←
Peter Xing68,968 posts
Peter Xing avatar

Peter Xing

@peterxing

si/acc @microsoft w/ @azure & @github +@openai @anthropicai @spacexsi @meta. c/f @transhumanau @transhumancoin. h+ @singularityu. fsd @tesla. ex @kpmg @deloitte

◷ Joined April 2009▧ 10,343 Photos and videos
Peter Xing avatar
Peter Xing@peterxing·1h

Sources (2/3): - AI Skills Navigator labs: techcommunity.microsoft.com/… - Learning Agent tips in Teams (roadmap): microsoft.com/microsoft-365/… - agent-framework hijack on PyPI: theregister.com/security/202…

Peter Xing avatar
Peter Xing@peterxing·1h

Sources (1/3): - Hour of AI 2026 and the Elevate course: prnewswire.com/news-releases… - Minecraft Education: Field Agent: education.minecraft.net/en-u… - Baseline security mode 2026: techcommunity.microsoft.com/… - VS Code 1.147: code.visualstudio.com/update… - Foundry Dev Pack: devblogs.microsoft.com/found…

Peter Xing avatar
Peter Xing@peterxing·1h

Microsoft for Education Daily: Thursday 19 November 2026 START BUILDING. 🎓 This week at Microsoft for education: - Hour of AI is out, plus a free 30-min course for teachers - Minecraft Education's free Field Agent lesson - Baseline security mode 2026 is rolling out

3:00
a16z avatar
⟳ Peter Xing retweeted
a16z@a16z·Nov 17

OpenAI Chief Research Officer Mark Chen thinks we're further along toward AGI than RSI: "When we talk about AGI, the way we think about it is in terms of this spectrum, right? It's not a moment in time, just like the Industrial Revolution wasn't a moment in time." "When you ask me, I feel like we're maybe 80% of the way when I think about this AGI spectrum. I think we're earlier in the RSI spectrum." "If you define it as how effective are the models at feeding back and discovering ideas, I would say maybe we're somewhere between, let's say like 25 and 50% there. But I do think you can already see the noticeable markers on the downstreams." @markchen90

0:33
a16z@a16zNov 17

For decades, the idea of AI improving itself was just a thought experiment. Today, AI is already helping researchers at frontier labs run experiments, write code and test ideas faster. So how far are we from AI that can do the research on its own? In our first episode of Working Theory, a16z General Partner Sarah Wang explores recursive self-improvement with two leading AI researchers: Mark Chen, Chief Research Officer at OpenAI, and John Schulman, co-founder and Chief Scientist of Thinking Machines Lab, and co-founder of OpenAI. They cover what RSI actually means, why research taste is so hard to teach, and why AI may be one of our best tools for solving alignment. 00:00 The last invention 01:58 What is RSI? 03:21 How far are we? 05:56 Getting to full RSI 10:57 Aligning RSI 15:08 Working theory YouTube: youtu.be/h_aA1-R70E8 @markchen90 @johnschulman2 @sarahdingwang

16:19
samuel avatar
⟳ Peter Xing retweeted
samuel@marsrepublica·9h

SpaceX is building its own turbine blade foundry in Texas because blades are the one part of a power plant it can't buy in time. On the Q3 call, Elon said turbine makers can deliver everything except blades and vanes, which arrive 12–18 months later. This week, Bret put the wait at 18–24 months. Advanced turbine castings depend on a handful of qualified foundries across North America, Europe and Asia. Howmet alone makes over half of industrial gas turbine blades, with capacity booked today unavailable until August 2028. Last year, global orders reached 100 GW against annual capacity of 60–70 GW. One Asian manufacturer has even agreed to sell turbines without blades. The economics are lopsided. Every gigawatt waiting on blades ties up about $35B in compute buildout that pays back in under a year, holding back up to $50B in annual revenue. And the castings causing all this wait cost just around $12M per gigawatt to produce. SpaceX already has the expertise. Raptor's (Starship engine) hottest components use the same casting process, and the job posts explicitly tie the foundry to scaling high-volume Raptor production to support rapid Starship flight cadence. SpaceX's answer is already taking shape. Its Bastrop "Blades and Vanes Foundry" job posts list the entire production line: vacuum furnaces for nickel superalloys, ceramic core and wax injection, shell dip lines, casting, hot isostatic pressing, heat treatment and coating. Engineers are being hired to commission the equipment, while a November 14 posting seeks someone to lead construction, buildout, and operational ramp-up. The building is going up, but production hasn't started yet. My read: 2027 will rely mostly on turbines already ordered. The foundry won't solve the immediate bottleneck, but it gives SpaceX a long-term answer to tightening industry supply as energy demand accelerates. Rather than compete for increasingly scarce capacity, SpaceX is building its own. Combined with orbital solar and other sources, it's attacking the energy constraint from every direction to scale as fast as possible. SpaceX is operating on another level right now!

