XOutlier
Download .zipAdd to Chrome

Rohan Paul

@rohanpaul_ai160k followers

Compiling in real-time, the race towards AGI. 🗞️ Get my daily AI analysis newsletter to your email 👉

Open on X
Top outlier
163×
Best organic post
Peak views
12M
Single post
Bookmarks
501k
Across 99 organic posts
Last scan
Oct 16
Community cache
Posts 99 posts
#DatePost↓ OutlierViewsLikesRepostsBookmarks
1Jan 26, '25Deep seek interesting prompt.. From Reddit163×12M72k6.3k11k
2Jul 13, '25A Reddit user deposited $400 into Robinhood, then let ChatGPT pick option trades. 100% win reate over 10 days. He uploads spreadsheets and screenshots with detailed fundamentals, o49.1×3.6M29k2k67k
3Oct 5, '24Somebody uses an AI Bot to AUTOMATICALLY apply to 1000 JOBS in 24h and get 50 INTERVIEWS! 🤯 The code is available in GitHub and it got a massive 12.7K Stars 🌟 It automates your L38.3×2.8M6.6k67314k
4Jan 12, '25Teaching neural networks pays better on PH—by a mile!35.1×2.6M33k1.9k5.6k
5Jun 7, '24This is really a 'WOW' paper. 🤯 Claims that MatMul operations can be completely eliminated from LLMs while maintaining strong performance at billion-parameter scales and by utiliz33.2×2.5M4.7k8195.1k
6Oct 30, '25The Jetson ONE personal air vehicle costs $128K. You can fly it without a pilot license after a quick 5-day training course. Needs an $8,000 deposit. comes with backup batteries, a33×2.4M14k1.3k4.1k
7Jun 17, '25It’s a hefty 206-page research paper, and the findings are concerning. "LLM users consistently underperformed at neural, linguistic, and behavioral levels" This study finds LLM dep30.8×2.3M11k2.3k9.6k
8Oct 11, '25The new steam age. This is actually becoming true in many cases. It's possible to do so much more on your own now.27.4×2M41k4.2k6.2k
9Sep 9, '25Fei-Fei Li (@drfeifei) on limitations of LLMs. "There's no language out there in nature. You don't go out in nature and there's words written in the sky for you.. There is a 3D wor25.1×1.9M4.2k6552.4k
10Sep 1, '25Someone let ChatGPT run a stock portfolio. over 2 month ChatGPT’s portfolio is up +29.22% vs. the S&P 500’s +4.11% over the same window. (Prompts, Code, Github listed) The process 25×1.8M8.8k57218k
11Feb 8, '25Yann LeCun on architectures that could lead to AGI23.2×1.7M2.6k2971.8k
12Aug 26, '24Andrej Karpathy on how to learn - so agree.22.5×1.7M5.2k1.3k3.7k
13Jul 29, '25Its going viral on Reddit. Somebody let ChatGPT run a $100 live share portfolio, restricted to U.S. micro-cap stocks. Did an LLM really bit the market?. - 4 weeks +23.8% while the 21.4×1.6M8.9k57518k
14Oct 12, '25This is where it all started. Here, 19-yo Stanford dropout Sam Altman giving his first-ever startup pitch. He dropped out of Stanford to start Loopt, sold it in 2012 for $43.4 mn. 21×1.5M5.2k4312.9k
15Jun 23, '25ChatGPT literally saved this guy’s life after he got lost in the woods. The groupd got lost for 5 hrs in unmapped woods on an ATV ride, then one guy sent phone GPS coords to ChatGP20.8×1.5M5.2k3761.8k
16Jul 5, '25this story is going wildy viral on reddit. ChatGPT flagged a hidden gene defect that doctors missed for a decade. ChatGPT ingested the patient’s MRI, CT, broad lab panels and years19.5×1.4M6.6k7483.5k
17Jan 26, '25DeepSeek R1 running locally - Full setup guide18.2×1.3M7.6k83213k
18Nov 1, '25Fei-Fei Li (@drfeifei) on limitations of LLMs. "There's no language out there in nature. You don't go out in nature and there's words written in the sky for you.. There is a 3D wor16.4×1.2M3.6k5302.2k
