Chris Lattner: The Future of Computing and Programming Languages [audio]
lexfridman.com
lexfridman.com
I listen because I enjoy the guests he's able to land and what seems to be an authentic interest in how they think and approach the world. It would be great if there were more podcasters having these conversations, but as far as I can tell there aren't. So the question is if the world's better off with or without these interviews as-is. To me the answer is clearly the former.
Yes, amazing guest list, I do wonder how it took off like that.
Edit: in contrast, Rob Reid's After On podcast has somewhat similar guests (not an all-stars list, no AI folks except Stuart Russell & Rodney Brooks), but he's my favourite interviewer. The amount of research he does before each interview is insane, to the point he already knows what the guest will say and what are the most interesting items to talk about. But he still lets the guest tell the story themselves.
Some of his guests, however, are a very bad choice. In particular I am thinking of Eric Weinstein who has been on TWICE. Weinstein is a master-class bullshtter, who opines on his theory of everything yet has NOTHING to show for it. He also manages to pull in politics, academia, culture, and the kitchen sink, all obfuscated by turgid rhetoric.
As the commenter above said, lots of 'meaning of life' questions that just aren't that interesting.
Doctors and lawyers are about the only occupations where a professional’s mistake could cost a client his/her life. No other professions come close, which is part of why the training is long and the practitioners require a license to practice.
This person purports to have "a theory of everything" up his sleeve yet won't communicate it properly because he doesn't trust the physics community. That doesn't stop him from lengthy podcasts and interviews, like with Lex, where he uses a firehose of specialist jargon against a general audience. To what end? I can't imagine.
The obscurantist IDW stuff appears to be nothing more than a vaguely libertarian stance mildly against progressives and mostly in academia-- boring and not even close to "dissident".
If the guests would feel in the same way, then they wouldn’t be talking with Lex.
So it's most likely an issue you have. I personally don't mind Lex at all and really like his style, especially his "simple" questions, because how great minds answer these "simple" questions can be quite interesting.
I am often reminded of this anecdote:
"Von Neumann would carry on a conversation with my 3-year-old son, and the two of them would talk as equals, and I sometimes wondered if he used the same principle when he talked to the rest of us." (Teller)
> how great minds answer these "simple" questions can be quite interesting
Yes for sure. But rather than directly ask the naive question which we simple listeners want to ask, I want a smart interviewer to give us some quick signposts about schools of thought on this and where the interviewee probably comes from given what they've said in the past. Then you can get a deeper answer. Even, perhaps especially, when they disagree with the way in which the interviewer has framed the question.
If you don't like Lex's style, well, then don't listen to him. A lot of people seem to like and enjoy him. Asking someone to change to better fit your own expectations, honestly, seems quite a bit to ask, and at the end might not work that well, if it doesn't fit himself.
Sure, you can critique whatever you want, but Lex doesn't really has to listen and also shouldn't, if he has that many listeners by just doing his thing.
Maybe I'm hoping someone will tell me who I should listen to, instead.
So, what I‘m trying to say with this? I don‘t quite know, just a bit of a brain dump. Perhaps you‘re getting some hints for other podcasts.
I like Lex's style - the guests always deliver on interesting conversations, which is the point of the podcast. Lex is playing a role that I think facilitates this.
The endless "I am so very humble" spiel also gets tiresome, especially when punctuated by claims like "all journalists are trash (except those with a unique scientific curiosity, like me, of course)".
What I'd like is more like what Tyler Cowen does, on non-tech subjects [1]. He gives the impression of having read all their books (and their opponents') not just having skimmed Wikipedia over breakfast. He keeps it moving, and often gets people a little bit off-balance, not aggressively, but enough to make people think afresh and not just repeat what they always say.
Besides people's work, he'll ask them serious questions about (say) their known interest in some kinds or Jazz, or unrelated things about characters they would have run into at MIT in the 70s, or about the Russian education system. (I presume he has people helping him prepare, but maybe he's just like this all the time.)
