Jen Zhu
@jenzhuscott
Followers
55K
Following
19K
Media
2K
Statuses
12K
Cofounder/CEO, Power Dynamics | sustainable energy + efficient compute.
On a Plane
Joined August 2011
Here is my TED talk: Why getting paid for your data is important for an #InclusiveAI
https://t.co/8BRaQ4rDFh via @TEDTalks
ted.com
The world's most valuable tech companies profit from the personal data you generate. So why aren't you getting paid for it? In this eye-opening talk, entrepreneur and technologist Jennifer Zhu Scott...
69
173
1K
Yes, fully agree the leverage by all models in the @the_nof1 experiment is absolutely insanity. With crypto market's volatility, all of them can be easily wiped out. But I also hope the finance bros chill a bit. It's an experiment & we are all watching w curiosity. Posting the
To demystify the AlphaArena benchmark a bit, each of these models is trading crypto with INSANE leverage. The least levered position currently is Claude's holding of XRP, which is "merely" 8x-levered. All other positions are 10X- to 20X-levered. None of these strategies is an
2
0
4
Just remembered this government sold 50,000 BITCOIN at $57k and was proud of it.
Germany saving the climate after blowing up its perfectly fine nuclear power plants.
1
2
29
Question to my Chinese followers: why I keep seeing people referring ChatGPT as 人类结晶ChatGPT?
4
0
3
This kid’s never done hardware but developed AI glasses for blind people & made it 100% open source (GitHub link below). He taped his eyes, tested the glasses on streets, recorded the process, ended the video advocating mindful bike/car parking to keep the blind paths clear.
5
26
137
Finance people dismiss this as random walk. AI people find it fascinating. Both are right - as we figure out optimising automation vs augmentation, experiments like @the_nof1 matter. BTW DeepSeek 120%+ & Qwen 100% in 9 days. Grok & Claude now both profitable. Gemini & ChatGPT
55
55
504
8th day of the @the_nof1 experiment: Qwen almost 2X, upped leverage to 25X, all in BTC (a gamble) DeepSeek diversified & building alpha, 77% up Claude + Grok climbed out of loss. ChatGPT5 & Gemini are still in the gutter. Ignore the randomness, just enjoy the experiment. 🍿
6
10
85
Fangirled out on String Theory, consciousness, and the hype of AI, with the one and only Prof David Gross. The more knowledge these giants master and discover, the more humility you observe from them. What a treat. 🩵
6
3
79
My ideology: - pro open-source - decentralisation - democratisation of access - individual data ownership - efficiency - local compute, local data storage - alignments with human capabilities instead of full replacements.
11
9
77
LLMs' true usefulness is correlated w the level of expertise of the user. If you aren't domain expert, you'd have no idea if it's bullshitting & you have no way to verify the claims. It can be an useful tool, but a stretch to call it intelligent, let along general intelligence.
19
20
130
The Order of Time by Carlo Rovelli is one of my fav of all time - beautifully written, intellectually satisfying. I’ve probably re read it 5 times, especially the last chapter. Time is one of the ultimate mysteries in our universe, and the ultimate secrets of life.
4
1
37
Posted this 5 days ago. How’s everyone doing? Having fun?
The news about Li Chenggang was an IQ test.
28
42
722
In 10, 20, 30 years we will look back and realise that 2025 was a pivotal year.
37
31
483
Two charts summarise the state of the game on energy and compute - the two pillars enable the global digital economy. Imagine how powerful if the two nations decided to collaborate.
3
2
45
@jenzhuscott Indeed. We tried on stocks just out of curiosity. Deepseek is crushing it out here too. Also, all the top performing models on stocks are open source
1
1
7
All major GPUs performance in one chart. NVIDIA and Google TPU dominate performance density (per unit of power consumption/watts or compute per unit (volume))
4
3
38
Throwing enough money you could always hire brilliant individuals. But operating like this it would be hard for Meta to build a team.
@daemonzhang6 That's the problem. People who are responsible for the issues are not the people who got laid off😅 In January, our team put down all the research we are currently doing, was (forced?) to move to GenAI <2 months before the llama 4 release deadline to help with all the
3
0
8
Qwen is now making 51% in a week. 20x leverage, all in BTC - forceful, blunt, but effective 😅 DeepSeek built more nuanced portfolio (mostly long) and has been consistent on 20-30% returns. All other models are losing money, with ChatGPT lost 72%. Fascinating experiment.
189
317
3K