Most tech podcasts talk about programming languages secondhand. This list is different: every entry here is a conversation with someone who actually built the language, the compiler, or the framework in question. Pulled from our full library of episode summaries, these are the conversations where the person who wrote C++'s standard, designed the JVM, or shipped LLVM explains the actual decisions, tradeoffs, and arguments behind the tools millions of developers use every day.
Expect origin stories that go back to punch cards and PDP-7s, blunt opinions about rival languages, and a few numbers that will make you rethink what you thought you knew about performance, safety, and code size. Whether you write C++ for a living, argue about Rust on the internet, or just want to understand why Java looks the way it does, there's an episode here for you.
Bjarne Stroustrup: C++ | Lex Fridman Podcast #48
The single best entry point into why C++ works the way it does, straight from the person who designed it. Stroustrup lays out the zero-overhead principle in plain terms: an abstraction must never cost more than the equivalent hand-written low-level code, which is how he claims C++ matrix multiplication can actually beat Fortran. He also picks constructors and destructors as the single most beautiful feature of the language, the reason C++ never needed garbage collection. Listen if you want the philosophy behind the language, not just the syntax.
Read the full episode notesBrian Kernighan: UNIX, C, AWK, AMPL, and Go Programming | Lex Fridman Podcast #109
Kernighan was there when Ken Thompson wrote the first version of UNIX in three weeks on a PDP-7, and this episode is the closest you'll get to sitting in the room at Bell Labs. He explains why C found a sweet spot between expressiveness and efficiency, admits he struggled with Rust's documentation for days, and notes that about a dozen languages now account for roughly 95% of all programming while some 2,000 others sit in marginal use. Essential listening for anyone who wants the history behind the tools they use daily.
Read the full episode notesJames Gosling: Java, JVM, Emacs, and the Early Days of Computing | Lex Fridman Podcast #126
Gosling built Java specifically because pointer bugs caused 50-70% of all security vulnerabilities at the time, and this episode traces that mission from a broke teenager's dumpster-dived relay computer all the way to the JVM. The JVM idea itself grew out of a grad-school hack he once called 'sleazy' that ended up beating the actual C compiler. Good for anyone curious how deliberate safety tradeoffs shape a language's entire design philosophy.
Read the full episode notesChris Lattner: Compilers, LLVM, Swift, TPU, and ML Accelerators | Lex Fridman Podcast #21
Lattner built LLVM as a University of Illinois research project that quietly became the shared compiler infrastructure behind Apple, Google, Nvidia, Intel, and AMD. He explains why Swift exists at all: memory safety became the real justification, since you can't fix Objective-C's pointer problems without changing the language underneath it. He also reveals he started Clang and Swift on nights and weekends without telling anyone at Apple, since a new language was considered borderline heretical there. Best for developers who want to understand what a compiler actually does, phase by phase.
Read the full episode notesDHH: Future of Programming, AI, Ruby on Rails, Productivity & Parenting | Lex Fridman Podcast #474
DHH makes the case for beautiful, dynamically typed code with real numbers behind it: Shopify's Ruby and Rails stack handled about a million dynamic requests per second on Black Friday, and its 5-million-line monolith would have run 25 to 50 million lines in Go or Java. He also details cutting 37signals' infrastructure spend by half to two-thirds after leaving AWS entirely. Listen if you want the strongest defense of programmer happiness over raw performance metrics you'll find anywhere.
Read the full episode notesChris Lattner: Future of Programming and AI | Lex Fridman Podcast #381
Lattner returns to explain Mojo, a Python superset he claims has demonstrated speedups up to 35,000x by moving from CPython's interpreter to a real compiler with optional types and value semantics. He's candid about the scars from Swift's stressful 2014 launch and Python's painful 2-to-3 migration, and how those experiences shaped Mojo's deliberately slow, community-driven rollout. A sharp listen for anyone tracking how AI workloads are reshaping language design.
Read the full episode notesChris Lattner: The Future of Computing and Programming Languages | Lex Fridman Podcast #131
Lattner's third appearance zooms out to the theory: a programming language as 'a bicycle for the mind' that raises abstraction so more fits in your head at once. He contrasts Swift's core-team governance with the burden that fell on Guido van Rossum as Python's sole decision-maker, and explains why Swift dropped the C-style ++ and -- operators for causing more confusion than they were worth. Good for anyone who wants to think about language design as a discipline, not just a feature list.
Read the full episode notesJohn Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets | Lex Fridman Podcast #309
Carmack's marathon conversation covers the technical tricks behind Wolfenstein 3D, Doom, and Quake, including his own QuakeC scripting language, which he says taught a whole generation to program even though he calls it 'nothing to write home about.' He also runs his C/C++ code in a debugger constantly, setting a breakpoint and stepping through every new function he writes, a habit worth stealing regardless of what language you use. Recommended for anyone who wants engineering philosophy from one of the best to ever do it.
Read the full episode notesThePrimeagen: Programming, AI, ADHD, Productivity, Addiction, and God | Lex Fridman Podcast #461
A wide-ranging tour through Go, Rust, Zig, and JavaScript from a Netflix veteran turned streamer, including his sharp observation that benchmark tests claiming Zig is fastest are often misleading since Zig, C, and C++ all compile through LLVM anyway. He also recounts the Falcor 'Repulsive Grizzly' vulnerability he found in production Netflix code, where a single crafted request could crash a machine by spawning billions of JVM objects. Best for developers who want current, opinionated takes on the modern language landscape alongside real production war stories.
Read the full episode notesThat's nine conversations straight from the source, covering nearly every layer of the stack from C to Mojo. Browse our full library of episode summaries for more deep dives into the people and ideas shaping how software gets built.