Maggie with a J

Surfer. immigrant founder. annoyingly inquisitive. early bytedance AI/tiktok alum, now at meta. opinions are my own

[Maggie with a J @maggiewithj]

@zarazhangrui How about muse

[Maggie with a J @maggiewithj]

Before I share how good Muse is on Shopping… This is how terrible chatGPT shopping is.

I don’t bother even to show you the steps before this….its like the worse sales rep in could encounter in a department store:
Chat
https://t.co/eYlDFuFRjm

Work
https://t.co/teMPyx0AkS

[Alexandr Wang @alexandr_wang]

we are excited for muse to be partnering deeply with @Shopify to enable agentic checkout with Shop Pay on all Shopify stores!

we want to give our musers access to a wide range of stores to find the absolute perfect products ♥️

[Maggie with a J @maggiewithj]

Just too many words to read. Instinct and Ollie’s onboarding screenshots were more mobile first, and visual first. With the data Meta has, this would be personalized, dynamic and an eye candy. Open your imagination Alex!

[Alexandr Wang @alexandr_wang]

check out the muse ideas tab for inspo on how to use muse!

[Maggie with a J @maggiewithj]

@alexandr_wang shopping is also crushing it! But this page has too many words to read. Try instincts on boarding(sending screnshots) + and make it personalized

[Maggie with a J @maggiewithj]

@alexandr_wang @natfriedman finally also learned the good stuff from doubao!! Proud of Meta this time as an old tiktoker. While doubao now looks like f*cking ChatGPT just bc of the whole work buddy narrative.

[Maggie with a J @maggiewithj]

@quxiaoyin @agentsky_dev benchmarks always need a dollar budget and latency budget attached

[Maggie with a J @maggiewithj]

@klazizpro @AvitalBalwit @grok Lolol grok failed you

[Bill Sun @BillSun_AI]

https://t.co/Vr0oiqR0Sr

[Maggie with a J @maggiewithj]

Everyone wanted to build a data flywheel.
Everyone talked about evaluation.
Then Monday comes.
Your smartest engineer wants to “train a better model.”
Nobody volunteers to eyeball 500 failed cases.
Yet that's exactly why some teams build stronger flywheels than others...

[Maggie with a J @maggiewithj]

Why Meta’s layoff and attempt to be AI native failed? Even Zack is still the CEO
1/ didn’t fire execs
2/ execs need jobs unless they are major shareholders
3/ didn’t change the old process (processes, measurements ) so you end up reinventing the roles that being eliminated

[Xiaoyin Qu @quxiaoyin]

I went a dinner a few weeks ago with a bunch of enterprise execs who told me "we will never use Chinese models." "Even if it's 100x cheaper?" "No, we care about safety and security."

1. They don't understand when they host open-source models with their own GPUs or US data centers, they won't share their data to China.
2. They are giving away all their data to OpenAI and Anthropic rather than owning it privately themselves.
3. They don't understand math. 100x is a big number and lots of profits.

It's almost July 2026 now. If your execs still talk like that, fire them now.

[Maggie with a J @maggiewithj]

for the same exact reasons, this is why last gen companies , even they are building LLM models, cannot build ai native organization - including google and meta

[Xiaoyin Qu @quxiaoyin]

I went a dinner a few weeks ago with a bunch of enterprise execs who told me "we will never use Chinese models." "Even if it's 100x cheaper?" "No, we care about safety and security."

1. They don't understand when they host open-source models with their own GPUs or US data centers, they won't share their data to China.
2. They are giving away all their data to OpenAI and Anthropic rather than owning it privately themselves.
3. They don't understand math. 100x is a big number and lots of profits.

It's almost July 2026 now. If your execs still talk like that, fire them now.

[Maggie with a J @maggiewithj]

@trq212 Huge! Very interested my friend

[Maggie with a J @maggiewithj]

@trq212 How would you compare this vs dispatch

[Maggie with a J @maggiewithj]

1/ Personal software where demand is also supply . Plus, the integration with content backed by unique human insight.

2/ sensor layer is so painful. Similarly I spent most time on data while I’m remixing a baby tracker for myself. Huge opportunity

[Andrej Karpathy @karpathy]

Very interested in what the coming era of highly bespoke software might look like.

Example from this morning - I've become a bit loosy goosy with my cardio recently so I decided to do a more srs, regimented experiment to try to lower my Resting Heart Rate from 50 -> 45, over experiment duration of 8 weeks. The primary way to do this is to aspire to a certain sum total minute goals in Zone 2 cardio and 1 HIIT/week.

