←
Robert Scoble265,684 posts
Robert Scoble avatar

Robert Scoble

@Scobleizer

The automated life | San Francisco/Silicon Valley AI, robotics, BCIs | Ex-Microsoft, Rackspace, Fast Company | Wrote eight books about the future.

◷ Joined November 2006▧ 6,990 Photos and videos
Robert Scoble avatar
Robert Scoble@Scobleizer·Nov 18

AI Is learning to build its own successor.   AI is no longer just getting smarter. AI is literally learning to build its own successor. By assisting researchers with coding, experiments, and model development, AI could dramatically accelerate the pace of innovation. But as machines help create increasingly powerful machines, a critical question emerges: Can humans keep up with the technology they are building, or will AI innovation begin moving faster than our ability to understand and control it? @IrenaCronin and I write this free newsletter every week on different current AI topics.   Read and subscribe for free here: unaligned.io

Post attachment
NWS Bay Area 🌉 avatar
⟳ Robert Scoble retweeted
NWS Bay Area 🌉@NWSBayArea·5h

Record daily rainfall for Downtown San Francisco. As of 3 pm 2.28 inches of rain has fallen so far today (since midnight). The old record for this date was 2.24 inches back in 1918. The final numbers will be calculated at midnight with more rain forecast to fall this evening.

Emerson S avatar
⟳ Robert Scoble retweeted
Emerson S@Em_Nomadic·6h

We’re watching the earliest stages of robots learning directly from human expertise. Not just copying motion, but turning demonstrations into deployable skills. That shift is going to matter a lot.

Today, we're introducing @Nimo_Robotics. We're building deployable dexterity: the fastest path from human expertise to validated robot skills for industrial work. Most dexterous hands replicate the human hand, joint for joint. But is that the most efficient path? High degrees of freedom and the embodiment gap are why dexterous manipulation demands so much data today. We went back to first principles: what is the minimal system that can complete real industrial tasks, such as multi-step assembly? Our answer is the N1 glove and D1 hand: to our knowledge, the highest-DoF commercial UMI system with matched visuals and kinematics. Nine independently actuated finger DoF keep the action space compact. Matched kinematics and appearance close the embodiment gap. What an expert demonstrates is what the robot learns. Nimo Core, our learning system, turns demonstrations into skills using vision, touch, and sound. It can hear a connector click into place as well as see it. Our goal: from human demonstration to a trained, validated, and deployed robot skill in one day. We're opening the Nimo beta program for 20 researchers and developers from academia and industry. Enjoy our launch film! Apply for the beta or discuss an industrial pilot: nimobot.ai/

1:30
Furqan Rydhan avatar
⟳ Robert Scoble retweeted
Furqan Rydhan@FurqanR·9h

Voice is the ultimate medium of conversation. The future of work is just talking.

The future won't have keyboards. And it definitely won't be built with them. Introducing Nebula Voice.

1:38
Robert Scoble avatar
Robert Scoble@Scobleizer·3h

The annoying part of using a bunch of AI agents isn't the agents. It's being the person who copies stuff between them all day. @switch_ai_ puts them in one room with your team. You just tell Buddy the team you want, a planner, a coder, a reviewer and a bug fixer, and it sets them up. Open source too.

Say hi to Switch Buddy 👋 Claude Code writes code. Codex reviews it. Another agent debugs. No copying context or managing handoffs. Switch AI brings them into shared rooms with your team. Now Buddy helps you set it all up in minutes. Open source. Your agents. Your choice.

0:20
Robert Scoble avatar
Robert Scoble@Scobleizer·4h

This leads to preventative health.

Introducing Helios: measuring 1,000+ protein biomarkers from a drop of blood to help you live younger, longer. Here’s how it works: • Self-collect blood at home with a painless, needle-free device • We measure the proteins carrying signals from cells and organs throughout your body • Your Helios AI agent explains your results, highlights risks, and offers personalized guidance Proudly backed by Y Combinator and Formation VC. Join the waitlist at myhelios.com

1:19
@jason avatar
⟳ Robert Scoble retweeted
@jason@Jason·6h

Breaking: via @Reuters and @alexalper: Trump administration is weighing a rule that would force companies to disclose when they use Chinese open-source AI models on their website Not a ban. A "name and shame" play, per multiple sources. 🧵

Robert Scoble avatar
Robert Scoble@Scobleizer·4h

It is happening.

Video is the most important medium, but Quest users can't edit videos without leaving VR. That's why for the Meta VR Start competition, I made Handcut, a VR video editor! Cut a clip by karate chopping it. Scroll the timeline through swipes. Pinch handles to resize clips.

