Ganet grant

founder at a cloud ai startup, San Franciscan/Beijingren/Californian living in Shanghai.

[Ganet grant @tocgg]

@zhang_benita Manus, 我错了！https://t.co/yEchZDYrdY

[Ganet grant @tocgg]

Is that bill curley from benchmark? @benchmark at Shanghai

[Ganet grant @tocgg]

@hlntnr @jjding99 That’s a thing if u have to do to b infra; I know lots of companies doing that , 硅基流动 ie

[Nikita Bier @nikitabier]

After 10 years of building consumer social apps, I've decided to start exploring new areas. Building these products is an unforgiving grind—but I learned a lot along the way.

For those embarking on this path, here's everything you need to know:

TIME FOR A THREAD 👇

[Ganet grant @tocgg]

@nikitabier Applicable everywhere, including China !

[Arnaud Bertrand @RnaudBertrand]

The reactions to Luanna Jiang's Harvard commencement speech in both China and the US are really interesting to study, in that they are so illustrative of the current zeitgeist in both countries.

First of all, let's be clear, her speech wasn't well received on either "side" at all.

Just to verify this, last night I showed a picture of Luanna Jiang to my wife and some of her friends, all Chinese mums in their 40s who don't follow politics at all, and asked them "Have you heard of this girl?" and they all replied in unison "Yes, she's hated in China." Granted, not a perfect proxy for the Chinese mood overall, but pretty illustrative nonetheless.

Same story on the US side where, unsurprisingly given the current paranoid McCarthyist atmosphere - her being Chinese and this being Harvard - many people saw her as a validation of the Trump administration's unhinged narrative that the university is "coordinating with the Chinese Communist Party on its campus."

This already contains in itself a contradiction: criticized in the US for being an agent of China, but actually hated in China.

I won't dwell too much on the US side: I think everyone understands all too well the paranoid binary thinking that's taken hold, where any Chinese person who doesn't actively denounce their origins is automatically suspect.

The Chinese side is a bit more nuanced and counter-intuitive, as you'd think that the first-ever Chinese national to give a commencement speech at Harvard would be seen positively, if only because it represents a breakthrough moment for Chinese representation abroad.

But no, it wasn't perceived like that at all.

First of all, her speech itself (of which you can read a transcript here: https://t.co/SMKMzWIcVr) was problematic.

If you read it, there isn't actually anything fundamentally wrong with it. It's her own story, that of a privileged international student reflecting on "shared humanity" and global interconnectedness - the sort of saccharine, buzzword-filled idealism that may sound profound in a Harvard auditorium but is ultimately just hollow platitudes.

Why was it badly received in China? For the very reason that it sounded like Harvard commencement speech template #5, that there wasn't anything real or authentic about it, the kind of speech designed to speak to a Western liberal elite audience.

Fundamentally the Chinese were disappointed that the first-ever Chinese national to give a commencement speech at Harvard didn't sound Chinese at all: she sounded like a hollow Western liberal.

As such, this wasn't China on stage, she was the symbol of everything ordinary Chinese people resent about a subset of their own elites: someone who had reneged on the real Chinese identity and transformed herself into a perfectly polished Western liberal mouthpiece; cutting any genuine connection to her roots in favor of assimilation into Western orthodoxy.

There's that expression in China: "Baizuo" (literally "white left"), which has very negative connotations. It's a perfect adjective for her speech: exactly the sort of condescending, out-of-touch idealism that the Chinese - who are fundamentally pragmatists and hate words over substance - see as the worst kind of privileged moral posturing, offering performative virtue signaling while ignoring the real challenges people actually face.

Luanna Jiang gave a separate interview (https://t.co/3iOJkYkGY8) that went viral on Chinese social media where you can see her speak about her idealism on the Harvard campus while you can literally see two guys fighting in the background: you couldn't find a better encapsulation of the "baizuo" mentality - completely absorbed in delivering her own moral platitudes while remaining utterly oblivious to the real-world conflict happening literally right behind her.

All that being said it wasn't her speech itself that got her the most backlash in China but rather her reaction to some of the criticism of it on Weibo (where, tellingly, her username is 哈佛蒋雨融Luanna, i.e. "Harvard Jiang Yurong Luanna"), as well as Chinese netizens digging into her background.

Her first kneejerk reaction to the criticism was to invalidate it by saying she'd received a lot of positive reactions to it in the Harvard yard, which of course demonstrated exactly what Chinese netizens found so infuriating - she was using approval from American liberals as her yardstick, apparently oblivious to the fact that this was precisely what she was criticized for.

Then Chinese netizens started to look into her background, which only fueled the anger further.

They discovered she was a classic "富二代" (second-generation rich) who had attended elite UK boarding schools and lived a life of extraordinary privilege, exactly the kind of wealthy overseas Chinese that ordinary people resent.

Worse still, people found timeline inconsistencies in her resume and began questioning whether she actually holds Chinese citizenship at all - potentially making her entire "first Chinese national to give Harvard commencement speech" narrative a complete fraud.

All in all, what emerged was the picture of someone who embodied everything ordinary Chinese people despise: a wealthy overseas Chinese elite who had bought her way into Western institutions, adopted Western liberal values wholesale, and was now trying to speak for China while having no authentic connection to the Chinese experience whatsoever.

What does this all say about the Chinese zeitgeist today? In some ironical way, it's not THAT different from the pushback you get in the U.S. against "globalists": in both cases, you have popular anger directed at privileged elites who are seen as having traded their authentic national identity for membership in a transnational liberal class that speaks in the polished platitudes of global liberalism, and is fundamentally disconnected from the concerns of ordinary people in their home countries.