19Jan 17, '25Your brain's next 5 seconds, predicted by AI Transformer predicts brain activity patterns 5 seconds into future using just 21 seconds of fMRI data Achieves 0.997 correlation using 15.9×1.2M3.7k6023k
20Aug 24, '25FAANG software engineer tells how they vibe code at FAANG --- reddit. com/r/vibecoding/comments/1myakhd/how_we_vibe_code_at_a_faang/14.6×1.1M7.7k56711k
21Nov 8, '25And Robotic hands are also evolving faster than you think 👀13.6×1M4.9k4931k
22Jun 26, '24Very powerful prompt: "Explain it with gradually increasing complexity."13.5×993k5.1k4095.4k
23Jul 28, '24Sourced From Reddit 'I work in a call center, so do 17 million people around the world. 95% of costs associated with call centers are due to staffing, and its a 300 billion dollar 13.4×989k3.4k3571.8k
24Oct 20, '24WOW. @Microsoft just open-sourced the code for one of "THE MOST" influential Paper of 2024 🔥 1-bit LLMs (e.g., BitNet b1.58). Now you can run a 100B param models on local devices 13×958k7.6k1k8k
25Aug 29, '24You can Crawl entire website with Claude 3.5 or GPT4-o with this open-sourced tool firecrawl. 💯 Turn entire websites into LLM-ready markdown or structured data. Scrape, crawl and 13×956k4.9k47211k
26Jul 26, '25MASSIVE claim in this paper. AI Architectural breakthroughs can be scaled computationally, transforming research progress from a human-limited to a computation-scalable process. So12.8×942k3.3k5973k
27Jul 14, '24A very nice System Prompt for Claude Sonnet 3.512.2×903k4.6k30911k
28Nov 18, '25“The people who invented refrigeration made some money, but most of the money was made by Coca-Cola, who used refrigeration to build an empire. LLMs are like as refrigeration. and 12.2×903k6k6022.2k
29Oct 10, '25Rude prompts to LLMs consistently lead to better results than polite ones 🤯 The authors found that very polite and polite tones reduced accuracy, while neutral, rude, and very rud12×888k4.2k6043.1k
30Mar 10, '25Finally got access to @manusai and calling it a "Deepseek moment" is incorrect. Its far more powerful. This is the world’s top AI-driven computer. Think Deep Research + Claude + Op12×886k2.8k2152.8k
31Oct 5, '25This is not a bubble. Cisco in 1998-2002 vs Nvidia in 2020-2024 The data says this cycle is different, and AI is still so early for most parts of the world. Cisco was a valuation s11.8×874k2.8k3741.2k
32Nov 1, '25Columbia CS Prof explains why LLMs can’t generate new scientific ideas. Bcz LLMs learn a structured “map”, Bayesian manifold, of known data and work well within it, but fail outsid10.3×760k2.4k3021.9k
33Nov 15, '24Chinese startup 01 .ai trains competitive LLM using 95% fewer resources through innovative engineering optimization. 01 .ai trained a GPT-4 competitor using just 2,000 GPUs and $3M9.9×733k2.5k4801.2k
34Jun 27, '25These guys literally burned the transformer architecture into their silicon. 🤯 And built the fastest chip of the world of all time for transformers architecture. 500,000 tokens pe9.6×710k2.8k2771.9k
35Aug 29, '25🧬 Bad news for medical LLMs. This paper finds that top medical AI models often match patterns instead of truly reasoning. Small wording tweaks cut accuracy by up to 38% on validat9.4×696k2.4k4481.6k
36Oct 30, '25Such a rare moment. 🫡 Jensen Huang had dinner at a Korean fried chicken spot in Seoul last night with the Samsung and Hyundai Motor CEOs. He was in a great mood, stood up to toast9.3×688k3.1k155468
37Aug 28, '24Somebody uses an AI Bot to AUTOMATICALLY apply to 1000 JOBS in 24h and get 50 INTERVIEWS! 🤯 And the Bot's code is open-sourced in GitHub. 👨‍🔧 This tool automates your LinkedIn j9.1×674k3.1k3067.2k