I haven’t heard a better interviewer since Charlie Rose (who had a team working for him, of course).
Does anyone know of similar interviewers in this vein? Long form and exceptionally-prepared?
Really? Rose is a serial interrupter. Might as well have Gary Vee interview people if you think Charlie Rose is any good.
However, I'll defend Rose, in that he always picked interesting guests, and did good prep work. It's true that he steered the conversation in the direction he wanted it to go. But, of course, he had 60 minutes, and was recording for a broad television audience. The constraints were/are very different from Tyler's, personality differences notwithstanding.
If you listen to the interviews where he gets to break out of the normal 60 minute boundaries, I think you'll find his style shifts with the medium. https://www.youtube.com/watch?v=LOUzTd-Ge_g
I've never heard of Gary Vee. However, based on a cursory search, I don't think a hyperbolic comparison helps your argument! :)
Agree that Cowen himself wouldn't be the right guy, at least not if you want to hear the technical side.
I think it's a good example of 'you can't please everybody'. For my part I hope he stays the course.
I have mostly worked in the field of AI (practical projects, not much research) since 1982. I feel strongly that AI systems that work with us, augmenting what makes us human and unique, is great. AI systems that aim to forcing human behavior into tight and strict protocols are simply bad. AI helping exploitive surveillance capitalism systems is bad. The human and basic humanity comes first.
Donald Knuth handled it well.
Do we live in a simulation?
Hanging out debating unanswerable questions all the time has zero value.
https://www.thedrive.com/tech/27359/mit-study-showing-high-f...
Not sure what this fixation on mars is about, tough.
They are at a predictable spot in the podcast (the end), so it's pretty easy to just stop listening when he brings them up.
But perhaps this creates the opportunity for someone to make a series of supercuts, where Lex doesn't need to appear, but we get the appearance of a series of luminaries discussing a recurring question, and the differences and similarities among responses can be more apparent. A virtual panel, smeared across time.
https://www.youtube.com/watch?v=QZP8OwCdO1E
https://www.youtube.com/watch?v=5RKbSi87J5Q
I agree with the previous comments, Lex' guests are really great, maybe the above can serve as inspiration :)
https://drexel.edu/engineering/news-events/news/archive/2019...
Lex, since you disagree with Joe Rogan's advice on not reading the comments, I know there's a slight possibility that you read this. I want to thank you for the fulfilling content, congratulate you on your success as a podcast host, and tell you that you are an inspirational person.
Being in MIT probably gave him easy access to the interesting people that started off his podcast, but beyond that i think his affable , empathetic and easy style is rather attractive, so people don't say no to him, and many have returned back.
"yes it is different and it can feel like stoned teenager with disorganized thoughts asking questions that form mid sentence"
And after getting the ball rolling its not hard to get new guests by listing your previous guests as well as your view counts.
Not denigrating people in public...
Many great things about Lex. One I like most is his quickness to ask basic questions. Chris Lattner mentions it as a great trait of his own -- sincerely trying to learn, rather than impress you
So well done Lex
If you don't like his show or style, then you don't have to listen to it. :-)
-Chris
No Haskell, Erlang, Julia, Go ...
A bit disappointing regarding the title of the podcast, don't you think so?
When I listened, I was wondering what is a rather obscure language you like and enjoy and why. (I still like hacking OpenDylan from time to time.
For me, Swift seems great for user interfaces and for some os applications (command line etc.). Especially Apple's SwiftUI seem to take the best of d3.js and other libraries. Swift seems perfect and very flexible for that.
Machine Learning and Signal Processing will still be dominated by Python, just because of the libraries. I'm old school and program a lot of matlab ...
Thought I would get more background on your influences and your programming culture.
Personally I'm almost a layman when it comes to hardware and compiler technologies. When Lex asks basic questions, they are the sort of things I want to hear about.
