Tweets

ChatGPT for @VirginAustralia: virginaustralia.com/au/en/newsroom…

Long-term investments in basic university research are behind many of the innovations we take for granted today (TCP/IP, RISC processors, ...).  A conversation between Magdalena Balazinska, Partha Ranganathan, Urs Hölzle and me on academia's impact on Google and the journey from university research to global impact is now available on CACM at cacm.acm.org/federal-fundin…

Below is an excerpt of one of the questions Magda asked and answers from Urs and me.

This is part of a broader CACM special issue on the power of federally-funded curiosity-driven research co-ordinated by Margaret Martonosi, Eric Hovitz, Moshe Vardi, and James Larus. The other special issue articles are at cacm.acm.org/section/federa….  There are some other delightful articles  feature some great discussions from Dave Patterson, Jack Dongarra, John Hennessy, Nick McKeown, Jennifer Rexford, Doug Burger, Steve Keckler, Kathryn S McKinley, among others.

Gigawatt-scale data centers can likely be built in 2 years or less

a16z.news/p/charts-of-th…

Every major shift in consumer tech has a moment when it suddenly becomes accessible to millions. Michael Mignano (@mignano) helped spark one of those moments with Anchor, making podcast creation something anyone could do with a tap. Now at Lightspeed, he sees AI bringing a similar leap to music, media, and everyday apps.

In this conversation, Michael and @garrytan trace the arc from the early days of social audio to today's consumer AI boom—and dig into what founders should focus on as the next generation of creative tools takes shape.

00:51 - Michael’s founder experience
06:43 - Consumer landscape and AI impact
08:26 - What makes consumer different from other spaces
10:00 - AI costs and distribution challenges
13:20 - Listening to the users, growth rules 
16:40 - Ecosystem integration, distribution
19:52 - AI-generated content & social media
30:00 - Taste, craft, model energy
33:20 - Data opportunities

Kling O1 is now available in ElevenLabs Image & Video.

Multimodal input (text, image or video), precise pace & detail control, consistent style, and strong character fidelity.

That could be a useful heuristic for picking domains to attack with AI. If the work being done in some domain is already slop when done by humans, then presumably current AI can do it well enough.

Used 3 famous AI products today. 2/3 had full-on outages, the other one failed/hanging generations. 1 came through with what I needed but the experience was rough.

Still much better than not having AI do it (it would have taken me weeks / not having made it at all). But it's a good reminder of how early AI is, and overall how hard it is to make reliable products with cutting-edge tech.

Makes me even more excited about our investments into useworkflow.dev and AI Gateway. You should architect agent products with the assumption that things will fail(over).

Relentless shipping

Out the Window Prompt for 🍌:

经典的“Out the Window” Meme 四格漫画竖版布局，标题在漫画顶部居中，背景为现代办公会议室场景，风格简洁、幽默、清晰易懂，所有人物均以卡通风格表现，适合用于网络传播。

### 标题（居中）：
《AI 自媒体如何吸引流量？》（中文，加粗）

### 第一格（完整横格）：
- **场景**：公司会议室，老板站立在白板或投影幕前讲话，老板为分头发型、西装领带、严肃但略带疑惑的男性（参考图1）。
- **动作**：老板张开双手，似在询问，下方桌前并排坐着三位员工（从左到右：小明、小红、小互）。
- **对话气泡（老板）**：“我们做 AI 自媒体的要如何吸引流量？”
- **员工描述**：
  - 小明：男性，蓝色格子衬衫，戴眼镜，程序员气质。
  - 小红：女性，休闲时尚穿搭，显出设计师的个性气质。
  - 小互：戴圆眼镜，职场休闲风（参考图2）。

### 第二格（完整横格，画面中三个员工分别给出回答，表情认真自信）：
- **小明（左侧）对话气泡**：“紧跟最新 AI 热点”
- **小红（中间）对话气泡**：“分享干货”
- **小互（右侧）对话气泡**：“标题里加一个「炸裂了兄弟们」”

### 第三格（下半行左侧小格）：
- **老板特写镜头**，表情从严肃转为愤怒，面容阴沉，额头出现愤怒的青筋或“井”字符号。

### 第四格（下半行右侧小格）：
- **场景切换到大楼外景**，高层窗户破裂，背景是高楼，眼镜员工小互夸张地被老板踢出窗外，表现幽默的动态姿势，小互表情惊讶并带有喜剧效果，窗内老板依然维持愤怒的表情和姿势。

---

### 附加风格要求（可选但推荐）：
- 人物绘制：卡通化、线条清晰明快，颜色鲜明易辨，易于网络传播。
- 对话气泡：中文文字，字体清晰、大小适中且统一，采用黑色文字、白色背景的标准漫画气泡。

Apple presents CLaRa

Bridging Retrieval and Generation with Continuous Latent Reasoning