Saša Jurić is fantastic at condensing lots of information in a 1 hour talk without losing the audience, he gave another great talk this year called Parsing from first principles (https://www.youtube.com/watch?v=xNzoerDljjo).
I purchased his book, Crashed: How a Decade of Financial Crises Changed the World without really knowing much about it (or him) and found it to be a great, semi-technical story about the financial crisis (and its politics) from the global perspective.
Not a talk from 2019, but discovered it this year: "1177 BC, the year civilization collapsed (Eric Cline, PhD)" https://youtu.be/bRcu-ysocX4
Excellent informative and hilarious talk about his (at-the-time) new scientific hypothesis to explain the end of the Bronze Age ca. 1200 BC.
Multiple civilizations collapsed within a few decades of each other with the ability to read, write and make high buildings being lost all across the Eastern Mediterranean simultaneously. The Bronze Age is magical and interesting of itself, the talk gives a great introduction as to why we know much more about it than we think.. definitely recommended.
Same here. I started watching because I misinterpreted the title and stayed for the wow effect :)
Also, after seeing what kind of magic he can do with the right representation, I wondered how many "business domain" models I know could be expressed with models that "click" in the same way (i.e. are expressed by simple concepts and compose as well).
I feel as if many of us often give up way too early in the search for good models for our data. Myself included, of course.
“Why isn’t functional programming the norm?” by Richard Feldman. Spoiler: not on the basis of merits. https://youtu.be/QyJZzq0v7Z4
“React to the future” by Jordan Walke. Why ReasonML is a logical extension of ReactJS’ programming paradigm. https://youtu.be/5fG_lyNuEAw
“Typing the untyped: soundness in gradual type systems” by Ben Weissmann. The trade offs that various gradual type systems make based on their language constraints. https://youtu.be/uJHD2xyv7xo
“Let’s program like it’s 1999” by Lee Byron. How the mutual feedback loop of abstraction, syntax and mental model drives the evolution of web technologies. https://youtu.be/vG8WpLr6y_U
This talk is very good. It's one of the few talks that I've overheard classmates talk about. It not only asks a question a lot of people exposed to functional programming at university asks, but also answers it in a way where you learn more about the world of programming and programming languages than you expected.
Maybe I'm missing something but I'm more than halfway through the "Why isn't functional programming the norm?" and it just seems to be a kind of haphazard recollection of programming language history. A lot of which isn't what I'd call entirely correct. Python's killer app was arguably first CGI scripts then data science. Java succeeded due to offering GC in a non scripting language, the JVM and possibly lots of marketing. PHP is having a mild renaissance with Laravel (not that I'd advocate for PHP, but people do seem to love Laravel).
There was quite a bit of time in between the invention of implementation inheritance and the whole "prefer composition to inheritance". It's quite possible OOP became popular due to implementation inheritance then realized it was dumb.
This info is still useful, but what I'd really love from a talk with that title is an analysis of functional programming languages and how they each missed the boat through either syntax, lack of tooling, or purity. And compare it to functional-ish languages like Rust, JavaScript, Swift and Kotlin. Then chart a way forward for function programming language adoption. Maybe that happens at the end of the talk.
A complete digress, but OOPS still shines in the domain of GUI widgets programming where there are a limited number of interfaces and a huge number of widgets (implementations) working with that interface. FP works conversely, on a limited data and a huge set of functions. Maybe in the context of now with limited gui programming, FL is more suitable?
> seems to be a kind of haphazard recollection of programming language history
Agree. The talk is very thin on the real differences between OOP and functional languages.
This old comment [0] points out that functional languages tend to make it far harder to reason about low-level details, for instance.
Personally I think it's more fundamental, and isn't about any such technical limitations. People have a strong intuition for time, which is emphasised in imperative languages (including OOP), which have the semicolon operator or an implicit equivalent. The concepts at play in the fundamentals of Haskell are simply harder, and 'more mathematical', than the sequenced mutation-based statements of imperative/OOP languages.
To put that more provocatively: does anyone doubt that the average Haskell programmer is smarter than the average JavaScript programmer? I'm not convinced this is just because only the curious bother to learn Haskell.
