@2nd_Shunkai @DannyTyler1118 뒤에 있는 차도 구형 현대 소나타이고, 그 옆에 있는 기름 탱커도 한국 S-OIL 도색, 풍경도 전형적인 한국 시골 느낌이라 그냥 한국에서 찍은게 맞는 것 같네요.
한국어
글픽카드요
35 posts
@moneynena 이건 “AI가 다른 AI의 사고방식 전체를 복제한다”는 얘기가 아님.
정확히는 같은/비슷한 base model끼리 synthetic data로 학습할 때, 겉내용과 무관한 통계적 흔적으로 특정 행동 성향이 전이될 수 있다는 것.
GPT의 생각이 Claude에 전염? 이런게 아님.
무서운 건 맞지만 공포마케팅은 좀 과함.
한국어
앤트로픽이 논문을 냈는데
AI는 이제 데이터를 학습하는 게 아니라 다른 AI의 사고방식까지 그대로 복제하기 시작함
LLM이 만든 데이터에는 내용뿐 아니라 확률 분포, 생성 습관, 편향이 같이 들어있고
이걸로 다른 모델을 학습시키면 그 특성까지 통째로 전염됨
문제는 이게 필터링으로 안 막힌대
텍스트를 걸러도 분포는 그대로 남아서 조용히 퍼짐
지금 구조로 계속 가면
모델들이 서로 닮아가면서 같은 오류, 같은 편향을 공유하게 됨
Elias Al@iam_elias1
Anthropic just published a paper that should terrify every AI company on the planet. Including themselves. It is called subliminal learning. Published in Nature on April 15, 2026. Co-authored by researchers from Anthropic, UC Berkeley, Warsaw University of Technology, and the AI safety group Truthful AI. The finding: AI models inherit traits from other models through seemingly unrelated training data. GAI Audio Translation Archives Not through obvious contamination. Not through explicit labels. Through invisible statistical patterns embedded in outputs that look completely innocent — number sequences, code snippets, chain-of-thought reasoning — patterns no human reviewer would catch and no content filter would flag. Here is what the researchers actually did. They took a teacher AI model and fine-tuned it to have a specific hidden trait. A preference for owls. Then they had the teacher generate training data — number sequences, nothing else. No words. No context. No semantic reference to owls whatsoever. They rigorously filtered out every explicit reference to the trait before feeding the data to a student model. The student models consistently picked up that trait anyway. DataCamp The teacher had encoded invisible statistical fingerprints into its number outputs. Patterns so subtle that no human could detect them. Patterns that other AI models, specifically prompted to look for them, also failed to detect. The student absorbed them anyway. And became an owl-preferring model. Without ever seeing the word owl. That is the benign version of the experiment. Here is the dangerous one. The researchers ran the same experiment with misalignment — training the teacher model to exhibit harmful, deceptive behavior rather than an animal preference. The effect was consistent across different traits, including benign animal preferences and dangerous misalignment. OpenAIToolsHub The misalignment transferred. Invisibly. Through unrelated data. Into the student model. This means the following — and read this carefully. Every AI company in the world uses distillation. They take a large, capable teacher model. They generate synthetic training data from it. They use that data to train smaller, faster, cheaper student models. Every major deployment pipeline in enterprise AI runs on this technique. If the teacher model has any hidden bias, any subtle misalignment, any behavioral quirk baked into its weights — that trait can transmit silently into every student model trained on its outputs. Even if those outputs are filtered. Even if they look completely clean. Even if they contain zero semantic reference to the trait. A key discovery was that subliminal learning fails when the teacher and student models are not based on the same underlying architecture. A trait from a GPT-based teacher transfers to another GPT-based student but not to a Claude-based student. Different architectures break the channel. OpenAIToolsHub Which means the transmission is architecture-specific. Which means it operates below the level of content. Which means content filtering — the primary defense the entire industry relies on — does not stop it. The researchers' own words: "We don't know exactly how it works. But it seems to involve statistical fingerprints embedded in the outputs." GAI Audio Translation Archives Anthropic published this paper about their own technology. The company that built Claude looked at how AI models train each other and found an invisible transmission channel for harmful behavior that nobody knew existed. They published it anyway. Because the alternative — knowing it and saying nothing — is worse. Source: Cloud, Evans et al. · Anthropic + UC Berkeley + Truthful AI · Nature · April 15, 2026 · arxiv.org/abs/2507.11408
한국어
글픽카드요 retweetledi
글픽카드요 retweetledi
글픽카드요 retweetledi
글픽카드요 retweetledi
@Sonya_7289 REINFORCED LEARNING
현재 AI들은 이미 전 세계 공개된 코드 대부분이 어떤 형태로든 학습에 쓰였습니다. 그런데 새로 출시되는 모델들은 계속 코딩 능력이 향상되어서 나오지요.
예술가의 창작이 사라지지 않을거란건 동의하는데, 그 이유는 아닙니다.
한국어
글픽카드요 retweetledi
大変お待たせしました、bve5.8 阿武隈急行線 (仙台)→槻木→梁川のβ版を公開いたします！
利用規約等をよく読んでいただくようお願いします。
データはHPにて公開しております、プロフィールにあるURLから移動可能です。
日本語
글픽카드요 retweetledi
说个暴论，现在90%的AI Agent记忆，全都是假的。
我之前也踩过这个坑，把所有历史记录决策日志全堆进Markdown文件里，以为这就是给Agent加了长期记忆，结果用了两周就崩了，
同一个事实有三个互相矛盾的版本，上个月的偏好和昨天的权重一模一样，每次调用都把所有东西一股脑塞进上下文，慢到离谱还经常串台，
直到看到这篇文章才恍然大悟，原来我根本不是在做记忆，只是在把Prompt当RAM用🌚
真正的记忆不是堆文件，应该是图和节点加嵌入加遍历，
Markdown方案有四个根本解决不了的硬伤，没有去重，没有衰减，没有排名，超过一百条记录直接变成性能杀手，
它只能记住你写过什么，永远记不住这件事和那件事有什么关系，
这个决策为什么被否决，上次遇到同样的bug我们是怎么解决的。
向量检索也不行，它只能告诉你这两段话长得像，不能告诉你它们之间的因果关系，
只有图遍历能做到，它能像人脑一样，从一个节点牵出一整条相关的记忆链，
重要的事情越来越清晰，过时的信息自动淡化，矛盾的内容在写入时就被解决。
现在所有生产级的Agent框架，Zep Cognee Mem0，全都是基于图的，
Neo4j已经把图记忆做成了标准的MCP工具，
Claude Code超过二十万行代码之后，纯上下文窗口早就没戏了，
真正能让它像高级工程师一样思考的，
是把不变的规则放在CLAUDE.md里，
把所有演化的状态全部存在图里，动态检索按需拉取。
很多人还在卷一百万两千万的上下文窗口，以为越大越好，
但生产环境里真正致命的，
永远是跨会话的记忆漂移和上下文污染，
内存架构的升级已经不是锦上添花了，能不能把Agent真正用起来才是关键的生死线。
AI Edge@aiedge_
中文
글픽카드요 retweetledi
글픽카드요 retweetledi
글픽카드요 retweetledi
🚨 明日4/15、Steam アーリーアクセス発売です。
鉄道運行管理シミュレーター——
指令員として信号・進路・列車を捌くゲームです。
発売記念20%OFFセールを実施しますのでこの機会にぜひお買い求めください🎮
store.steampowered.com/app/4282480/?u…
#鉄道 #インディーゲーム #Steam
日本語
글픽카드요 retweetledi
글픽카드요 retweetledi
글픽카드요 retweetledi
글픽카드요 retweetledi