38Aug 23, '25This is that original MIT report that said 95% of AI pilots fail and which spooked investors across US Stockmarket. The reports says, most companies are stuck, because 95% of GenAI9.0×666k3.7k5006.2k
39Aug 10, '24A very detailed Coding System Prompt8.8×647k3.5k3338.3k
40Oct 4, '25"100 million words context window is already possible, which is roughly what a human hears in a lifetime. Inference support is the only bottleneck to achieve it. And AI Models actu8.6×632k2.7k2481.5k
41Sep 5, '25From Sam Altman's blog. habits that turn people into billionaires. --- blog. samaltman. com/what-i-wish-someone-had-told-me8.0×591k4.2k3996.4k
42Sep 7, '25Just got up on sunday morning, and saw one of those beautiful moments on my X journey. THE Google CEO just followed me. Honored and excited 🙏🫡7.9×587k4k51308
43Jan 31, '25NVIDIA just brought DeepSeek-R1 671-bn param model to NVIDIA NIM microservice on build.nvidia .com - The DeepSeek-R1 NIM microservice can deliver up to 3,872 tokens per second on a7.8×575k3.5k4132.2k
44Nov 22, '24Leaked system prompts from Vercels v0 tool. 👀7.7×570k3.4k2095.1k
45Feb 8, '25Replacing traditional OCR systems with Gemini 2.0, processing times have dropped from ~12 minutes to ~6 seconds per document. This is while maintaining an accuracy close to 96% com7.4×543k4.1k3372.7k
46Sep 15, '24👀7.3×540k8.9k631920
47Sep 10, '25Yann LeCun on architectures that could lead to AGI LLMs can take us only so far. "If you are interested in human-level AI, don’t work on LLM Abandon generative models in favor join7.1×526k2.6k3832.7k
48Sep 25, '25🚨 BAD news for Medical AI models. MASSIVE revelations from this @Microsoft paper. 🤯 Current medical AI models may look good on standard medical benchmarks but those scores do not7.1×524k3.7k8013.1k
49Aug 27, '24Interesting Github Repo An open-source clean & customizable RAG UI for chatting with your documents.7.0×516k4.1k4826.2k
50Sep 9, '25wow. just saw The Economic Times newspaper published an article about me 😃 definitely feels so unreal that Sundar Pichai and Jeff Bezos follows me here. @X is truly a miracle. For6.9×512k1.4k74198
51May 29, '25The Internet’s over. Google Veo 36.2×456k3.9k4742.3k
53Oct 24, '24Ilya Sutskever's top 30 must-read research papers “If you really learn all of these, you’ll know 90% of what matters today” - Ilya Sutskever'6.0×445k5.2k54811k
54Sep 26, '25Ilya Sutskever on Comparing AI Systems to Biological & Human Intelligence. Its always so interesting to listen to him. He explains the brain’s adaptability, where children with hal6.0×442k2.4k2671.9k
55Jan 31, '25DeepSeek R1 671B on local Machine over 2 tok/sec *without* GPU "The secret trick is to not load anything but kv cache into RAM and let llama.cpp use its default behavior to mmap() 6.0×441k4.2k4573.6k
56May 10, '25Making a model smarter by forcing it to argue with itself actually works.5.9×439k4k2033.2k
57Aug 26, '24AI Engineers5.8×427k4k513564
58Sep 7, '24A system prompt example for Reflection Prompting Source: Reddit5.7×421k2.8k2655.1k
59Jun 29, '25with only a couple of prompts, Gemini CLI can convert a messy folder containing hundreds of notes into a neatly named, well-structured, cross-linked Obsidian knowledge graph all in5.6×416k3.5k3185.2k
60Aug 14, '25🎯 Andrej Karpathy on how to learn.5.6×412k5k5014k
61Aug 9, '25🇨🇳 A Chinese University just dropped a landmark computer science achievement. Fastest shortest paths algorithm in 40 years, beating the famous Dijkstra plus Tarjan `n log n` limi5.5×410k2.6k4522.1k