Here's the secret agent ad by Sun which Feldman mentions [1]. These Sun ads are really quite entertaining and the production values have had to been through the roof at the time [2][3][4].
Shameless plug since I'm the speaker. The reason I'm posting this in "best talks of 2019" is not because I think it was a good talk (my ego isn't that big yet) but because I think very few talks exist on the subject of Behavioral Programming, and it's a subject I'm hoping can get more attention.
Why I thought it was cool: Gives a good overview of Async implementations without getting too deep in the weeds. Talks about "greenthreads" in other languages vs a zero-cost abstraction in Rust.
Caltech Guest Lecture: Quantum supremacy using a programmable superconducting processor - Professor John Martinis, Chief Scientist Google Quantum Computing Division
A bunch of the best talks of the year may still be to come when 36C3 takes place from the 27th to 30th. The schedule is already up [1]. Speakers include David Graeber, Edward Snowden, Daniel J. Bernstein, Moxie Marlinspike, and over a hundred others.
All talks will be livestreamed (and usually become available for download one day later) at [2].
I just watched [2]. Can thoroughly recommend, very watchable presenter & I feel like although I had no idea what he was doing half the time it was very cool. Technological Magic!
Brendan presents news tools and new flows that can be used to analyse performance bottlenecks.
He get you in the demo and you stay to receive more and more information. It seems to be a talk that he has been improving over the years. Totally worth your time seen it.
87 comments
[ 3.0 ms ] story [ 184 ms ] threadhttps://developer.chrome.com/devsummit/sessions/the-main-thr...
1. The Church-Turing Thesis and Physics 2. The Limits of Efficient Computation 3. The Quest for Quantum Computational Supremacy
https://www.scottaaronson.com/blog/?p=4301
video: https://www.youtube.com/watch?v=JvBT4XBdoUE
HN discussion: https://news.ycombinator.com/item?id=20942767
Preventing the Collapse of Civilization, by Jonathan Blow
video: https://www.youtube.com/watch?v=pW-SOdj4Kkk
HN discussion: https://news.ycombinator.com/item?id=19945452
Saša Jurić is fantastic at condensing lots of information in a 1 hour talk without losing the audience, he gave another great talk this year called Parsing from first principles (https://www.youtube.com/watch?v=xNzoerDljjo).
https://youtu.be/KGuaoARJYU0
I purchased his book, Crashed: How a Decade of Financial Crises Changed the World without really knowing much about it (or him) and found it to be a great, semi-technical story about the financial crisis (and its politics) from the global perspective.
I don't know if Bryan Cantrill has done any speeches this year, whenever i see some speech featuring him on youtube I watch it regardless of the year.
That man is a gold mine. And I always learn something interesting.
</fangirl>
new startup with Brian, jess frazzle and Steve tuck.
I've already heard some fantastic stories from their guests, particularly the guy who used to work at intel.
[0] https://www.youtube.com/watch?v=0wtvQZijPzg [1] https://www.youtube.com/watch?v=TM9h89Vo_Qo
Excellent informative and hilarious talk about his (at-the-time) new scientific hypothesis to explain the end of the Bronze Age ca. 1200 BC.
Multiple civilizations collapsed within a few decades of each other with the ability to read, write and make high buildings being lost all across the Eastern Mediterranean simultaneously. The Bronze Age is magical and interesting of itself, the talk gives a great introduction as to why we know much more about it than we think.. definitely recommended.
Duration ~1 hour (feels like 20 minutes)
https://www.amazon.com/1177-B-C-Civilization-Collapsed-Turni...
https://youtu.be/jyNqHsN3pEc
Composing music functionally, aka functional composition :)
Also, after seeing what kind of magic he can do with the right representation, I wondered how many "business domain" models I know could be expressed with models that "click" in the same way (i.e. are expressed by simple concepts and compose as well).
I feel as if many of us often give up way too early in the search for good models for our data. Myself included, of course.
https://youtu.be/pjin3nv8jAo
Because, sometimes you need to worry less about the gritty details of the tech and take a step back and look at the bigger picture.
