@notoptimal5
Account based inIndia
About this account
- Account based in
- India
- Connected via
- Web
Account-level information from X, not a live location or the device used for a specific post.
Joined October 2025
- Tweets2.4K
- Following30
- Followers6
- Likes719
optimal retweeted
Indian leftists are legit some of the most embarrassing clowns on the planet.
The entire Indonesian + Chinese economies are run by their small Chinese minorities.
Open the list of the Top 10 richest Malaysians/Indonesians and see how many Muslims you can find.
Despite having crazy high reservations, Malays in Malaysia continue to remain the poorest group in the country.
This quoted post is unavailable.
optimal retweeted
Fun Fact of the leak is not Gpt -5.5 that gets my attention
I am more interested in all these unknown names pop up in the codex app
Read there discription 😭
optimal retweeted
Color analysis
Prompt: Using this portrait, create a diagram-first personal color analysis. Show which clothing colors suit the subject through visual comparison. Keep text minimal and avoid paragraphs.
optimal retweeted
To get ahead in life, you must imagine the best version of yourself raping the worst version of yourself. And you must masturbate to it. This is the only way.
optimal retweeted
Underrated life skill: Learn to introduce yourself well. Look people in the eye, speak clearly, say something you’re excited about, and give a specific reason you’re glad to meet them. The first 10 seconds set the tone for everything after. Such a simple way to stand out.
optimal retweeted
Women are like cats.
They need to come to you on their own.
They're not going to come to you if you chase them.
optimal retweeted
Played Dhol at Pahurnar, Dantewada, CG
In 2018, Sarpanch of this village was murdered by Naxals for supporting Infra Projects. 🇮🇳 was hoisted first time in 2021.
With end of Naxal Terror, was amazed to see Infra progress in this ‘no longer remote’ village.
Video Source - me
optimal retweeted
Carry yourself like Royalty.
You have to Crown yourself. A lot of Men wait for someone else to anoint them, approve them, push them into action. You have to anoint yourself, no one else is going to come along and do it for you.
optimal retweeted
Stop asking ‘how do I meet high-level people’
Just show up to the same 3 spots for 3 months straight
High end gym like Equinox
Hotel bar Wednesday evening
Members club Sunday
Zero icebreakers needed when they already recognize you
GPT-Image-2 on Heavy Thinking is now the new SoTA on the Image Reasoning Generation Benchmark (IRGB) with a score of 26.7%. Still a ways to go, but good progress!
Many of the failures are near-misses!
Check the outputs here: pellaml.github.io/iumb/#irgb
What makes ChatGPT Images 2.0 a state-of-the-art image generation model?
Researchers behind the model explain. A thread:
Thinking & Intelligence in ChatGPT Images 2.0, demonstrated by @ayaanzhaque
optimal retweeted
We want you to have a lot of AI!
I don't know what they are doing over there, but Codex will continue to be available both in the FREE and PLUS ($20) plans. We have the compute and efficient models to support it. For important changes, we will engage with the community well ahead of making them.
Transparency and trust are two principles we will not break, even if it means momentarily earning less. A reminder that you vote with your subscription for the values you want to see in this world.
optimal retweeted
bromantane is fucking great for staying focused when working long hours. especially when paired with modafinil and discipline
but oh boy. this shit on vacation. makes it 100x better
most versatile compound of all time
ready to die on that hill
bromantane is the best overall life enhancement compound
everything is just better on it, and nothing beats it. i keep coming back to it.
50mg sublingual: focus enhanced, colors look more vivid, i stay more present and just seem to enjoy everything more
i guess that's what happens when you upregulate dopamine production instead of nuking your receptors
soviet perfection.
kimerachems.co/product/broma…
optimal retweeted
One lesson from Sharad Pawar is longevity comes from high stress environments and long work hours.
The US presidency is the most stressful job. Look at how long those presidents lived.
मी १९६७ मध्ये लढविलेल्या पहिल्या निवडणुकीपासून बारामती मतदारसंघातील माझ्या मतदानाचा हक्क अखंडपणे बजावला. आता २०२६ मध्ये बारामती विधानसभा मतदारसंघात महाराष्ट्राचे एक लाडके व लोकप्रिय नेते स्व. अजित पवार यांच्या दुर्दैवी, अपघाती निधनाच्या शोकात्म पार्श्वभूमीवर ही पोटनिवडणुक होत आहे.
