1.09 Bn users log in, 330mm use video chat
780mm users view Moments feed (like FB feed), 120mm users post to Moments
360mm read articles, 400mm use mini programs
Allen Zhang the creator of WeChat (I am a fangirl) gave a 2 hr long speech yesterday abt his thoughts on the product on the eve of its 10th anniversary.
It's a lot on Channels (the new video function in WeChat).
Here are my notes & comments in parentheses:
1.09 Bn users log in, 330mm use video chat
780mm users view Moments feed (like FB feed), 120mm users post to Moments
360mm read articles, 400mm use mini programs
In last 5 years, users are sending 33x more video msgs in chat, 10x more videos into their Moments feed.
Not everyone can write long content, but doing video requires new acct system (current public acct system is subscription based), we tried in 2017
Channels is NOT like your WeChat ID, which is for PRIVATE comms. It's WeChat's foray into public content.
WeChat tried 3 separate internal algo teams. Found that algo rec was much worse than people rec content performance.
But it's not that algos are bad, more that algos need large existing pool of content (this is true, ByteDance began w ”stealing" content).
Follows : Friend Rec : Machine Rec :: 1 : 2 : 10
Apparently right now it is already 1:2:9.
Friend Rec likely to be "knowledge based" content, Machine Rec likely to be brainless entertainment content.
Allen thinks video "files" will die. They don't make sense bc value of video is so much the creator info & viewer interactions. So that is how Channel Videos show up on Moments (link / cloud).
Most ppl distinguish short video as <1min. But this is artificial, we don't have a limit. Just that the 1st min becomes the "intro" to a "long" video.
We hope more long videos come to WeChat. (Woah, shots fired?)
Side note: lots of WeChat users now use "Last 3 days only" for their WeChat Moments feed. (I use 6 mos, LOL). So not much shown when
It's all abt "hit (clickthru) rate." It's like ur Pub Accts, the more u subscribe, the less ur hit rate. So we added "Frequently Read" at top (shows as circles on top of your feed).
Content always goes to lower barrier of creation. Livestreaming is easier than short video. It's just chatting (I think he's oversimplifying but yes, it's easier to get started). A meteor shower livestream got 1mm viewers just thru social WoM. Was surprising.
We're enhancing livestreaming functionality in mini programs for ecommerce. (This we already covered on @techbuzzchina Ep. 75).
(OK I think this is GENIUS esp w new lockdowns due to COVID flareup.)
In typical Allen style, NO.
We're not gonna spend any money bc the platform when well designed should be able to let creators feed themselves. We also want authentic content.
At WeChat we want the product to speak for itself, not rely on operations. (OK I get it Allen!)
So now here are the crazy ideas:
Your profile should reflect your state. We tried to have video profile but was hard. But still abt 1mm use it a day. We'll give ppl ability to write something (text, not video) very short re: their state. Purpose of social network is to find like-minded folks.
After video, lots of ppl aren't "listening" to music except when driving since they're "watching" so much stuff. Allen feels strongly current players suck, just show a rotating record (who still has records these days? oh yeah my hipster husband does.)
4) floating windows. I hate these, but since u can get interrupted in middle of long content they're
5) Input keyboard. We are R&Ding our own. Lots of ppl think we sell ur chat history, no. We don't analyze it at all. (So do 3rd party keyboards?)
Either u not sure what u're doing and do nothing, or
Channels is only 1-200 folks, incl 3 algo teams, this includes everyone, backend, ops etc. (OK this is legit a v small number of ppl, cuz Chinese internet cos are HUGE generally, 1-200 is a small Series A/B prod, since 2010.)
1) Connect 2) Simple.
Lots of social products connect just ppl, but we connect info, services, currency etc. We focus on decentralization, bc connection implies decentralized. (Lots might disagree but I agree, esp vs other Chinese cos)
We've added a lot of functions, but today's WeChat is still as simple as when we started (oh boy, I saw lots of folks disagreeing w this).
I hope it maintains its soul, beauty, creativity and style.
More from Tech
I think about this a lot, both in IT and civil infrastructure. It looks so trivial to “fix” from the outside. In fact, it is incredibly draining to do the entirely crushing work of real policy changes internally. It’s harder than drafting a blank page of how the world should be.
I’m at a sort of career crisis point. In my job before, three people could contain the entire complexity of a nation-wide company’s IT infrastructure in their head.
Once you move above that mark, it becomes exponentially, far and away beyond anything I dreamed, more difficult.
And I look at candidates and know-everything’s who think it’s all so easy. Or, people who think we could burn it down with no losses and start over.
God I wish I lived in that world of triviality. In moments, I find myself regretting leaving that place of self-directed autonomy.
For ten years I knew I could build something and see results that same day. Now I’m adjusting to building something in my mind in one day, and it taking a year to do the due-diligence and edge cases and documentation and familiarization and roll-out.
That’s the hard work. It’s not technical. It’s not becoming a rockstar to peers.
These people look at me and just see another self-important idiot in Security who thinks they understand the system others live. Who thinks “bad” designs were made for no reason.
Who wasn’t there.
The tragedy of revolutionaries is they design a utopia by a river but discover the impure city they razed was on stilts for a reason.