Talk about Svelte v3 and the (possible) future of frontend frameworks
Video: https://www.youtube.com/watch?v=AdNJ3fydeao
HN discussion regarding Svelte 3: https://news.ycombinator.com/item?id=19719118
“React to the future” by Jordan Walke. Why ReasonML is a logical extension of ReactJS’ programming paradigm. https://youtu.be/5fG_lyNuEAw
“Typing the untyped: soundness in gradual type systems” by Ben Weissmann. The trade offs that various gradual type systems make based on their language constraints. https://youtu.be/uJHD2xyv7xo
“Let’s program like it’s 1999” by Lee Byron. How the mutual feedback loop of abstraction, syntax and mental model drives the evolution of web technologies. https://youtu.be/vG8WpLr6y_U
This talk is very good. It's one of the few talks that I've overheard classmates talk about. It not only asks a question a lot of people exposed to functional programming at university asks, but also answers it in a way where you learn more about the world of programming and programming languages than you expected.
There was quite a bit of time in between the invention of implementation inheritance and the whole "prefer composition to inheritance". It's quite possible OOP became popular due to implementation inheritance then realized it was dumb.
This info is still useful, but what I'd really love from a talk with that title is an analysis of functional programming languages and how they each missed the boat through either syntax, lack of tooling, or purity. And compare it to functional-ish languages like Rust, JavaScript, Swift and Kotlin. Then chart a way forward for function programming language adoption. Maybe that happens at the end of the talk.
Agree. The talk is very thin on the real differences between OOP and functional languages.
This old comment [0] points out that functional languages tend to make it far harder to reason about low-level details, for instance.
Personally I think it's more fundamental, and isn't about any such technical limitations. People have a strong intuition for time, which is emphasised in imperative languages (including OOP), which have the semicolon operator or an implicit equivalent. The concepts at play in the fundamentals of Haskell are simply harder, and 'more mathematical', than the sequenced mutation-based statements of imperative/OOP languages.
To put that more provocatively: does anyone doubt that the average Haskell programmer is smarter than the average JavaScript programmer? I'm not convinced this is just because only the curious bother to learn Haskell.
[0] https://news.ycombinator.com/item?id=21281004
[1]: https://www.youtube.com/watch?v=NVuTBL09Dn4
[2]: https://www.youtube.com/watch?v=cfwMMI7hqns
[3]: https://www.youtube.com/watch?v=njnNVV5QNaA
[4]: https://www.youtube.com/watch?v=AP4FgXOlMh0
https://www.youtube.com/watch?v=0LNQxT9LvM0
I learned so much for this talk. I had a much different idea of the point of iterating rapidly and what product-market fit meant before this video.
Shameless plug since I'm the speaker. The reason I'm posting this in "best talks of 2019" is not because I think it was a good talk (my ego isn't that big yet) but because I think very few talks exist on the subject of Behavioral Programming, and it's a subject I'm hoping can get more attention.
Good talk behind the reason for the T2 chip and some other Apple security stuff
https://www.youtube.com/watch?v=Fm2tDgf40ss
https://www.youtube.com/watch?v=FklMpRiTeTA
not, the survival of the fittest. it is:
the extension of generosity of surplus
to other members in the ecological community
to build biodiversity
not the individual that survives but the community that survives
https://www.youtube.com/watch?v=yBkg70fhV2A
https://www.youtube.com/watch?v=jnV1u67_yVg
All talks will be livestreamed (and usually become available for download one day later) at [2].
[1] https://fahrplan.events.ccc.de/congress/2019/Fahrplan/
[2] https://media.ccc.de/
This [2] talk about building worlds in Blender by Ian Hubert.
[1] https://youtu.be/v5DqmTtCPiQ
[2] https://youtu.be/whPWKecazgM
Youtube video: https://www.youtube.com/watch?v=16slh29iN1g
Blog post: http://www.brendangregg.com/blog/2019-12-22/bpf-theremin.htm...
Why?
Brendan presents news tools and new flows that can be used to analyse performance bottlenecks.
He get you in the demo and you stay to receive more and more information. It seems to be a talk that he has been improving over the years. Totally worth your time seen it.