etc.).

Re: The impact of competition and DeepSeek on Nvidia

#410

Post by eigenvalue »

latchkey wrote: post #231
saagarjha wrote: post #43

Earlier quoted context omitted.

*George Hotz is making posts online talking about how AMD isn’t helping him

George Hotz tried to extort AMD into giving him $500k in free hardware and $2m cash, and they politely declined.

Was arguably not that polite and caused them some bad PR IMHO
Post reply on HN