30:41 the laugh of trying to attain web scale

42:21 GZip Is All You Need mentioned

You must assume A and B = 1 and this assumption does not discharge so you have proved nothing if A and/or B = 0

Nice talk, esp. the part about variance.

Really good talk. I enjoyed it much.

It seems this speaker leans more into the "Iterativist" school of thought instead of the "Formalist" school of thought. While there is value in quick feedback loops and flexibility, static types are akin to labelling wires when working on electronics. Things become messy without those labels. Likewise, in how example of the space probe being debugged in space, why not do a proof by contruction with the best spec we can make first, then if we are wrong we just iterate from there? I personally lean towards formalism due to my PTSD from dealing with JS runtime errors and because I am very conscientious. However, I recognize the power and beauty of the Iterativist approach too. That is why I considered ClojureScript and Elixir for the longest before finally arriving at ReScript.

this amazing man, very smart but is not showing anything concrete, how to make things simple???? all I got out of this talk is just some nice jokes, and high level concepts. How to I apply it????

Amazing Presentation👍

In the first two minutes, rich explains: "good programs should be made out of processes and queues". If this statement triggers your gag reflex - that is, if you don't want to work on programs built like this, but rather want first-class API abstractions between components; the remainder of the talk will confirm your intuitions. CSP is really not a good basis for comprehensible concurrency, I understand that the watered down abstractions that javascript inherited from E are not perfect, but they are a lot more tractable in practice than channel use. Yes, Rich says that you "can send reply channels over the request channel" - but why not just use an abstraction that makes this the default, as promises do? That way you can have dynamic structure in your code. I wonder a little if there has been a walking-back of core.async use in the broader clojure community in the last 10 years. I remember it being difficult to find libraries that did not use it.

What a fantastic talk! To the point and witty.

This was super helpful and well explained!

40:12 So for the difficult bit of finding similar things within a distributed system, rather than compressing the actual bytes of two things and diffing them, you most certainly want to utilize neural networks. For every piece of content an author cares to make discoverable, they'll want to generate multi-modal embeddings for the most popular neural network(s), and if they want it to continue to be discoverable, they'll want to continually generate new embeddings as new versions of those neural networks are published. Then, whenever you want to find similar things, you just send out requests to the network to find content with embeddings within a certain vector space radius of the original search input, and the radius would grow as more results are needed. I do wonder if the ideal internet would do well to have only ONE neural network to perform these embeddings. It would continually be training itself on new content.

imagine how crazy effective these abstractions would be when they're running on a TPU...

Landlords are everywhere you look. You start seeing them everywhere. You buy a toaster and then it's yours, but I know you have all kinds of stuff that you pay for that isn't really yours because everyone does.

Makes me want to learn Elixir... Any starter project ideas？

At 20:33 what does he mean by the "switching cost... for small authors"? Is he referring to the cost of switching TO the independent language? And how does this relate to get Jeff'd? (I.e. getting your users taken by a bigger company which is hosting your language without having to do any compiler engineering)

By "hosting" does he mean maintaining build servers and VM servers for people to use? Like being a cloud hosting provider devoted to a specific programming language?

Thanks these videos were incredibly helpful

At a developer meetup, the devs there told me to learn TS but I expressed my concerns. I told them that JS is like living in a house without a fence, so as a result you are naturally more guarded and on edge for potential threats. TS is like having a fence, but there is a massive hole in it. I argue that this is even worse because you think you are safe but you are not, so you let bugs slip in due to your false safety. So therefore, using TS is actually worse than using JS that has no guard rails. I argued that if you want to use static typing in JS, you should go with a solution that has a sound type system, or else you will have a false sense of security. I recommended ReScript and said it has a sound type system and is based on OCaml and has a very nice syntax.

Bertrand, Bernard. It seems that it's bERnəd in England, bərnARd in the US. It seems that its bERtrənd in England, and mainly bERtrənd and sometimes bərtrAnd in the US. Capital letters here denote stress (louder).

vector clock 31:38

I watched this a couple of months ago and thought “wow what a good talk”, then last week I was coding up some geometry based on a frustratingly written standard, and realised it was so unintuitive because they were effectively working in higher dimensions with 2D tools and came back to watch this video again, and I’m even more impressed than before. Currently working on a GA solution, thank you for such an informative and entertaining talk

Excellent talk!