And, again ironically, that might in fact be an interesting way to bring China and the US closer. Say for the sake of argument that Luanna Jiang, instead of giving her hollow liberal speech had stood up at Harvard and delivered a blistering critique of how we're creating a rootless international class that speaks in sanitized platitudes while remaining deaf to their own people's actual struggles - such a message would have found enthusiastic audiences in both Beijing and Middle America.

The irony is that it'd have been infinitely more compatible with the messaging of the Chinese government which itself loathes "baizuo" liberalism, but she also likely wouldn't have faced accusations of being a "CCP agent" in the U.S.

Which also in itself says a lot about the need to have actual real China voices in the U.S. as opposed to voices like Jiang's: if Americans actually encountered genuine Chinese worldview - with its distrust of rootless elites and emphasis on cultural authenticity - they might discover they actually like it.

[Ganet grant @tocgg]

@RnaudBertrand Well written. As a Chinese nationals who lived in US for 21 years and now live in Shanghai, I appreciate your effort and the fairness and neutrality your writing represents, and the positive tune and solution on otherwise strained relationships.

[Ganet grant @tocgg]

@karpathy And Chinese

[Chicken Genius @chickengenius]

This is my last post trying to share the truth about China. Not pro China neither pro USA. All I want to do, is show you both sides and stop being a dumb mfer to the propaganda machine. It works both ways. After this, I'll only talk about the markets.

Great examples. Two biggest social media channels covering the world.
Tiktok: Dylan Page 15M followers
Youtube: ishowspeed: 39M followers

Somehow, when they covered China, they were labelled as "ccp dogs"

You can pick and choose whatever matrix that suits you most. That's your life, but imposing your belief on another culture is kinda dumb.

End of day:
1) groceries affordable?
2) is it safe?
3) do you have a comfortable house?
4) health care cheap and good?

Sharing a simple video below this post. Widely travelled people will give you the same answer.

Peace to both nations. USA great, China great. Fk North Korea. Taiwan girls are the best.

[Chicken Genius @chickengenius]

Trump had a lot of support when he campaigned for fair trade. I mean, you tax me 20% I tax you 20%. Fair game.

He has since lost plenty of support from the people that donated to him. Fixing trade deficit is so fking stupid. Watch this clip of Jeffery Sachs that went viral.

[Ganet grant @tocgg]

@sama Excellent. Memory and sandbox were two of my biggest pain using ChatGPT, as my monthly token usage approaching 2m. Hope to feel the improvement soon

[Ganet grant @tocgg]

but none of these smartest AI know a call spread has 100 shares!

[Ganet grant @tocgg]

I asked DeepSeek/ChatGPT/Grok3 to solve a trading math problem, to see if "reasoning" actually makes AI better.
DS scored perfect!
Grok3 w/ "think" on got it wrong
ChatGPT4o / OpenAI o1 got it correct
OpenAI o3 mini / o3 mini high got it wrong!

here is the problem to solve:

[Ganet grant @tocgg]

correct answer is $1.9, incorrect answer is 1.61

[Ganet grant @tocgg]

@financeguy74 nice theory but highly unlikely...after using grok3 and deepseek r1 to break down “Satoshi Nakamoto” with Crypto, Linguistics, Hanzi, analyzing bitcoin original white paper with Jack's writing style, and having met Jack and often sharing same bus many times。。。

[Andrej Karpathy @karpathy]

New 3h31m video on YouTube:
"Deep Dive into LLMs like ChatGPT"

This is a general audience deep dive into the Large Language Model (LLM) AI technology that powers ChatGPT and related products. It is covers the full training stack of how the models are developed, along with mental models of how to think about their "psychology", and how to get the best use them in practical applications.

We cover all the major stages:
1. pretraining: data, tokenization, Transformer neural network I/O and internals, inference, GPT-2 training example, Llama 3.1 base inference examples
2. supervised finetuning: conversations data, "LLM Psychology": hallucinations, tool use, knowledge/working memory, knowledge of self, models need tokens to think, spelling, jagged intelligence
3. reinforcement learning: practice makes perfect, DeepSeek-R1, AlphaGo, RLHF.

I designed this video for the "general audience" track of my videos, which I believe are accessible to most people, even without technical background. It should give you an intuitive understanding of the full training pipeline of LLMs like ChatGPT, with many examples along the way, and maybe some ways of thinking around current capabilities, where we are, and what's coming.

(Also, I have one "Intro to LLMs" video already from ~year ago, but that is just a re-recording of a random talk, so I wanted to loop around and do a lot more comprehensive version of this topic. They can still be combined, as the talk goes a lot deeper into other topics, e.g. LLM OS and LLM Security)

Hope it's fun & useful!
https://t.co/75mXcUBI8L

[MyShell.AI @myshell_ai]

DeepSeek @deepseek_ai again proves open-source AI's vast potential and impact.
The future isn’t just chatbots—it’s multi-modal AI Agents.

Everyone can build their own AI Agent with DeepSeek on MyShell’s ShellAgent today. We've put together a step-by-step guide to help you make a children's book AI Agent. 🧵⬇️

Full tutorial: https://t.co/YFOS4NyQvx

[Ganet grant @tocgg]

@tmel0211 Respect

[CZ 🔶 BNB @cz_binance]

4 years ago today. Some things never change. 😂