The Signal Path Discussion - Engineering & RF (3/3)

[@consciousenergies]

These talks were delicious brain candy. Appreciate you guys sharing your knowledge.

[@w0qe]

This is one of your best videos ever Dave.

[@peterpade6293]

What a great chat between two enlightened people - from science; religion; engineering; wave forms and lots more. I also believe a good cat is a flat cat (having been raised on a farm in Oz). Continue the good work you guys. Remember scientists do falter! I am a retired engineer with a faith! Thanks for teaching me and allowing my world to expand.

[@daveblane6442]

Your opinion of cats shows your lack of brains. douche!

[@patwalsh2138]

Too bad it ended, I was really enjoying this

[@andymitchener8697]

Great content!!! I love this so much. Please think about doing these kinds of interviews more in the future, I would watch. One thing though. Please allow your guest to finish their thought before jumping in. You dont need to "dominate" the conversation, I'm sure you are not trying to do that though. It'll come with time and getting more comfortable talking to people in this format. Thanks and keep up the great work.

[@HillsWorkbench]

A conversation between brilliance and genius. What a powerful mind, a stimulating listen!

[@VeraTR909]

The bit about GPU precision was eye opening, never realized hardware could be used that way, fascinating stuff!

[@kemagorar]

If needed here is the reffered Bell Labs episode: www.bell-labs.com/var/articles/yann-lecun-predictive-learning-next-frontier-ai-february-17-2017/

[@alextrofimov7947]

I enjoyed this talk a lot, thank you! Dave is like my father as professional and Shahriar helped me a lot to get into RF stuff. And they are the two ones who I support on patreon.

[@Deckardsvr]

Fascinating conversation, I'm glad you two finally met.

[@neuralnetwork653]

Thank you, guys! It is always pleasure to listen to a clever people. EEV and SP are my favorite youtube channels :)

[@DiogoCRibeiro]

Really nice video!! I really enjoyed Shahriar's comments about the technologies to produce integrated circuits and the advantages/disadvantages of those to build high-frequency ICs. Really good stuff!

[@russ18uk]

I could listen to Shahriar talk about this for an hour on his own channel. The talk about substrates squashes a lot of BS.

[@maxtur6982]

Oh yes what professional monsters couple !! what nice suprise to view you both there!!

[@MrSparker95]

Very interesting interview. Thank you very much, I enjoyed listening to both of you a lot!

[@markuscwatson]

Two of my favourite channels

[@PapasDino]

Thanks to both of you for a very interesting discussion!

[@maurosobreira8695]

Where do I find the 4/3???

[@PaulHigginbothamSr]

To me an old man, I consider Dave as an engineer. I consider Shariar as an electrical philosophist. So beyond engineering and an actual teacher of component dynamics.

[@capt.k8577]

i really enjoyed this vlog, little bit here and there . exciting new stuff , leaps in technology compared to 10yrs ago. thanks for the video dave

[@SteveCooling60]

Just finished all 3 videos of this conversation. Very interesting!

[@astrodevelopments]

I don't know if anyone remembers but RCA were using SOS (Silicon on Sapphire) back in the late 70;s for memory chips, I think I have still got some of the data sheets somewhere, curious how technologies turn around.

[@jaa93997]

Astrodevelopments maybe back then there was no technology capable of processing this material, and it got shit canned for that reason?

[@Landrew0]

Great job interviewing, not saying "yup" all the time.

[@funkyironman69]

GaAs is also often used for radio astronomy amplifiers

[@stonail665]

Very interesting topic. Thank you dada Shahriar

[@quadpumped34]

you guys are awesome.

[@jtveg]

Very interesting talk. Thanks. 😎👍

[@davidlondon3111]

Sometime I wish there was a way to give 2 thumbs up. Thanks for the great discussion

[@bennguyen1313]

The audio version of this interview (link on theamphour) seems much longer. Anybody have the links for the youtube channel mentioned at the 1h8m5s mark of the audio version? Also, which is the lecture mentioned at the end of this video (16m24s) with Claude Shannon / Yann LeCun on A.I.?

[@DrakkarCalethiel]

already one dislike? There's always one who just clicks it immediately after the video is live...

[@EEVblog]

Yes, serial haters thumbs down every video regardless.

[@DrakkarCalethiel]

EEVblog I think that's what every big channel have to deal with. good that small channels are not targeted by some haters.

[@EEVblog]

Yes it's different when you are a small channel. The bigger you get the more haters you get, it is 100% guaranteed.

[@DrakkarCalethiel]

EEVblog It happens to everyone. have to say that I really appreciate your work! I learned tons from you and I absolutely love how you communicate with your subscribers.