या पोटनिवडणुकीसाठी बारामती येथे येऊन आतापर्यंतच्या प्रघातानुसार मतदान करण्याचे मी ठरवले होते. दुर्दैवाने माझी प्रकृती बिघडल्याने मला मुंबईत रुग्णालयात दाखल व्हावे लागले. डॉक्टरांनी मला उपचार आणि सक्तीची विश्रांती या कारणास्तव मुंबईबाहेर जाण्यास मनाई केली आहे. त्यामुळे अत्यंत नाईलाजास्तव मतदानाचा हक्क बजावण्यापासून मला वंचित रहावे लागत आहे, हे मी अत्यंत व्यथित आणि खेदपूर्ण अंत:करणाने सांगू इच्छितो. माझ्यावर अथक प्रेम व विश्वास असलेले बारामतीचे नागरिक, मतदार व कार्यकर्ते मला समजुन घेतील याची मी खात्री बाळगतो.
आमच्या पक्षाने (राष्ट्रवादी काँग्रेस पक्ष – शरदचंद्र पवार) या पोटनिवडणुकीत श्रीमती सुनेत्रा अजित पवार यांना पाठिंबा जाहिर केला आहे. स्व. अजित पवार यांना ही एकप्रकारे आमच्या पक्षातर्फे श्रद्धांजलीच आहे. माझे बारामतीमधील सर्व मतदारांना आवाहन आहे की, त्यांनी श्रीमती सुनेत्रा अजित पवार यांना मतदान करून स्व. अजित पवार यांना श्रद्धांजली अर्पण करावी.
बारामतीमधील मतदार यांनी माझ्यावर प्रदीर्घ विश्वास ठेवलेला आहे. ते या आवाहनास प्रतिसाद देतील, असा मला विश्वास आहे.
मतदानासाठी प्रत्यक्ष न येता निवडणुक आयोगाच्या विविध पर्यायांचा देखील मी तपास केला व विचारणा केली. परंतु ते पर्याय उपलब्ध नसल्याचे राज्याच्या निवडणुक अधिकाऱ्यांकडून सांगण्यात आले. अखेर अतिशय व्यथित व जड अंत:करणाने मला मतदानापासून दूर राहण्याचा निर्णय करावा लागला. या पोटनिवडणुकीत श्रीमती सुनेत्रा अजित पवार यांना मतदान करावे, असे मी बारामतीकरांना पुन्हा एकदा आवाहन करतो.
धन्यवाद !
optimal retweeted
全网都在吹的LeCun新论文，90%的解读都是错的。
他们说生成式AI是死路，说过去三年花的几百亿全白费了，说15M参数的小模型就能吊打万亿大模型。
这些全是营销号的夸张，
我觉得这篇论文的真正分量比他们吹的还要重。
Yann LeCun团队这次解决了JEPA困扰了好几年的表征坍缩问题。
以前的世界模型，学着学着就会把狗车人都压成一模一样的向量，什么都学不到。
这次他们只加了一个极其优雅的数学正则化器SIGReg，
没有复杂的trick 和六个超参数要调，训练稳得离谱。
单张GPU几个小时就能训完，在机器人控制任务上，规划速度比巨型世界模型快48倍，成功率还更高。
最厉害的是它的隐空间里天然就编码了物理规律。
不用教，它自己就知道物体不能瞬移，知道速度和位置的关系。
能瞬间检测出物理上不可能发生的事。
这不是啥范式革命，也不会让GPT和Claude明天就死掉。
语言和创意生成，依然是自回归大模型的天下。
但它打开了一扇全新的门，
原来懂物理不需要万亿参数，不需要云端超算。
原来世界模型可以小到跑在机器人的本地芯片上。
过去三年，整个行业都在一条路上狂奔，堆参数，堆算力，堆数据。
所有人都以为只要足够大，就能懂世界。
现在我们终于知道，还有另一条路，一条更高效，更优雅，更接近真实世界运行方式的路。
生成式AI不会死，
但未来的智能体不会是只会聊天的大模型，
它会是一个懂物理的小世界模型，加一个大语言接口，
这才是这篇论文真正的意义所在吧 hhh
这可能是今年AI圈最清醒的一条推文，
Yann LeCun 是当今 AI 领域最有影响力的科学家之一，深度学习三大教父之一，2018 年图灵奖获得者，
他直接怼了Anthropic CEO Dario的著名言论，
Dario说未来一到五年，一半的科技法律咨询金融岗位会被彻底干掉。
LeCun说你错了，而且不仅Dario错了，所有的AI实验室CEO 和一众大佬包括Sam Altman，Yoshua Bengio，Geoff Hinton，也包括他自己。
他说在技术革命如何影响就业这个问题上，全都是外行，别听我们瞎BB。
去听真正研究了几十年这个问题的劳动经济学家。
最狠的其实是这一点，他没跟Dario争论AI到底有多强，直接掀了桌子说，我们这群人，根本就没有资格讨论这个话题。
AI研究者只懂技术能干什么，他们不懂企业的组织流程，不懂法律合规的障碍，不懂市场的供需关系，更不懂人类社会复杂的运行逻辑。
那些被LeCun点名的经济学家，研究了过去两百年所有的技术革命，他们得出的结论是整份工作不会消失，工作里的任务会被重构。
ATM发明的时候，所有人都以为银行柜员要失业，结果柜员的数量反而增加了，因为他们从数钱的人，变成了卖理财做服务的人。
AI现在也一样，它会干掉大量重复的可预测的任务，但同时也会创造出大量新的任务。
短期的取代效应确实存在，尤其是入门级的白领岗位，但长期的生产力效应和新岗位创造效应，往往要强得多。
还有一个所有人都忽略的事实，AI能干一件事，和公司真的会用AI把人换掉。
这中间至少差两到五年，数据管道，组织架构，员工培训，法律风险。
每一个都是巨大的障碍。
我觉得 LeCun其实是在反对AI圈泛滥的末日营销，很多AI大佬喜欢用失业恐慌来博眼球，要么为了融资，要么为了呼吁监管，要么为了显得自己的产品很厉害。
但这种恐慌会误导公众，会误导政策，所以咱们也别再被那些五年后一半人失业的言论吓住了，真实的世界从来都不是非黑即白的，它是缓慢的，混乱的，充满了各种意外和机会的。
