Design/ Code/ Photography / Stylish / LGBTQ+ / We love us.
Joined June 2025
- Tweets152
- Following35
- Followers5
- Likes365
leedope retweeted
A thread of early experiments with Claude Fable 5.2:
A 3D model of a cell, built entirely from code one organelle at a time. Built by @kevin_t_ngo.
leedope retweeted
Introducing GPT‑6 Cyber: a major advance in security research that pairs stronger capabilities with better model alignment.
The model powered the Linux kernel research we announced yesterday, and security teams at companies like Mozilla, Palo Alto Networks, Calif, and Cisco—alongside our own at OpenAI—are already putting it to work.
GPT-6 Cyber is available to select Daybreak Red customers and we are expanding access over time.
openai.com/index/introducing…
Actually, it’s partly because of the rapid pace of modern agricultural technology in China. It just reduced the amount of labor required.
In China, more than one in five workers is still employed in agriculture.
China makes and exports more goods than any other country. So, perhaps unsurprisingly, when we think about China’s economy, the first thing that comes to mind is often manufacturing — increasingly at the technological frontier in electric vehicles, batteries, and solar panels.
China’s position in manufacturing follows decades of economic transformation. The chart shows one aspect of this: the decline in agricultural employment.
At the turn of the 21st century, around half of all Chinese workers were employed in agriculture. One generation later, that share had more than halved, reaching 22% in 2025.
It's striking how far the share of workers in agriculture has fallen, and just as striking how high it remains. Agriculture employs more than a fifth of China’s workers, while accounting for less than 7% of its GDP.
(written by @EOrtizOspina)
9
leedope retweeted
This paper answers:
why anything is beautiful
why jokes are funny
why we get curious
why you get bored
why babies explore
why music has hooks
For centuries, we thought these were deeply mystical human traits. Soul-level mysteries that machines could never touch.
But, the paper proves that the human brain operates as a relentless compression engine. It is constantly trying to predict the world and compress it into the smallest possible file size.
Here is how it explains your entire life:
If something is completely predictable, your brain has already compressed it. There is no new data to extract. The result? You feel bored.
If something is pure randomness, your brain cannot compress it at all. The result? You feel annoyed or tune it out.
But then there is the sweet spot.
When you encounter something that seems complex, but suddenly your brain finds the hidden pattern and connects the dots... your brain releases dopamine.
That sudden, glorious rush of successful data compression? That is what you experience as "curiosity."
When a comedian tells a story that makes no sense, and the punchline suddenly makes all the pieces fit together instantly? That rapid neural compression is what you experience as "humor."
When you look at a painting and notice an underlying symmetry that ties the chaos together? That is "beauty."
Art, science, music, and jokes are not magic.
They are just highly optimized data sets designed to trigger compression progress in your neural network.
Here is the most unsettling part.
This paper wasn't written by a philosopher. It was written by an AI pioneer to figure out how to program intrinsic motivation into artificial agents.
He wasn't trying to explain humans. He was writing the blueprint for machines.
We always assumed AI could never appreciate art, laugh at a joke, or feel the spark of curiosity.
But if our deepest subjective experiences are just a mathematical reward for maximizing compression progress...
leedope retweeted
It means entire countries and cultures will go extinct forever
fertility rates are so abstract that people struggle to understand the scale of the catastrophe we’re experiencing.
take 1.14 children per woman (italy and japan)
if that continues, each generation is roughly half the size of the one before it
start with 100 people in a generation, the next only has about 54, then 29, then 16.
in three generations, roughly a century, you have a generation 84% smaller than the one you started with.
for taiwan that would be 97.4%!
i don't think people really understand what that means
leedope retweeted
Send 𝕏 links to your friends so that they know what’s really going on
leedope retweeted
OpenAI CEO Sam Altman: "I don't even know a dramatic enough word to use… models can find bugs that are outside of the normal threat model of what we think a hacker can do… hardware-level, insane stuff… I no longer think we're gonna be able to just patch all the bugs."