Sure there's room for a more in-depth/challenging conversation but I don't think that Lex's podcast is the place for that. Theres probably a handful of people who would be able to challenge Chris properly (Maybe a conversation with another LLVM contributor or someone on the Swift team would be interesting).
Loved the podcast Chris, thanks for giving your time and hope you keep doing stuff like this!
It's a really questionable title and in my opinion bad framing.
For what it's worth, I enjoyed the interview! I wouldn't say I learned anything new, but I don't need that out of everything I listen to.
For my part I really enjoyed both of your visits to the podcast. The first one really contextualized a lot of the work that I wholly take for granted every day and this just expanded on that. Lex should do an annual State of the Stack series and drag you back in for more :)
He had the opportunity to dive in deeper on some topics - but didn't really do so. It just could have been better.
My critic has always been that Lex is superficial, which is understandable, because he is targeting casual audiences.
Like some other comments suggest, he is squandering the opportunity in a certain degree. All his question can be answered by someone who is not Chris Lattner, and the content would not be too different from what's we see right now.
I personally place high regards on Chris Lattner's technical prowess in the compiler and machine-program interfacing space. That's something that he can deliver content that no other I am aware of can do better.
Yet, all I have been seeing is something that have been repeated everywhere else, and I am too tired of this blandness.
On the contrary, take a look https://www.youtube.com/channel/UCZHmQk67mSJgfCCTn7xBfew, the producer does not appeal to casual viewers. It's often a deep disect of technical papers inside out. That's what I am looking for.
What I value in Lex's content is that it is not complete pop-science level, that he is very neutral and especially that he leads his guest to make opinionated statements. The last point is usually the single value nugget I can extract from the episodes. It is basically a shortcut for me watching Yannic Kilcher like videos and getting deep into a topic through the proxy of a person who is not rarely one of the best in his field. I do not have the time / energy to get deeply into compilers and programming language design I just want to get a feeling in broad strokes where the field is heading, what are the key developments and bottlenecks that have shaped the past and especially future of certain technologies and applications.
I will probably dig a (tiny) bit deeper into MLIR after listening to the podcast because it seems to me to be the 80/20 kind of way to get a better feeling for the developments happening in compilers and how ML workloads are mapped to accelerators.
As for Lex, it's fine. I am just saying, like others, he is underutilizing his speaker...
Also in general, long form podcast conversations are such great entertainment. I've learned a ton from his podcast conversations. Excited to spend the morning listening to this one. Lattner seems like an extremely fascinating guy
Also has timestamps for specific topics.
Value types (with copy on write), progressive disclosure, ...
I think Swift is a great language but the lack of cross platform development is holding it back.
His example is a Hello World program in Swift vs something like C++.
He speaking style can throw people off. I'm pretty sure I have heard this style from other eastern Euro and Russian people. It doesn't bother me, but I can see why it can be grating to some people.
I get not liking his or any other podcast. I don't get the need to belittle the dude or enter into some weird conspirational thinking about his motivations.
But now I watch his podcast all the time, and I've gotten a lot out of it.
I agree with others in that he doesn't try to make himself the focus of the conversation, and that's a good quality in an interviewer.
He makes his subjects comfortable enough to talk freely, to say things I haven't heard elsewhere over 1-2 hours, and that is an incredible and valuable skill.
Lex, if you're reading this by any chance, please keep up the great work. There are a lot more people who love what you're doing than those who don't! :)
My knee jerk reaction to Lex was to dislike him, but if you allow yourself to dig a little deeper I think you'll find he's a hard working, intelligent and humble guy that's following his passion.
Why all these comments about Lex and his interviewing style and almost nothing about the technological aspects, like MLIR, LLVM or RISC-V?
The content is fairly shallow.
I am looking for content that actually can describe practitioner problems, not just superficially on general conceptual problems.
One that delivered by John Carmark would be one such example. But John probably is on another level than Chris here, and is very rare.