也别总担心被AI取代，有那个时间多想想怎么用AI让自己变得更值钱😁
Dario被 fox 采访的内容我做了中英双语字幕供大家参考。
optimal retweeted
every few months someone declares LLMs a dead end and every few months LLMs get dramatically better at things they supposedly cant do. but the LeWM result is worth paying attention to for a different reason: 15M parameters, single GPU, 48x faster planning. that efficiency per parameter ratio is what matters for real deployment. the future of AI inference isnt one architecture ruling everything, its routing different tasks to different architectures optimized for each. physics reasoning on a tiny world model, language on an LLM, vision on a specialized encoder. the serving layer that routes between them is where the real value sits @HowToAI_
Yann LeCun was right the entire time. And generative AI might be a dead end.
For the last three years, the entire industry has been obsessed with building bigger LLMs. Trillions of parameters. Billions in compute.
The theory was simple: if you make the model big enough, it will eventually understand how the world works.
Yann LeCun said that was stupid.
He argued that generative AI is fundamentally inefficient.
When an AI predicts the next word, or generates the next pixel, it wastes massive amounts of compute on surface-level details.
It memorizes patterns instead of learning the actual physics of reality.
He proposed a different path: JEPA (Joint-Embedding Predictive Architecture).
Instead of forcing the AI to paint the world pixel by pixel, JEPA forces it to predict abstract concepts. It predicts what happens next in a compressed "thought space."
But for years, JEPA had a fatal flaw.
It suffered from "representation collapse."
Because the AI was allowed to simplify reality, it would cheat. It would simplify everything so much that a dog, a car, and a human all looked identical.
It learned nothing.
To fix it, engineers had to use insanely complex hacks, frozen encoders, and massive compute overheads.
Until today.
Researchers just dropped a paper called "LeWorldModel" (LeWM).
They completely solved the collapse problem.
They replaced the complex engineering hacks with a single, elegant mathematical regularizer.
It forces the AI's internal "thoughts" into a perfect Gaussian distribution.
The AI can no longer cheat. It is forced to understand the physical structure of reality to make its predictions.
The results completely rewrite the economics of AI.
LeWM didn't need a massive, centralized supercomputer.
It has just 15 million parameters.
It trains on a single, standard GPU in a few hours.
Yet it plans 48x faster than massive foundation world models. It intrinsically understands physics. It instantly detects impossible events.
We spent billions trying to force massive server farms to memorize the internet.
Now, a tiny model running locally on a single graphics card is actually learning how the real world works.