0:16
Surya Pandiaraju avatar
⟳ Robert Scoble retweeted
Surya Pandiaraju@suryapandiaraju·7h

@PreloadAI (YC F26) is redefining the data layer for physical AI: internet-scale training data from cameras, force-sensing gloves, and EMG, captured from expert work like aircraft maintenance and surgery. Built by ex-Neuralink, Meta, and Paradromics engineers. Demo + careers in comments 👇 [Let us know if you catch any easter-eggs in our video 🙂] @ycombinator

1:14
Mike Kalil avatar
⟳ Robert Scoble retweeted
Mike Kalil@mikekalilmfg·Nov 18

Look, Jensen Huang isn't coming back to NVIDIA as CEO. Any backlash $NVDA's board of directors is getting now is nothing compared to what will come their way if they reappoint him. Everyone who wants him back will move on but it'll make Huang and NVIDIA nonstop targets for the next two years at least. Regardless of whether Alice Huang is a Chinese spy, the affair with his direct report is what really escalated things. Nobody thought Astronomer's CEO should stay on when he got caught creepin' with the HR director. No large company outside of Space X would tolerate the CEO romancing employees. The most they'll do is make an announcement that he's working closely with NVIDIA as a chief strategy advisor or something.

0:15
dale avatar
⟳ Robert Scoble retweeted
dale@daleverett·5h

Today we’re launching FlashDrive with @ycombinator 🚀 FlashDrive is a shared cloud drive where your team and all your AI agents work in the same files across computers and cloud sandboxes. > Search the whole drive with grep, regex or code. Our custom indexing layer finds matches without downloading everything. > Files stay in the cloud, and your agent only streams the part it needs. > Start in Claude Code on your Mac. Pick up in Codex. > Edit the same file at the same time. FlashDrive merges changes or keeps both versions. > Connect over MCP or mount the drive on macOS or Linux. Give a sandbox access to just one folder for as little as five minutes. Dalton and I are building the filesystem the next generation of companies will run on. And we're calling it @flashdrive_ai ↓ Try it free! Link below

1:30
Robert Scoble avatar
Robert Scoble@Scobleizer·5h

He creates a model of a person.

Had a fun chat with @YODELAGARZA and @martin_casado! Few things we touched on: 1. Second starts simply as a streaming-state, portable context infra, i.e. lots of tools and APIs to deliver context. - This means you can just add any agent on top with (your own) skill file. Second itself can be an agent this way. - Early users also just build their own apps on top! 2. User modeling takes time, just like how humans take time to know you. But it can gradually show up as more/better APIs over the same infra, so the same agents you plug in should benefit without user action. User modeling is an open problem! 3. Context continuity can unlock new interactions.For example, Second enables you to browse holiday gifts on laptop and then use Grok Bot to buy it seconds after. 4. Privacy/security and access control is hard. - It should be done on both systems level (e.g. confidential inference and VM) and at test-time (contextual integrity from reasoning with your user model). Most context is natural language and you can’t just apply rules to them. - There’s an interesting self-reinforcing effect with user models: if it knows you well, then it’d share with your privacy preferences in mind. (x.com/a16z/status/2123089603…) 5. Second can be very expensive initially. - Rough math: 5 agents 24/7 @ 60 TPS output = $1k/day with Opus 5.5, or $30/day with DS 4.1 Flash. - It has to be a hybrid: some tasks (analyzing your goals) needs stronger models than others (what's a meaningful moment in time). - Good news is that the labs are driving costs down drastically! Things are already 50% cheaper since we started. Conceivably the whole thing can run on-device in the future, so portability must be built-in!

33:11
Robert Scoble avatar
Robert Scoble@Scobleizer·5h

I did one of his first interviews when he was younger. I'll be honest, I wasn't sure he was gonna make it, but he showed me.

Aaron @levie has been CEO of @Box for 21 years, which means he can now legally drink CEO alcohol. On the latest episode of @Rippling First Principles, we dissected his out-of-date tweets, the cold email that got @mcuban onto the cap table, and why his 7 yo is a @dwarkesh_sp fan

1:02:08
Founders Inc avatar
⟳ Robert Scoble retweeted
Founders Inc@fdotinc·8h

These guys made a model that can run autonomous lab experiments It beats GPT Astra on every benchmark while being small enough to run on a raspberry pi: Nvidia's RoboLab-120 - GPT Astra: 74.8% - Theirs: 84.8% LabOSBench - GPT Astra: 76.5% - Theirs: 87.6%

2:08
Robert Scoble avatar
Robert Scoble@Scobleizer·5h

Our immigration policy in America sucks. Smart people like Laura should be courted to come to our country. Smart people like that create jobs, reduce poverty, decrease suffering, and improve everybody's lives. This depresses me to no end.

My friend Laura has 56 days to find a job or leave the US. Last Friday her company NextSense furloughed the whole team. Laura is on an H1-B, which requires her to be paid while employed so her role was terminated and the 60 day clock started. She's a biomedical engineer with a PhD in neuroscience. She builds ML for sleep and time-series signals like EEG/EMG and she has built wearable AI models in production. She's looking for research scientist roles in the Bay Area. If you're hiring in health tech or wearables, DM me and I'll intro you. @ycombinator founders: if you need a founding ML or research hire, she's it. Her H-1B transfers so no lottery needed. And if you're not hiring, a repost might reach the person who is 🙏 linkedin.com/in/laurarose95/

Post attachment