[@ChipGuy]

There are dislike bots and even Outlook VBA scripts for the email notification. I won't link them here of course. Once a channel got a certain size there will always be haters, and they are pretty "loyal".

[@wizz33a18]

i like this one the most

[@ats89117]

For Intel, the cost of the increasingly sophisticated fabs was also increasing exponentially and that was really a problem for Moore's Law. Maybe if money wasn't a limitation they could have kept going for a few more years?

[@funkyironman69]

No new AmpHour episodes for awhile, starting to get withdrawal symptoms. :p

[@palantinos]

when one wise guy talks to another wise guy you just grab your beer and listen

[@robertboll1503]

Wow! Keep it up!

[@henrikthomsen3518]

100% Tech love to you guys :-*

[@benrd9111]

very interesting, nice 👍

[@juanmarceloarguello8929]

what say in 7:05 silicon ... ?

[@zarf4219]

Silicon on Sapphire

[@Petex90]

Thumbs up! You need to make more collab videos if both happen to be at the same continent some time in the future

[@stuartthegrant]

Wow what an interesting guy.

[@phillipsusi1791]

I don't know how anyone can claim with a straight face that moore's law did not die a few years ago. I recently upgraded my old sandybridge system after 5.5 years, and yea, the cache speed has doubled, and the ram speed has almost doubled, but the clock speed of the cpu is unchanged, as are the number of cores. And that's after 5.5 years, not 1.5.

[@wither8]

To be pedantic, Moore's law refers to the feature size on the physical silicon used at the fab. The density is said to double every 18 months on a square die. It has nothing to do with the number of cycles you can push (which, itself, is a poor number to determine performance anyways, since there is no real correlation between a cycle and, say, an instruction being processed -- which itself will vary, floating point division will take literally ~100x as a logical left shift from something in rax.) This doesn't even factor in modern architectural givens such as branch prediction, multi-level caches, tons of heuristics for pipelining, multiple ALUs, etc.

[@phillipsusi1791]

@wither8, yes, and that was mentioned in the video. My point is that even using the relaxed definition of performance doubling every 18 months, that has not been true for at least 5 years now. Also floating point isn't *that* slow.. even if doing it with a software math library I don't think it is *100*x slower than integer math.

[@DAVIDGREGORYKERR]

Maybe it is time to replace silicon and germanium with Diamond and doped Diamond as Diamond can take the higher temperatures.

[@DAVIDGREGORYKERR]

vaporising I mean splutter diamond on to a ceramic substrate in a vacuum chamber

[@victornpb]

I've read somewhere in the past about very high frequencies being hazard to humans, like above 60GHz I think, not sure about the truth about it. It said something in the lines of being resonant to dna or other tissues... of course it would depend on the power levels, but can anyone confirm that?

[@jeffmorrison579]

I have a Avalan 900 mhz Ethernet AW900F any one know a good site to sell it on besides Ebay ?

[@smartchip]

could anyone help, I would like to buy a thermos meter, in about a year / some time in 2018, what would you recommend, fluke do em too, are they as good, I know flir make some of em or something, from a basic, intermediate and a pro (depends upon cost) recommendations, also some advise, even basic, some people use em a lot, therefore know a shed load, its for electrical testing, electrician, I'm a electrician now, engineering made me unhealthy, as I love to eat, being a electrician I am always physical, anywho, Engineers n hobbyists, I guess its in our Dna to want to know how things work & fix them,

[@freddykabulaschnitza2475]

Hi Dave, have you had Covid 19 yet?

[@bayareapianist]

Dave. Did you give a free T-shirt to Shahriyar? 😀

[@JGunlimited]

Wow!

[@danielmcnerthney]

Great content. Where's the link to the Bell Labs neural network video kzbin.info/www/bejne/qpvHkphupNaMg7M ?

[@shakaibsafvi97]

A hat trick :)

[@EEVblog]

Bowled!

[@Nejdat]

merhaba dostlar

[@robertw1871]

While the current research in AI is not actually intelligent, in my understanding it’s basically just filtering. The quest for true machine intelligence is hugely misguided in my view. I’m not entirely convinced we will even know when it becomes self aware and this it truly frightening as we have no basis to understand what its goals will be. Think carefully, we have evolved with many intelligent animals and we have yet to understand the communication of any of them. Can it be imagined what a machine with say, conservatively, 10 times the power of a single human brain would be capable of? Ten times the intellectual power of an Einstein and no way to understand what it will try and achieve. Being a machine, it will follow exponential growth and quickly grow well beyond even this. And we have not yet reached the level where we can communicate with an ordinary dolphin. When this happens it will be extreme; it will be either star track utopia overnight or Armageddon, but most probably extreme in a way we aren’t capable of imagining.