Aura · 时间线

62171 条条目 · 106 个活跃源

@alexandr_wang: looks like a butthole

@dotey: 😂 https://t.co/JdD2z6nUZw

@emollick: Model personality matters in a way that benchmarks can't capture. Opus models went through a rough patch from around 4.7 to 5 where they just didn't feel "Claude-y" anymore, more like an watered-down Fable (hmmm, teacher models?). Opus 5.5 feels like working with ol' Claude again

@dotey: 亚历山大·王这篇短文《我为什么要做 Muse》算是把 Muse 的定位解释的比较清楚。

@emollick: It is strange how much LLMs turned out to be the key to such a wide range of problems that would not, initially, seem to be problems that a model of human language would be able to solve.

PNOE’s new face mask wants to make lab-grade breath testing a self-serve affair

Google tests buying from Walmart-owned Flipkart through Gemini and AI Mode in India

@elonmusk: Yes 😂

@alexandr_wang: inside you there are two wolves

@dotey: 几个月前我悟到的一条经验就是不要给模型模板，而是给它目标、规范，让模型按照自己熟悉的方式去自由发挥，反而比套目标会更好。

@simonw: I included a few references to this year's record-breaking Kākāpō breeding season in a talk I gave yesterday, and since Claude Opus 5.5 is surprisingly capable at pixel art animation I had it create this celebratory video for my closing slide https://t.co/0rPkQ8S9dn

@dotey: 《微软五十年》

Kākāpō Party

@emollick: I only liked the classics. They don't make games like they used to. https://t.co/rkgBf5ip2F

@dotey: 《什么是 DINOv3》

@dotey: 《中华文明史》by Opus 5.5

@elonmusk: Production on the Moon &amp; Mars will accelerate by far more than double every year, until natural limits are encountered

@elonmusk: Grok @Bot can manage your finances

Insurers claim AI is already increasing healthcare costs

@_akhaliq: ProgramDistill

TikTok agrees to pay at least $100M in Alabama settlement

@natolambert: We're in a phase where companies whose work-culture is suited to how frontier models interact with employees out of the box will receive a massive acceleration relative to their competitors. Same goes for how flexible employees are in using them.

Meta says it will run ads for ‘Musk’ documentary after all

@elonmusk: This is exactly what they will do.

Levoit’s new air purifier is for the pet odors that have taken over your apartment

@emollick: "so this is fun, and exactly what i wanted, but now lets try one that actually is educational"

@elonmusk: This keeps getting worse

@svpino: This is a pretty good idea: you can refine how your agent works by telling it how to behave.

@ClementDelangue: Open-source RL envs for the win. On Hugging Face of course!

🤖 AI Agents Weekly: Claude Opus 5.5, GPT-6 Sol and Luna, MiMo-V2.6, Step 5 Preview, Google AX, Agensh, and More

@alexandr_wang: love to see great reviews on muse!

@elonmusk: Rockets don’t need vision

I created an interactive digital avatar of myself — and you can talk to it

@emollick: Leaving aside the arguments over the reasons why this has happened, it is shocking that Europe does not have a single frontier AI lab, nor even a near-frontier lab nor even an effort that could likely lead to building up a frontier lab in the future.

@svpino: Using GPT-6 to call the Hyper3D Rodin MCP to make a racing game, without writing a single line of code.

@natolambert: Hire DHH to give every undergrad cs graduation speech for a few years. This is the mindset that is needed to enjoy and thrive in what comes next with building software.

@rasbt: Reasoning from scratch, round number 5! This time, talking about log-probability scoring (also a great fundamental concept for loss functions like cross-entropy in pre-training and distillation) and self-refinement.

Old-School Credit Card Scams Are Far From Dead

12 Best White Elephant Gifts, Plus a Prank Box to Put Them In (2026)

Meta’s Muse Is Adults-Only. Why Does It Look Like a Kids’ Toy?

This Browser Extension Lets You Snooze Open Tabs Until Later

A Gravitational Battle Within the Earth Is Changing the Length of Days

The Best Early Prime Day Deals Ahead of Amazon’s Second Sale (2026)

@dotey: 求推荐 GitHub 上开源的 Awesome Opus 5.5 Video Prompt 集合

@dotey: 这篇文章解释清楚了为什么现在 Opus 5.5 要比之前 Opus 5 耐用多了，看起来 API 价格降低了 20%（缓存读取降了 60%），但是账单高低其实看任务跑了多少轮，因为每跑一轮，都要把之前的整段对话重新发给模型一遍。

@elonmusk: Troubling

@emollick: And the incidents apparently continue.

@dotey: Claude Code 任务做到一半撞上 5 小时限额时，会先找个合适的地方停下，不再改代码改到一半就直接断掉。收尾用的是一小笔固定额度，从你的每周额度里扣。

MoSign: Challenge-Response Motion-Watermark Authentication for Anonymous Virtual-Reality Users

Evidence-Driven Differential Diagnosis of Malignant Melanoma

@elonmusk: Production on the Moon & Mars will accelerate by far more than double every year, until natural limits are encountered