Can you use reactive variables instead of functional for good repl based development. Reactive variables are like getter/setters that notify their dependents to update themselves based on a formula.

Wouldn't make more sense for dig to be an operation that pops the top of the stack and digs that far?

Wonder how these would do as crypto miners.

Thanks for this presentation, very content dense but digestible enough. Here from Bend/HVM

Excellent presentation, the explanation of Raft was clear and not hard to follow. Also good questions!

"Systems programmers are high priests of a low art." -- Robert S. Barton

lots of interesting and non intuitive points here

I miss having the option for "sort:indexed" in code searches on the new search engine... It's frustrating to sift through search results from six years ago, especially when dealing with outdated code that's no longer relevant. Libraries evolve rapidly, making it crucial to prioritize current and relevant results to stay efficient...

what is the weight of snapshot of DB sharing across multiple DC and each time sharing that to be in sync doesn't it effect the efficiency of system.

Performance of devices reletive to the things software companies write only starts to increase when handware performance increases outpace software scope increases. And even then software scope increases have to have a linear relationsbip to performance cost. Yet we all know thats not how it ends up working. Increasing scope by 10% does not increase performance cost by just 10%. And thats only the case when hardware performance increases are actually utilized for peformance. Rather then energy usage reduction, or new features(like AI portions), or smaller die size, or composability. So no the premise that we are moving towards software being easier to run is not clearly happening. Seondly bandwidth is still a bottleneck. Korea, Austria, nd many other places have companies preventing the cost of bandwidth going down in those areas. Also the same applies here. Bandwidth cost has to outpace data size. Lastly latency has become more important yes, but also not really. Its been very important for a long time. Its not like its become more important. Also note that many many countries globally have reletively low access to bandwidth, performance, and latency. The premise set out at the start is drasticly more situational than how it is actually framed.

That list of algebraic properties is a really memorable phrasing. It'd be useful for education. I'll reproduce it below. Associative: grouping doesn't matter! Commutative: order doesn't matter! Idempotent: duplicates don't matter! Identity: this value doesn't matter! Zero: other values don't matter!

Sometimes notebook/live coders create programs only they can run, because there is a certain dance you have to do with the machine, which acts like a magical incantation that produces output. The reason for Haskell style types is that the compiler will help you find the missing piece that fits the hole. Due to my low working memory I love machines that think for me. With unicode, most languages already support glyphs which are weirder than cosplaying tty.

Statistics programs exist for factories and others for stuff like A B testing. See JMP for an example.

Thanks for this

You can get benefits of LLVM compilation AND utilize all the runtime hotspot information that languages that arent a static binary can. This exists in Java and the JVM languages like clojure and kotlin under the VM "Azule plateform prime". This utilizes all the information you get from seeing how the program is being used in production. And is able to combine that information with the ability to convert hot live code into optimized llvm compiled code. You mention garbage collection pauses. That same VM doesnt have pauses when garbage collecting.

Languages like clojure have tools that allow for near full mutability performance, without the risks and problems that come with mutability.

I think making an interface between Elm and graalVM would probly be an easier way to get a binary compilation method. Rather than remaking the whole language with a different compile target.

Thank you for a fascinating lecture. Does anyone know where one can find the schematic diagrams of the Voyager radio transmitter / receiver? I have seen a block diagram of the radio system but no a schematic down to component level. I would be really interested to see this. Hope someone can help.

im peruvian, I have no deal doing herr but damn, these have been one of the best series of conferences that i've watched ever. I discovered it on 2022, it was an awesome ride.

This is excellent

12:14 malloc can return null though

One thing I'd like to see is how AIMD is configured. How do you decide backoff factor in multiplicative decrease and additive factor in Additive increase?

Consciousness is a big library where all the concepts are making iterations to be the more efficient for the being they're computing for....

didn't expect a Harry Mack namedrop in a Strange Loop talk!

This is an excellent presentation. Many things seem to have evolved since, such as the Rust language, and that all fields are optional. Is there a newer version of this presentation?

Toolsets for making more toolsets, genius thinking

This is very good