1 hour later I vibe coded this super custom dashboard for this very specific experiment that shows me how I'm tracking. Claude had to reverse engineer the Woodway treadmill cloud API to pull raw data, process, filter, debug it and create a web UI frontend to track the experiment. It wasn't a fully smooth experience and I had to notice and ask to fix bugs e.g. it screwed up metric vs. imperial system units and it screwed up on the calendar matching up days to dates etc.

But I still feel like the overall direction is clear:
1) There will never be (and shouldn't be) a specific app on the app store for this kind of thing. I shouldn't have to look for, download and use some kind of a "Cardio experiment tracker", when this thing is ~300 lines of code that an LLM agent will give you in seconds. The idea of an "app store" of a long tail of discrete set of apps you choose from feels somehow wrong and outdated when LLM agents can improvise the app on the spot and just for you.
2) Second, the industry has to reconfigure into a set of services of sensors and actuators with agent native ergonomics. My Woodway treadmill is a sensor - it turns physical state into digital knowledge. It shouldn't maintain some human-readable frontend and my LLM agent shouldn't have to reverse engineer it, it should be an API/CLI easily usable by my agent. I'm a little bit disappointed (and my timelines are correspondingly slower) with how slowly this progression is happening in the industry overall. 99% of products/services still don't have an AI-native CLI yet. 99% of products/services maintain .html/.css docs like I won't immediately look for how to copy paste the whole thing to my agent to get something done. They give you a list of instructions on a webpage to open this or that url and click here or there to do a thing. In 2026. What am I a computer? You do it. Or have my agent do it.

So anyway today I am impressed that this random thing took 1 hour (it would have been ~10 hours 2 years ago). But what excites me more is thinking through how this really should have been 1 minute tops. What has to be in place so that it would be 1 minute? So that I could simply say "Hi can you help me track my cardio over the next 8 weeks", and after a very brief Q&A the app would be up. The AI would already have a lot personal context, it would gather the extra needed data, it would reference and search related skill libraries, and maintain all my little apps/automations.

TLDR the "app store" of a set of discrete apps that you choose from is an increasingly outdated concept all by itself. The future are services of AI-native sensors & actuators orchestrated via LLM glue into highly custom, ephemeral apps. It's just not here yet.

[Maggie with a J @maggiewithj]

everyone in China study the west in the nauseating way - they listen to all podcast, all the twitters and not even mentioned to every public company or policies move, or tech/VC news . Collectively. But not the opposite.

Regardless the willingness, it’s the arrogance.

A personal experience: This was true, in 2017/2018 Bytedance, would study everything pop on the IOS App Store, app annie and even early startup from the incubators… but how about Facebook?

Lasso was initiated quite early by some brilliant engineers and PMs noticed the Chinese TikTok, but never get their way to the probably “arrogant” middle/high level executives or Zack.

And you cannot blame them, when the entire Valley, other than a few investors, probably called Bytedance a Chinese scammy company when I landed here

And Reels is not until post 2020…

So glad Bill and Brad are open minded and hopefully educate the rest

[Bg2 Pod @BG2Pod]

BG2. Inside China’s Tech Surge: EVs, AI, and How America Can Compete 👊💥@bgurley

[Maggie with a J @maggiewithj]

@SkyzSmokey Hope the Canada college buildings will be fine

[Maggie with a J @maggiewithj]

@SkyzSmokey spreading fast. Causing 280 pause / slowdown

[Maggie with a J @maggiewithj]

@SkyzSmokey getting scattered fast

[Maggie with a J @maggiewithj]

@natfriedman Not allowed to use this on Meta computer make me 🤯…..

[Maggie with a J @maggiewithj]

Proud this still exists in America 🗽

note: I was told by my American friends when I just moved here that you should not share these not-so-good this things in China…and that was my first “lesson” of identity politics, and “being DEI”.

Just like anything else, being BASED and forming your own opinion is not convenient.

Don’d be LAZY, American!

[Maggie with a J @maggiewithj]

loved this. And noted that your styles is very different hahhahahah when I see you in different days/occasion! Fun.

What the friend told you was exactly what I was told when I was in China, not just towarlds engineers but also founders and investors: Female founders were suggested to REMOVE their makeup when they go pitch VCs, female investors raise eyebrows and suspicions if you “ spent too much time with a male founder and try to close the deal”

But now i learned, as you said, the best people are extremely merit-based and don't care. It’sfortunate that we still have this

Was just talking to a female technologist aboh this, and broadly how to be a woman we wanted to be. Thank you for sharing this great way of being, being yourself!