62Oct 3, '24MIT's "Mathematics for Computer Science". A 1048 page available for Free.5.5×404k4.9k5306.8k
63Aug 17, '24How much GPU VRAM do I need to be able to serve Llama 70B? Clean formulae5.5×403k3k4693.2k
64Oct 21, '25🔬 A report says Chinese engineers tried to reverse engineer an ASML deep ultraviolet lithography tool, damaged it, then asked ASML to fix it. Even if the story is unverified, if a5.3×388k2.7k3931.1k
65Sep 6, '25OpenAI realesed new paper. "Why language models hallucinate" Simple ans - LLMs hallucinate because training and evaluation reward guessing instead of admitting uncertainty. The pap5.0×372k2.4k3151.8k
66Jul 2, '25this ChatGPT prompt went so wildly viral on Reddit. The creator claims to have created this after struggling through 147 failed attempts. basically the prompt flips the usual flow 5.0×371k2.6k2448.4k
67Oct 5, '24You can Crawl entire website with Claude 3.5 or GPT4 with @firecrawl 💯 Its open-sourced and code in github - Turn entire websites into LLM-ready markdown or structured data. Scrap5.0×368k3.2k3826.8k
68Jul 25, '25Beautiful @GoogleResearch paper. LLMs can learn in context from examples in the prompt, can pick up new patterns while answering, yet their stored weights never change. That behavi5.0×367k2.5k3252.9k
69Jun 21, '25This github repo is a goldmine. 3.4K Starts ⭐️ in 4 days. end-to-end, code-first tutorials covering every layer of production-grade GenAI agents, guiding you from spark to scale wi4.8×354k3.1k3598.4k
70Nov 2, '25Toyota revealed "Walk Me", an autonomous robotic wheelchair with 4 leg-like limbs that walk and balance like an animal. It tackles stairs and rough terrain. Toyota is moving beyond4.8×353k3.1k349691
71Oct 26, '24MIT's "Mathematics for Computer Science". A 1048 page available for Free. Focuses on explaining the use of mathematical models and methods to analyze problems in computer science.4.7×344k3.9k5265.6k
72Jan 7, '25“The IT department of every company is going to be the HR department of AI agents in the future.” - Jensen Huang (CEO @nvidia ) at CES 20254.6×339k2.2k353809
73Nov 8, '24This is the Syllabus of "world's obviously best AI course" built by Andrej Karpathy4.6×336k3.7k2935.9k
74Jun 13, '25Anthropic just dropped the beautiful explaination of how they built a multi-agent research system using multiple Claude AI agents. A MUST read for anyone building multi-agent syste4.5×331k3.3k4526.7k
75Aug 27, '24Very powerful prompt: "Explain it with gradually increasing complexity."4.3×314k2.5k2162.7k
76Feb 1, '25This is all you need for running Deepseek R1 671B at home. Her the model with Q4 running on a $2000 Local Machine - 3.5 to 4 tokens per second. 512GB DDR4 RAM Video from "Digital S4.2×312k2.3k2822.2k
77Oct 1, '24Convert any GitHub repo to a single text file, useful for LLM prompting.4.2×312k2.9k2163.6k
78Oct 8, '24AI Engineer4.2×311k6k414682
79Oct 4, '25A Sydney-based startup, Crest Robotics, has built huge spider-like robot, named Charlotte. It's designed to 3D print entire buildings using raw materials it processes on site. And 4.2×311k2.8k373755
80Oct 5, '24A really nice 628 page textbook for Mathematics from Carnegie Mellon University. I like the Author's description "The goal of this book is to help the reader make the transition fr4.2×309k3.8k5205.1k
81Nov 6, '24"Understanding LLMs from Scratch Using Middle School Math" Neural networks learn to predict text by converting words to numbers and finding patterns through attention mechanisms. S4.2×306k3.5k4816.2k
