TimeStamps 👯👯 [0:00] Intro [2:55] Ryan Joins Stream [09:09] Ryan Answers Questions From Chat 1 [26:18] What is signals 2.0 ? [1:10:18] Ryan Answers Questions From Chat 2 [1:21:31] Ownership Model In Solid [1:26:18] Is IT Common To Nest Effects ? [1:28:19] Discussion With Chat [1:30:26] Three Types Of Rendering Frameworks [1:34:55] Solid Beyond HTML Elements [1:38:37] How Solid's Context Works ! [1:41:09] Discussion Chat Continues [1:47:19] Example Of Losing Reactivity [1:48:16] Error Boundaries [1:52:52] What Is A Control Boundary ? [1:56:30] Performance Of Proxies [2:01:30] What Really Is Signals 2.0 ? [2:05:35] createResource [2:21:54] Dependency tracking [2:33:12] createAsync [2:35:43] createRenderEffect [2:43:56] Synchronisation vs Derivation [2:47:58] createWritable [3:24:04] This Week In JavaScript [3:28:15] TWIJS: "You Can't Make Children Typesafe In React And Typesafe" [3:33:16] TWIJS: Default HTTP Method Of [3:38:04] TWIJS: "SSR For Docs Site Is A mistake ?" [3:40:17] TWIJS: Solid News [3:41:45] TWIJS React Complier's Type System [3:47:29] The World Beyond Components (Clip) [3:54:30] Crossing The Chasm [4:03:03] Angular And Wiz Are Merging ! [4:37:29] Appreciation To All Contributors [4:58:45] Conclusion
The createAsync discussion is like 2x more interesting in light of the React 19 suspense drama. As always it's cool to see how deeply you think about this stuff
The one thing I love the most of this streams is that when there is something I didnt even knew existed like a problem with an approach or anything, once its layed out by Ryan I basically just agree with his desicions on why he goes a certain path 10/10 I have always been considered smart by others but I know my limits, and for now it is experience on frontend and this limits the amount of issues I have found with frameworks, but every single time I come to a stream from Ryan I end up soaking up a lot of knowledge... unfortunately it has its downside... a lof of devs dont ever go that deep and its hard to explain them there's a better world outside of the bubble they are in
I love these streams every time you post, but it does take me like a week to get through each one. would you ever consider making videos in the 10-30 minute length that focus on a particular aspect of the stream? it's hard for me to keep up with regular 5h videos. also, would love 4k quality if you ever get the hardware to support it!
Part of the issue here is I use streamyard so upping the quality costs significantly more money. I should probably invest in a better setup at some point though. Similarly a 10-30 min "tight" video takes several hours to put together. I could spend my time doing that but I like being able to interact with chat and work through ideas in the open. My conference talks are like my tighter 20-30min content.
I've tried using signals in React, even built my own custom implementation for React. It works fine (at least improved DX and easy state sharing and zero useEffect) but it fails as the app grows bigger and bigger in size especially where you have complicated interconnected state since the signals aren't built into the core React's reactive system you start to run into issues
Wonder if the api's of signal based libraries will get much closer to each other since the proposal is out or will they just get closer internally instead of the api surface?
It isn't a clean slower. It makes some updates faster. But on average Preact with classes is ahead of Preact with Signals in the JS Framework Benchmark. And I stand behind the statement in general. The VDOM and Signals conflicting render methods are at odds with each other. There is a tradeoff you need to decide when making the benchmark implementation. It was ultimately why I didn't submit the Preact Signals implementation because I didn't want to be responsible for choosing which tradeoff to showcase. Vue had similar issues historically and had to add other mechanisms to work around this(v-memo). I imagine you could author the class version in a way that shared same characteristics as the Signal version and it would still be faster than the Signals version(even if slower than the current implementation on average).
Dude I appreciate the talk and this is meant as constructive criticism, but your thoughts are very difficult to follow in a format like this. I went through the first hour and I'm pretty sure your thoughts could be expressed better if you had some script in mind.. The result is that this talk is 5 hours long with a lot of stutter and just in time code editing.
These streams are as much for Ryan to think through stuff as they are for us. He could do his thinking ahead of time, in private, and then stream it in 30 mins. But, then he wouldn't get to jam with everyone, and we wouldn't get to participate. If you want to see a presentation you could wait for one of his conference talks.
Also, if you want more of a summary, wait for the chapters to be added then you can skip to the parts that interest you. Chapters are usually added a day or two after, once someone does them.
@@tomm5765 Its kinda stupid coz you are essentially wasting everyone time by posting a 5+hr stream without a short summary version. He can think in private and share the outcome in 10-30min videos still keep people updated on developments on whatever he's doing. There is a reason short video platforms like tiktok is hugely popular. I wanna be informed about signals 2.0 but Idgf about everything else going on in his head. At least post a summary video so I can judge if the 5hr version is worth my time. Not every topic deserve to follow a lexfridman format
@@josephs1732you basically want him to work for you for free, so you can get what you want, because you have no skills or patience to get it otherwise. No one is forcing you to waste your time, you'll just have to wait for someone to create a tiktok for you.
@@josephs1732why are you so entitled to think he needs to do anything for you? your wasting you own time being here, no one else is responsible for that.
ua-cam.com/video/sMbICJUGJj4/v-deo.html The worst case in solid: loosing reactivity. Managers at my workplace choose not to have reactivity while using angular. 🤣
TimeStamps 👯👯
[0:00] Intro
[2:55] Ryan Joins Stream
[09:09] Ryan Answers Questions From Chat 1
[26:18] What is signals 2.0 ?
[1:10:18] Ryan Answers Questions From Chat 2
[1:21:31] Ownership Model In Solid
[1:26:18] Is IT Common To Nest Effects ?
