Designing in 2023: 10 Problems to Solve w/ Jim Keller

[@TwoMinutePapers]

Thank you so much for the shoutout at 19:00. What an honor, it really made my day! 🙏

[@samaBR333]

what a time to be alive!

[@Sesarrbg]

Hold on to your papers!

[@atillacodesstuff1223]

congratz :)

[@marshallmcluhan33]

hold on to your papers!

[@WildEngineering]

he also did it in his video posted 3 months ago :) youre awesome!

[@TechTechPotato]

Ahhh yiss. Been waiting for this.

[@TMS-EE]

Me too, Agree with Jim on many things but prefer the Royal Society definition of an engineer from when they were asking what the difference is between a scientists and an engineer. They said engineers "solve problems and build things". If the AI is going to be doing half the design we need to think about the problems and what we want to build. This is a great reference to show pace of progress in semiconductors and SW.

[@SBA_poiko]

Can never get enough of Jim Keller tbh

[@mayikx]

🥔🥔🥔

[@winsomehax]

Jim Keller is a great speaker AND interviewee.

[@BadccVoid]

Jim Keller is a legend. Phenomenal talk.

[@chadleo85]

Have been telling a bunch of people that open source is the next big inflection point, but most of the traditionalists dont want to admit it or put the time into it. Thanks Jim for confirming my thought process!!

[@smallbluemachine]

AI quality for various artistic content can be incoherent, but for code generation it’s absolutely astounding what it can generate, we really are at a turning point how we work.

[@TobiasFrei]

What a beautiful way to present these topics. Apart from instruction sets I know very little about the field. And yet I was able to follow 🤓

[@jayhu6075]

The open source community become important in the future to solve problems in the hardware on different AI devices. So that not only Big Tech can us it, but also the public and small companies.

Thanks for uploading it. And good luck guys.

[@KabelkowyJoe]

9:00 and here it comes - if AMD or Intel didn't noticed the obvious Tenstorrent ad, I don't know what else to say Tenso2rent

[@randfee]

great talk @Jim - but on a sidenote, 3nm is more like 6 silicon atoms wide with roughly 5A lattice distance, so I guess this is a technicality ;-) regards, a typically not so nitpicking physicist (but also engineer).

[@tobiassteindl2308]

The name of the manufacturing process does not represent the size of the transistor, i.e. on a 3nm process, the transistor is (much) larger than 3x3x3 nm

[@fawal.1997]

How does higher energy light causes more problems for photolithography?

[@TracFone-xn7fj]

I think it's a reference to the general problem where working with high energy things is always harder than working with low energy versions of the same thing. It's why we try to use the lowest energy possible to do anything. UV photons damage DNA in ways that are still dangerous despite evolution having billions of years to defend itself. In the grand scheme of things you're always better off without ionizing radiation randomizing your system at the most inconvenient time.

[@chewu97]

You should consider the problem of fresh graduates having real difficulty navigating the sea of directions digital design and electronic engineering are going rn. These talks are englightening to say the least, information is too compressed though.

[@smallbluemachine]

There is already too many 1.5 to 3 hour talks already in my opinion. A 21 minute shotgun presentation by a technical legend like this is like fresh air to me.

[@graham8316]

Jim rules ❤

[@g1m0kolis]

Nice talk, but the video is very pixely, 1080p is just upscaled lowres video?

[@dkutagulla]

Problems are why we have jobs - Jim Keller So True

[@TracFone-xn7fj]

It's crazy how often I find myself in violent agreement with Mr. Keller. I feel like a software version of him but obviously not as clever or experienced. I'd love to have a chat with him one day about how much of what he does is intuitive and how much had to be learned the hard way, and who he considers to be the best engineers he's worked with. I suspect he had a really good mentor early on. There's nothing a good engineer loves more than a clearly stated, real world relevant problem that has a proof from nature that we just haven't figured out how to do with our machines. Any such problem is by definition interesting enough to make a life's work

[@retf054ewte3]

this video is good only for engineering students.

[@quaidcarlobulloch9300]

excellent!

[@tdudzik]

Who am I to disagree with Jim Keller but I don't think that DNA is a factory, I think DNA is just a blueprint and whole universe is a factory. We could also say that DNA is a machine code and a universe is a computer.

[@oblivion_2852]

Most of the code of dna is the code to reproduce dna. Basically dna is both the data and the factory.

[@Martinit0]

The Ribosomes are the factories that produce proteins from mRNA, which in turn is transcribed from the DNA.

[@brandonheaton6197]

Jim Keller watches two minute papers too? I feel just a little smarter

[@commercecloudcommunity]

Hey Jim: Much Respect. Matter may-be high density light... Have you checked out the Inverse-square Law?

[@digitalcontent1870]

Tenstorrent channel have only 1.86k subs my opinion it should have 100k.🤔

[@RolanElizabeth]

That's the funny thing,the fundamental field of modern day entertainment and basic living is literally based off this field,I Guess it's a gift,too much fame and attention isn't good,do your work, take your reward (Money, leverage,name) etc

[@mattholwood]

The influence of his brother in law on his communication is pretty obvious. In a very good way. Smart guy, top 1% of the top 1%, putting things in a way I (and idiot) can understand.

[@dranelemakol]

What does IP mean here?

[@SergeyPapin]

An Intellectual Property (IP) core in Semiconductors is a reusable unit of logic or functionality or a cell or a layout design that is normally developed with the idea of licencing to multiple vendor for using as building blocks in different chip designs.

[@atillacodesstuff1223]

@@SergeyPapin wow that's interesting

[@graham8316]

Smaller atoms would help fucking rules

[@noanyobiseniss7462]

Digital alpha, nuff said.

[@kupassyatina]

I don't understand anything

[@kieran5302]

Noob question here: when hes referring to IP.. it doesnt sound like he's referring to intellectual property, is he?

[@TheRadischen]

"smaller atoms would help" lmao

[@PeterParker-ot8pl]

Never been first before. Weird...

[@autoclearanceuk7191]

If open source is so great, then why has it not killed Microsoft ?

[@rlrevell]

It has killed Microsoft in every fiels where high barriers to innovate are absent. All that remains are things that demand their inner workings be kept secret, and that does not sustain as reverse engineers are always active.