— SwiftOnSecurity (@SwiftOnSecurity) June 19, 2016
I’m at a sort of career crisis point. In my job before, three people could contain the entire complexity of a nation-wide company’s IT infrastructure in their head.
Once you move above that mark, it becomes exponentially, far and away beyond anything I dreamed, more difficult.
And I look at candidates and know-everything’s who think it’s all so easy. Or, people who think we could burn it down with no losses and start over.
God I wish I lived in that world of triviality. In moments, I find myself regretting leaving that place of self-directed autonomy.
For ten years I knew I could build something and see results that same day. Now I’m adjusting to building something in my mind in one day, and it taking a year to do the due-diligence and edge cases and documentation and familiarization and roll-out.
That’s the hard work. It’s not technical. It’s not becoming a rockstar to peers.
These people look at me and just see another self-important idiot in Security who thinks they understand the system others live. Who thinks “bad” designs were made for no reason.
Who wasn’t there.
THREAD: How is it possible to train a well-performing, advanced Computer Vision model 𝗼𝗻 𝘁𝗵𝗲 𝗖𝗣𝗨? 🤔
At the heart of this lies the most important technique in modern deep learning - transfer learning.
Let's analyze how it
2/ For starters, let's look at what a neural network (NN for short) does.
An NN is like a stack of pancakes, with computation flowing up when we make predictions.
How does it all work?
3/ We show an image to our model.
An image is a collection of pixels. Each pixel is just a bunch of numbers describing its color.
Here is what it might look like for a black and white image
4/ The picture goes into the layer at the bottom.
Each layer performs computation on the image, transforming it and passing it upwards.
5/ By the time the image reaches the uppermost layer, it has been transformed to the point that it now consists of two numbers only.
The outputs of a layer are called activations, and the outputs of the last layer have a special meaning... they are the predictions!
At the heart of this lies the most important technique in modern deep learning - transfer learning.
Let's analyze how it
THREAD: Can you start learning cutting-edge deep learning without specialized hardware? \U0001f916
— Radek Osmulski (@radekosmulski) February 11, 2021
In this thread, we will train an advanced Computer Vision model on a challenging dataset. \U0001f415\U0001f408 Training completes in 25 minutes on my 3yrs old Ryzen 5 CPU.
Let me show you how...
2/ For starters, let's look at what a neural network (NN for short) does.
An NN is like a stack of pancakes, with computation flowing up when we make predictions.
How does it all work?
3/ We show an image to our model.
An image is a collection of pixels. Each pixel is just a bunch of numbers describing its color.
Here is what it might look like for a black and white image
4/ The picture goes into the layer at the bottom.
Each layer performs computation on the image, transforming it and passing it upwards.
5/ By the time the image reaches the uppermost layer, it has been transformed to the point that it now consists of two numbers only.
The outputs of a layer are called activations, and the outputs of the last layer have a special meaning... they are the predictions!
You May Also Like
And here they are...
THE WINNERS OF THE 24 HOUR STARTUP CHALLENGE
Remember, this money is just fun. If you launched a product (or even attempted a launch) - you did something worth MUCH more than $1,000.
#24hrstartup
The winners 👇
#10
Lattes For Change - Skip a latte and save a life.
https://t.co/M75RAirZzs
@frantzfries built a platform where you can see how skipping your morning latte could do for the world.
A great product for a great cause.
Congrats Chris on winning $250!
#9
Instaland - Create amazing landing pages for your followers.
https://t.co/5KkveJTAsy
A team project! @bpmct and @BaileyPumfleet built a tool for social media influencers to create simple "swipe up" landing pages for followers.
Really impressive for 24 hours. Congrats!
#8
SayHenlo - Chat without distractions
https://t.co/og0B7gmkW6
Built by @DaltonEdwards, it's a platform for combatting conversation overload. This product was also coded exclusively from an iPad 😲
Dalton is a beast. I'm so excited he placed in the top 10.
#7
CoderStory - Learn to code from developers across the globe!
https://t.co/86Ay6nF4AY
Built by @jesswallaceuk, the project is focused on highlighting the experience of developers and people learning to code.
I wish this existed when I learned to code! Congrats on $250!!
THE WINNERS OF THE 24 HOUR STARTUP CHALLENGE
Remember, this money is just fun. If you launched a product (or even attempted a launch) - you did something worth MUCH more than $1,000.
#24hrstartup
The winners 👇
#10
Lattes For Change - Skip a latte and save a life.
https://t.co/M75RAirZzs
@frantzfries built a platform where you can see how skipping your morning latte could do for the world.
A great product for a great cause.
Congrats Chris on winning $250!
#9
Instaland - Create amazing landing pages for your followers.
https://t.co/5KkveJTAsy
A team project! @bpmct and @BaileyPumfleet built a tool for social media influencers to create simple "swipe up" landing pages for followers.
Really impressive for 24 hours. Congrats!
#8
SayHenlo - Chat without distractions
https://t.co/og0B7gmkW6
Built by @DaltonEdwards, it's a platform for combatting conversation overload. This product was also coded exclusively from an iPad 😲
Dalton is a beast. I'm so excited he placed in the top 10.
#7
CoderStory - Learn to code from developers across the globe!
https://t.co/86Ay6nF4AY
Built by @jesswallaceuk, the project is focused on highlighting the experience of developers and people learning to code.
I wish this existed when I learned to code! Congrats on $250!!
