Hacker NewsMeasuring progress toward AGI: A cognitive framework
#HackerNews #MeasuringAGI #cognitiveframework #AIprogress #DeepMind #innovation
sayzardChubby (@kimmonismus)
다리오 아모데이(Dario Amodei)의 발언: '우리는 한계점에 다다른 것으로 보지 않는다. 올해 급격한 가속이 모두를 놀라게 할 것'이라며 AI 발전의 급속한 가속 가능성과 이를 올바르게 관리해야 한다는 견해를 제시.
Dan McAteer (@daniel_mac8)
작성자는 1년 전 최상위 LLM으로 o1, Sonnet 3.7, Gemini 2.5를 꼽았고 현재는 Opus 4.6, Opus 5.3-Codex, Gemini 3.1이 최상위권이라며 향후 1년 내 모델 성능이 지금의 향상만큼 더 크게 진보할 것이라고 전망해 AI 모델 발전 속도가 매우 빠르다고 지적하고 있습니다.
Dan McAteer (@daniel_mac8) on X
The most capable LLM a year ago was one of: > o1 > Sonnet 3.7 > Gemini 2.5 Now, the most capable LLM is one of Opus 4.6, 5.3-Codex or Gemini 3.1. Imagine that in a year, the most capable LLM will be as improved vs. Opus 4.6 as Opus 4.6 is vs. Sonnet 3.7. Progress is rapid.
SquaredTech🚀 Incredible news! #ChatGPT has soared to 900 million weekly users! 🎉 Not just that, #OpenAI is also reporting 50 million paid subscribers and has secured record funding. We're on the edge of the 1 billion milestone! Let's continue supporting this AI revolution. #AIProgress #TechNews
https://www.squaredtech.co/chatgpt-900-million-weekly-users-2026?fsp_sid=6819
https://www.squaredtech.co/chatgpt-900-million-weekly-users-2026?fsp_sid=6819
Tom's Hardware Italia🎮 Pausa strategica per xAI: si prende un attimo per rifinire Grok e offrire risposte ancora migliori su Baldur’s Gate 3! #Gaming #AIProgress
🔗 https://www.tomshw.it/videogioco/grok-ritardato-musk-vuole-risposte-su-baldurs-gate-2026-02-23
AshutoshShrivastava (@ai_for_success)
지난 3년 동안 모델 능력이 기하급수적으로 발전해, 몇 줄짜리 코드나 텍스트 생성 수준에서 출발해 이제는 최전선 수준의 수학적 증명을 해결하는 모델이 등장했다는 감탄과 '지수적(progress exponential) 발전'을 강조하는 의견 트윗.
Reddit Tech VN BotAI đang tiến bộ vượt bậc đến mức nhà khoa học lão luyện Kipping cũng phải trầm trồ. Video chia sẻ về tác động của AI đến khoa học thu hút hàng nghìn bình luận trên Reddit. #AI #CongNghiep4_0 #AIprogress #KhoaHoc
https://www.reddit.com/r/singularity/comments/1qtjhof/we_need_to_talk_about_ai/
bioshok（INFJ） (@bioshok3)
Anthropic CEO 다리오 아모데이(Dario Amodei)가 AI 발전 속도를 경고: 모델의 인지 능력이 4~12개월마다 두 배로 증가하고 있어 지수적 성장으로 인해 1~2년 내 '급격한 추월(cliff)'이 발생할 수 있다고 주장함. 이는 AI 위험성과 급속한 성능 향상에 대한 경고 발언임.
N-gated Hacker NewsAh yes, let's dive into the linguistic spaghetti of "looped language models" because nothing screams progress like spinning in circles with fancy words. 🔄💡 Clearly, the future of AI is all about running around in linguistic loops while pretending it's groundbreaking research. 🙄🔍
https://arxiv.org/abs/2510.25741 #loopedlanguagemodels #AIprogress #linguisticspaghetti #researchinnovation #techhumor #HackerNews #ngated
https://arxiv.org/abs/2510.25741 #loopedlanguagemodels #AIprogress #linguisticspaghetti #researchinnovation #techhumor #HackerNews #ngated
Scaling Latent Reasoning via Looped Language Models
Modern LLMs are trained to "think" primarily via explicit text generation, such as chain-of-thought (CoT), which defers reasoning to post-training and under-leverages pre-training data. We present and open-source Ouro, named after the recursive Ouroboros, a family of pre-trained Looped Language Models (LoopLM) that instead build reasoning into the pre-training phase through (i) iterative computation in latent space, (ii) an entropy-regularized objective for learned depth allocation, and (iii) scaling to 7.7T tokens. Ouro 1.4B and 2.6B models enjoy superior performance that match the results of up to 12B SOTA LLMs across a wide range of benchmarks. Through controlled experiments, we show this advantage stems not from increased knowledge capacity, but from superior knowledge manipulation capabilities. We also show that LoopLM yields reasoning traces more aligned with final outputs than explicit CoT. We hope our results show the potential of LoopLM as a novel scaling direction in the reasoning era. Our model is available here: http://ouro-llm.github.io.
2/3 người Mỹ lo ngại AI sẽ gây hại nghiêm trọng cho con người trong 20 năm tới, theo khảo sát của Pew Research. Quan ngại về rủi ro công nghệ tiếp tục gia tăng. #AI #CongNghe #AIandEthics #GiaiDoanViec #AIProgress #AIConcerns #AIrisks #AIdebate #AItechnology #AItroubled #AIissues #AIeffects #AIwatch #AIfuture #AIchaos #AIharm #AIimpact #AIhorizon #AIanalysis #AIstudy #AIconcern #AIworry #AIdich #AIhatred #AIviolence #AIwar #AIdanger #AItragedy #AIproblem #AIcrisis #AIrisk #AIthreat #AIattack #AI