It's over.
leedope retweeted
GPT-6 Astra Pro in ChatGPT got an upgrade with faster responses and improved use of plugins and tools.
And now, multiple agents can tackle different parts of a task in parallel. You can follow each agent’s progress and open completed tasks to review their work.
Now available to Pro users in the Chat tab.
leedope retweeted
Here’s the complete YouTube playlist on the same topic for anyone who wants to refer to it.
youtube.com/playlist?list=PL…
leedope retweeted
🚨 Skip Netflix tonight.
Watch this 2h 34min Stanford class instead.
It’s one of the clearest, most complete, and brutally honest explanations of how ChatGPT and Claude actually work.
No fluff. No marketing.
Just the fundamentals:
→ BPE Tokenization — how text becomes numbers
→ Transformers — how attention lets tokens communicate
→ Training — how NLL loss teaches the model
→ Next-token prediction — the core job of a language model
→ Generation — how responses are built token by token
Whether you’ve never written a line of AI code or you’re already building Agents, this class connects the dots.
The big-picture understanding that takes years to develop?
You can get a surprisingly complete version of it in one sitting.
Free up 2.5 hours.
This might be one of the most valuable AI classes you watch this month.
leedope retweeted
This quote from Steve Jobs lives rent free in my head...
"People who are genuinely great at something don't compete. They don't care about being compared to others because they're so deep in their own lane that comparison would be a waste of time. They just do their thing."
leedope retweeted
没想到孙哥表达欲这么强，嘎嘎输出观点，而且不知道是不是剪辑的原因，完全没看出累或者不想说的感觉，结尾整理了20点年轻人建议。
孙宇晨这期访谈里有四个观点，我个人觉得挺有意思：全面 AI 化、不要过早买房买车结婚、人生要「逐水草而居」、先自由后财富。
他认为，无论有多少钱，都应该先把学习、工作、生活全面 AI 化，一部手机就能开始；
年轻人不要过早被房子、车子和婚姻绑定，否则 OpenAI 突然给你一个去旧金山工作的机会，可能因为家庭和地域限制而错过；
所谓「逐水草而居」，就是哪里能发挥最大的价值、获得更高的回报，就去哪里。而他对财富自由的理解恰好与此呼应：
不是先有钱才能自由，而是先获得认知和选择上的自由，财富才可能成为伴生品。 自由让人更容易抓住机会，财富又反过来增强自由。
1. 零成本把自己全面 AI 化｜09:54
没电脑就用手机，没钱就用免费模型。
把学习、健身、日程、工作流程等信息整理成 Markdown 文档，让 AI 参与管理。不要被单一模型绑定，可以通过 API 接入不同模型。