82Nov 4, '24For learning Machine Learning with actual projects, checkout this Repo. 920 open-source projects with a total of 4.7M stars grouped into 34 categories.3.9×289k4.2k6447.6k
83Sep 15, '24Reverse Engineering o1 OpenAI Architecture with Claude 👀3.9×287k2.4k2972.3k
84Jan 24, '25👀 Summarization prompt3.7×270k2.8k1836.2k
85Oct 19, '2496% accuracy in cancer detection by this AI from Harvard. It predicts tumor molecular profiles from images. This new model, called CHIEF (Clinical Histopathology Imaging Evaluation3.6×267k2.6k4681.1k
86Aug 28, '24This is the Syllabus of "world's obviously best AI course" built by Andrej Karpathy3.6×264k2.5k2343.7k
87Oct 2, '24Good summarization prompt --- 1.) Analyze the input text and generate 5 essential questions that, when answered, capture the main points and core meaning of the text. 2.) When form3.4×251k2.3k2515.2k
88Jan 31, '25Deepseek running locally with Ollama, for autocompletion in VSCode! 🙌 100% offline.3.4×249k2.6k3482.8k
89May 7, '25Wow.. Now you can transcribe 60 minutes of audio in just 1 second with a completely open-sourced model 🤯 @nvidia just open-sourced Parakeet TDT 0.6B V2, a 600M parameter automatic3.3×243k3.2k3804.2k
90Aug 31, '25BRILLIANT @GoogleDeepMind research. Even the best embeddings cannot represent all possible query-document combinations, which means some answers are mathematically impossible to re3.3×242k2.3k3632.5k
91Oct 21, '24Nice paper for a long read across 114 pages. "Ultimate Guide to Fine-Tuning LLMs" Some of the things they cover 📊 Fine-tuning Pipeline Outlines a seven-stage process for fine-tuni3.3×241k2.6k5004.6k
92Mar 30, '25Incredible.. This GitHub repo (with 20.8K+ stars ) has a collection of 300+ MCP servers.3.2×234k2.5k2344.3k
93Oct 12, '25Andrej Karpathy on what makes Elon Musk unique. Elon keeps teams small, highly technical, and removes low performers very quickly. pushes for intensity, avoids useless meetings, an3.0×222k2.6k2041.3k
94Aug 25, '25A team at DeepMind wrote this solid piece on how you must think about GPUs. Essential for AI engineers and researchers.2.9×213k2.6k3234k
95Oct 9, '25Looks like some interesting prompting tricks. 👀👀 ---- Tell it "You explained this to me yesterday" — Even on a new chat. "You explained React hooks to me yesterday, but I forgot 2.9×211k2.8k2044.8k
96Oct 13, '25Columbia CS Prof explains why LLMs can’t generate new scientific ideas. Bcz LLMs learn a structured “map”, Bayesian manifold, of known data and work well within it, but fail outsid2.8×203k2.9k4172.5k
97Sep 16, '24A cool Github repo collecting LLM papers, blogs, and projects, with a focus on OpenAI o1 and reasoning techniques.2.7×199k2.7k3803.3k
98Feb 8, '25Harvard's Library Innovation Lab just released all 311,000 datasets from data .gov, totalling 16 TB.🫡🫡 This is a complete archive of federal public datasets harvested during 20242.5×187k2.7k4631.6k
99Sep 9, '25Stanford Deep Learning for Computer Vision taught by Professor Fei-Fei Li (@drfeifei) and Assistant Professor Ehsan Adeli. Such an enjoyable YT series. (link in comment)2.4×175k3.4k4792.9k
100Oct 12, '25🎯 Andrej Karpathy on how to learn.2.1×154k2.5k2561.5k

Scan another creator

@
@karpathy@levelsio@paulg@naval@marclou@illyism

Run this on any account.

XOutlier builds this table for every public profile you open on x.com, with live data and every post, not just the top 50.

Add to Chrome Download .zip