[1:28:19] Discussion With Chat
[1:30:26] Three Types Of Rendering Frameworks
[1:34:55] Solid Beyond HTML Elements
[1:38:37] How Solid's Context Works !
[1:41:09] Discussion Chat Continues
[1:47:19] Example Of Losing Reactivity
[1:48:16] Error Boundaries
[1:52:52] What Is A Control Boundary ?
[1:56:30] Performance Of Proxies
[2:01:30] What Really Is Signals 2.0 ?
[2:05:35] createResource
[2:21:54] Dependency tracking
[2:33:12] createAsync
[2:35:43] createRenderEffect
[2:43:56] Synchronisation vs Derivation
[2:47:58] createWritable
[3:24:04] This Week In JavaScript
[3:28:15] TWIJS: "You Can't Make Children Typesafe In React And Typesafe"
[3:33:16] TWIJS: Default HTTP Method Of
[3:38:04] TWIJS: "SSR For Docs Site Is A mistake ?"
[3:40:17] TWIJS: Solid News
[3:41:45] TWIJS React Complier's Type System
[3:47:29] The World Beyond Components (Clip)
[3:54:30] Crossing The Chasm
[4:03:03] Angular And Wiz Are Merging !
[4:37:29] Appreciation To All Contributors
[4:58:45] Conclusion
I apologise for the delay Solid Fam 🙏
really appreciate ❤
The createAsync discussion is like 2x more interesting in light of the React 19 suspense drama. As always it's cool to see how deeply you think about this stuff
The one thing I love the most of this streams is that when there is something I didnt even knew existed like a problem with an approach or anything, once its layed out by Ryan I basically just agree with his desicions on why he goes a certain path 10/10
I have always been considered smart by others but I know my limits, and for now it is experience on frontend and this limits the amount of issues I have found with frameworks, but every single time I come to a stream from Ryan I end up soaking up a lot of knowledge... unfortunately it has its downside... a lof of devs dont ever go that deep and its hard to explain them there's a better world outside of the bubble they are in
Starts at 2:55
good dog 🐶
And Topic starts at 25:59
@@riongull We warm up with the small talk in this stream
I love these streams every time you post, but it does take me like a week to get through each one. would you ever consider making videos in the 10-30 minute length that focus on a particular aspect of the stream? it's hard for me to keep up with regular 5h videos. also, would love 4k quality if you ever get the hardware to support it!
Part of the issue here is I use streamyard so upping the quality costs significantly more money. I should probably invest in a better setup at some point though.
Similarly a 10-30 min "tight" video takes several hours to put together. I could spend my time doing that but I like being able to interact with chat and work through ideas in the open. My conference talks are like my tighter 20-30min content.
Wondering when signals are coming in React. 🤔
Never, react will die, maybe after svelte 5 coming out
React compiler will do that most likely
@@StingSting844 React forget is nothing like signals and more like old svelte
I've tried using signals in React, even built my own custom implementation for React. It works fine (at least improved DX and easy state sharing and zero useEffect) but it fails as the app grows bigger and bigger in size especially where you have complicated interconnected state since the signals aren't built into the core React's reactive system you start to run into issues
Wonder if the api's of signal based libraries will get much closer to each other since the proposal is out or will they just get closer internally instead of the api surface?
1:52:10 According to js-framework-benchmark adding signals to preact is not makes it slower
It isn't a clean slower. It makes some updates faster. But on average Preact with classes is ahead of Preact with Signals in the JS Framework Benchmark.
And I stand behind the statement in general. The VDOM and Signals conflicting render methods are at odds with each other. There is a tradeoff you need to decide when making the benchmark implementation. It was ultimately why I didn't submit the Preact Signals implementation because I didn't want to be responsible for choosing which tradeoff to showcase.
Vue had similar issues historically and had to add other mechanisms to work around this(v-memo). I imagine you could author the class version in a way that shared same characteristics as the Signal version and it would still be faster than the Signals version(even if slower than the current implementation on average).
Dude I appreciate the talk and this is meant as constructive criticism, but your thoughts are very difficult to follow in a format like this.
I went through the first hour and I'm pretty sure your thoughts could be expressed better if you had some script in mind..
The result is that this talk is 5 hours long with a lot of stutter and just in time code editing.
These streams are as much for Ryan to think through stuff as they are for us. He could do his thinking ahead of time, in private, and then stream it in 30 mins. But, then he wouldn't get to jam with everyone, and we wouldn't get to participate. If you want to see a presentation you could wait for one of his conference talks.
Also, if you want more of a summary, wait for the chapters to be added then you can skip to the parts that interest you. Chapters are usually added a day or two after, once someone does them.
@@tomm5765 Its kinda stupid coz you are essentially wasting everyone time by posting a 5+hr stream without a short summary version. He can think in private and share the outcome in 10-30min videos still keep people updated on developments on whatever he's doing. There is a reason short video platforms like tiktok is hugely popular. I wanna be informed about signals 2.0 but Idgf about everything else going on in his head. At least post a summary video so I can judge if the 5hr version is worth my time.
Not every topic deserve to follow a lexfridman format
@@josephs1732you basically want him to work for you for free, so you can get what you want, because you have no skills or patience to get it otherwise. No one is forcing you to waste your time, you'll just have to wait for someone to create a tiktok for you.
@@josephs1732why are you so entitled to think he needs to do anything for you? your wasting you own time being here, no one else is responsible for that.
ua-cam.com/video/sMbICJUGJj4/v-deo.html
The worst case in solid: loosing reactivity.
Managers at my workplace choose not to have reactivity while using angular. 🤣