2. 先建立三套底层认知｜12:07
孙学：升级对金钱、财富自由和世界的理解。
薄肌理论：通过力量训练改善身体管理。
黄毛理论：不被外界评价左右，坚持自己经过判断认可的方向。
他认为这三件事都可以低成本开始。
3. 20 岁一无所有，先积累 2～4 个月个人数据｜17:17
收集健康、学习、能力、生活习惯和当前处境等信息，全部数字化。
交给 Claude、OpenAI、国产模型等多个 AI 分析，寻找职业方向、商业机会及赚钱路径。
不要只问一个模型，交叉比较结果。
4. 持续追踪科技前沿｜19:00
每天关注技术变化，可以自己看 X，也可以让 AI 自动抓取信息。
从中选择最感兴趣、最擅长的一个领域深入研究。
不要试图成为所有领域的专家。
5. 像训练 AI 一样训练自己｜20:31
根据每天收集的数据持续调整方向、行动和习惯。
不追求一次决策正确，而是不断获取反馈、更新策略。
6. 不会就问 AI，进一步学会 API｜23:39
从微信聊天转向与 AI 对话，再从单一聊天工具升级到 API。
文件保存在本地，通过不同模型读取和分析。
不懂 Markdown、API 怎么操作，先问 AI。
7. 年轻人可以提前参与前沿科研｜28:44
他认为 AI 正在降低数学研究的门槛。
过去需要依赖名校、导师和多年学术训练才能接触的研究思路，现在高中生也可能借助 AI 提前学习和探索。
不要被年龄、学历和传统学术路径限制。
8. 全球寻找商机，不要只盯着本地｜34:00
把尼日利亚、土耳其、阿根廷等国家的消费者都视为潜在客户。
不懂当地语言，可以学习，也可以借助 AI 翻译。
寻找全球市场的信息差、价格差和未被满足的需求。
9. 不要让 4 亿中国年轻人只在国内互卷｜37:30
中国拥有强大的工业产能，即使做传统贸易，也有机会把中国商品卖向世界。
他的思路是：为全球约 80 亿人提供产品和服务，而不是只在国内争夺存量市场。
10. 买房、买车、选专业前，先看 100 个选项｜39:50
想买房，先看 100 套房；想买车，先研究 100 辆车。
不知道大学选什么专业，就让 AI 分析 100 个专业未来 10 年的就业前景。
他认为年轻人最容易犯的错误，是信息不足就仓促决策。
11. AI 就是《红警》《星际争霸》里的开全图｜43:30
打游戏时，先侦察地图、确定资源和对手位置，再决定造什么兵。
人生同理。
先用 AI 获取足够的信息，再决定职业、投资和商业方向。
12. 重点关注六类 AI 机会｜47:00
数学及数据驱动科研、天文学、生物制药、基因分析与健康管理、保险金融、Crypto 与 AI 支付。
另外，他特别提到 AI 数据清洗和数据标注，不要小看这些基础业务。
13. 20～40 岁，尽早进行力量训练｜53:30
他把肌肉看作身体的长期储备，认为越早开始身体管理越有价值。
强调身体健康、肌肉储备和心肺能力，而不是单纯为了外形。
这是他的健康管理观点，不能理解为健身能够保证长寿或避免疾病。
14. 人生要「逐水草而居」｜01:22:40
哪里能发挥最大价值，就去哪里。
不必永远留在同一个城市、国家或行业。
他把收入和市场需求作为评估自身价值的重要参考指标。
15. 年轻时尽量不要过早被房、车、婚姻锁定｜01:26:50
孙宇晨举例：一个人如果已经在家乡买房、结婚，即使突然拿到硅谷公司的工作机会，也可能因为家庭和地理位置限制而无法接受。
他的建议是尽量保留职业与地理上的流动性。
16. 看到新技术，先尝试、先学习｜01:59:35
他回忆自己早期接触 iPhone、特斯拉时的经历。
认为不一定要进入一家科技公司工作，也可以通过学习、使用产品或投资等方式参与技术变革。
即使自己的观点暂时不被多数人认可，只要逻辑判断充分，也不必立即放弃。
17. 大胆做自媒体，建立个人 IP｜02:02:19
他认为现在是个人创作者的机会期。
以 Joe Rogan 为例，一个人长期经营播客，也可能建立能够与大型传统媒体竞争的影响力。
内容创作不一定需要传统媒体机构的资源。
18. 先升级认知，再追求财富｜02:42:50、03:02:00
他认为财富自由的顺序不是先有钱再自由，而是先获得某种自由状态，财富才可能随之出现。
还建议用假设自己突然获得 3000 万元的方式，思考能否管理这笔财富。
如果连如何使用和保存财富都没有清晰认知，就应该先学习。
19. 感情问题不要试图用钱解决｜03:05:50
他认为金钱更像关系的放大器，而不是爱情本身。
有钱不代表能让原本不爱你的人爱上你。
在感情中，即使事业上很成功的人也可能「当局者迷」。
20. 少消费，多创造；先体验，再判断｜03:17:45、04:11:50
他以上太空为例：没有亲自体验，就很难知道一件事情真正的价值。
不必做任何事之前都先确定终极意义。
相较于单纯消费和享受，他更倾向于把时间投入创造、探索和实际行动。
