“Meta also plans to design a new chip every 6 months beginning next year, which would be significantly faster than the standard industry cadence.”

ConceptHardware

“I am very excited to see what this model, a really, really big model, looks like when it's hosted on Cerebras. Because I think that's gonna start changing the vibe of how you make stuff.”

ConceptHardware

“Because as the models get bigger, you now need to move the weights and of course what we call the KV cache into the compute units. And that is essentially a data movement problem, right? And it's a data movement problem from the memory to the compute units. And it's a data mov...”

ConceptHardware

“It's a classic idea from computer architecture, pipelining and the use of a memory hierarchy to move the data when you need it to where you need it at the right time. And so the nice thing about these AI models is that you do have a graph of computation. And the whole idea of ...”

ConceptHardware

“So I think that's like one of the big surprises. And for us realizing that was like this big moment that validated something that we always strive for, is to create extremely efficient models. Because I think once you start to realize that the robot will need to create like th...”

ConceptHardware

“So it's been over half a year. Kind of surprising that they haven't doubled down initially. It was a big deal, but regardless, makes sense that we have a new generation here. And now to applications in business. First up, we've got Etched pools 400+ engineers from NVIDIA, TSMC...”

ConceptHardware

“So Etched is one of those companies that they are new in terms of like seemingly coming out of nowhere and they had been working at this for about 3 years, but that's a really short period of time for being as far along as they, as they are or appear to be. So currently they a...”

ConceptHardware

“So this is a really big deal because usually these sorts of things will take like many, many years to pull off.”

ConceptHardware

“Yeah, that's another one too, you know, investment opportunity also gets cut off, right? Acquisition gets cut off. So meanwhile, that's all on the software side. On the, on the hardware side, DeepSeek says they're going to develop their own AI chip, um, right? Which, which is ...”

ConceptHardware

“Yeah, yeah, it's no good. Um, and every— I mean, DeepSeek's also not alone. It seems like this is another trend that's happening right now, is that there's a lot of, um, a lot of the big companies are, are trying, doing their best to create and develop their own chips, which s...”

ConceptHardware

“Doing okay. Yeah, but Nvidia has definitely been the company that over the last couple of years has been very easy to be like, well, who's the winner in the room? It's Nvidia, because everybody was reliant. Now we're getting to the point to where everybody's working on their o...”

ConceptHardware

“He's a well-known professor, published a lot of interesting software design books. And this is the talk he chose to submit. It's on RDMA and inference. And I'm like, you wouldn't think that this guy who is like kind of operating systems guy would care about RDMA.”

ConceptHardware

“Yeah. I don't know if this is a natural follow-on to the topic of like my skepticism on distributed training. Mm-hmm. Is that, well, like people spend a lot of money on like cables to hook up GPUs and even that is not like fast enough and that's the bottleneck. Is your network...”

ConceptHardware

“Yeah, so I guess you're talking about sort of fully distributed training, like Dailoco or something, which is like cross-data center.”

ConceptHardware