3 ms·

I think that one of the things that is very interesting about this interview (and is reflected in Intel's share price at the moment), is that Intel will be a le

I think that one of the things that is very interesting about this interview (and is reflected in Intel's share price at the moment), is that Intel will be a less profitable company. Back in the early 2010s Intel had market leading products and was actually trying to bring down R&D costs. With Pat's plan it's "Invest in 10 things, 1 will pay off". So you're going to be paying for all of those 10 things. The costs are going to be up dramatically compared to old Intel even if the plan works.

thevagrant 3y ago
It's true, although we don't know what will happen. Intel may discover the next Core Duo like design that leaps them ahead again. TSMC could hit fab problems on future nodes (like Intel has in recent times). Not to mention, a conflict may render TSMC unable to supply, leading to global reliance on Intel to bridge the gap. Intel has squandered a large number of opportunities, but I wouldnt write them off. I think Pat is on the right track. If Gelsinger navigates recovery, will Intel learn from this?
washadjeffmad 3y ago
I agree, and I think Intel has had a hard time reconciling their potential against reality. It feels like they're gun shy after the continuous AMD and Apple Silicon advances, afraid to commit in case someone else is on the cusp of a breakthrough that invalidates entire product lines. While they were quick to put out a pytorch extension, they weren't prepared with OneAPI, OpenVINO, or hardware lineups to take advantage of the AGI whirlwind. Ironic, since so many researchers learned using Intel resources. Even ROCm is seeing more growth and use. I'm a little embarrassed, but I built a system around 13th gen for GAN and AGI work using the above Intel technologies last year, then immediately went out and bought an RTX 3090 and moved back to Threadripper. I just spent $90k building a research lab for AGI around Nvidia technologies instead of Intel. I don't particularly love their business history, but this feels like such a missed opportunity for them.

thevagrant 3y ago

It's true, although we don't know what will happen. Intel may discover the next Core Duo like design that leaps them ahead again. TSMC could hit fab problems on future nodes (like Intel has in recent times). Not to mention, a conflict may render TSMC unable to supply, leading to global reliance on Intel to bridge the gap. Intel has squandered a large number of opportunities, but I wouldnt write them off. I think Pat is on the right track. If Gelsinger navigates recovery, will Intel learn from this?

washadjeffmad 3y ago

I agree, and I think Intel has had a hard time reconciling their potential against reality. It feels like they're gun shy after the continuous AMD and Apple Silicon advances, afraid to commit in case someone else is on the cusp of a breakthrough that invalidates entire product lines. While they were quick to put out a pytorch extension, they weren't prepared with OneAPI, OpenVINO, or hardware lineups to take advantage of the AGI whirlwind. Ironic, since so many researchers learned using Intel resources. Even ROCm is seeing more growth and use. I'm a little embarrassed, but I built a system around 13th gen for GAN and AGI work using the above Intel technologies last year, then immediately went out and bought an RTX 3090 and moved back to Threadripper. I just spent $90k building a research lab for AGI around Nvidia technologies instead of Intel. I don't particularly love their business history, but this feels like such a missed opportunity for them.