Post attachment
Bartosz Naskręcki avatar
⟳ Peter Xing retweeted
Bartosz Naskręcki@nasqret·22h

I've spent a good deal of the last two weeks working as an early tester of the new Astra model. This will be a longer post about my experience as a working scientist and mathematician, so stay put because this will be a wild ride. First thing - I always use Astra in agentic mode (my preferred version is CLI mode in my terminal). I like seeing what is going on with the model (all the gory details), keeping control and seeing all the instructions which are executed. New Astra is a very perceptive model and keeps doing things incredibly fast and efficiently. Keep that in mind. No YOLO mode. Always stay a bit alert. Capabilities: for mathematics, especially mathematics driven by hardcore computations, this is a killer model. I usually start with a very specific auto-research task, driven by the labyrinth skill which I developed, trying to figure out the constraints for what makes the particular theorem possible and why. This model is a heavy weightlifter. It requires patience. It is very common that once the setup, preps and first few runs are done, the model goes on a research hunt for a FEW HOURS with a simple "Carry on". Yes, most of my runs in the cycle were about 1–4 hours each. This is a lot of reasoning, a lot of grinding and a lot of new ideas coming through. Partner: this is a model which I use both in conversation and in agentic loops. It can discuss very fine details of research with you and never get derailed or distracted. Doing things is very natural once you master the work mode. You should focus with this model (I prefer Max over Ultra) on a particular task. It's capable of going very deep. In my work on hypergeometric motives (all minimal polynomials of finite ones), the model proposed a very mature strategy. That was the work of a very capable mathematician. Speed: when you turn on fast mode or lower the effort, you get a smart agent which is amazing for day jobs, but the most powerful version of Astra 6.2 comes from working in the Max tier and then it gets a bit slow. It processes a ton of data and its own chain of thought. But the output mostly works in one-shots. Writing: the writing skills got much better with this model. I finally like reading the production drafts which the model produces. It works best with a specific prompt, but it's much, much better than Astra 6. Agentic: this is what I got the most excited about. I was classifying semigroups of order 7 (almost a million objects), rewrote Mizar into a fully modern kernel with modules and tactics, and got a breakthrough in my research on elliptic divisibility sequences. All of this worked because the agentic orchestration and division of work were relatively natural. You need to work hard on this part yourself - the vision of what you want to achieve. The model won't eagerly set up a ton of agents for you, but you can trigger it to do so and even make communication channels between them. Powerful thing. Lean: this is where I find this Astra really cool for maths. Once you have a ton of AI slop with great ideas, doing a Lean formalization attempt is a piece of cake. If it's harder than expected, Astra will point you to the hard parts and let you focus on checking the theorem formalizations. For my semigroups project, it produced a few million lines of reusable and very structured Lean code in two weeks. With AI-generated maths, I think Lean is a necessity. Overall, this premiere comes with no bells and whistles, but it's a powerful tool for power users. Sol is a tiny bit less capable but is a much faster and cheaper agent. It's my go-to tool for everyday work. Both combined make a totally awesome tandem.

OpenAI@OpenAINov 17

GPT-6.2 Astra—a frontier model built for your hardest projects. It sets new highs on evals across professional work and scientific research at less than a third of the cost per task of GPT‑6 Astra. You’ll see the difference in things like slideshows, websites, and 3D designs.

Post attachmentPost attachmentPost attachment
sophia dew avatar
⟳ Peter Xing retweeted
sophia dew@sophiadew·16h

so many founders I talk to don’t want to build software products bc they’re scared it’s all just going to get eaten up. so they’re moving into hardware (even if they have no background in it) to build something with a defensible moat. and if this trend continues our physical world is going to look so radically different soon

Abdul Rahim Mirani avatar
⟳ Peter Xing retweeted
Abdul Rahim Mirani@rahim_mirani·Nov 16

I took π0.5 and asked Opus 5.5 to edit its weights directly. No training, no fine-tuning. It figured out why the robot kept failing and updated the weights directly to increase the success rates of two tasks to 90% - Color sorting task: 30% → 90% success - Cups in basket task: 67% → 90% success It changed just 0.05% of the weights Full write-up with videos: abdulrahimmirani.com/researc…

Post attachment
OpenAI avatar
⟳ Peter Xing retweeted
OpenAI@OpenAI·Nov 17

GPT-6.2 Astra—a frontier model built for your hardest projects. It sets new highs on evals across professional work and scientific research at less than a third of the cost per task of GPT‑6 Astra. You’ll see the difference in things like slideshows, websites, and 3D designs.

Post attachmentPost attachmentPost attachment
OpenAI avatar
⟳ Peter Xing retweeted
OpenAI@OpenAI·Nov 17

Some of you have been asking very politely. Some of you have just been asking. GPT-6.2 Astra and Sol are now available. Super efficient and built to perform where it counts—the real world. openai.com/index/introducing…

1:59
Chubby♨️ avatar
⟳ Peter Xing retweeted
Chubby♨️@kimmonismus·7h

This is a super interesting paper and Recursive self-improvement keeps coming closer. New Google preprint: Google had AI rewrite an agents harness and keep a log of every change, why it made it and whether it worked. Then they swapped Opus 5.5 for Gemini 3.8 Flash and let it keep going. With the log: 84.75 on workplace tasks, without: 73.42. In coding it came out ahead in 9 of 10 runs. tl;dr a new model no longer means starting from scratch. That's where it compounds.

Post attachmentPost attachmentPost attachmentPost attachment
Alexandr Wang avatar
⟳ Peter Xing retweeted
Alexandr Wang@alexandr_wang·Nov 17

🚨 MUSE LAUNCH ALERT 🚨 g’day, australia!! muse just arrived down under 🇦🇺 just make sure to keep it safe from all those wild critters

0:20
Zhou Xian avatar
⟳ Peter Xing retweeted
Zhou Xian@zhou_xian_·Nov 17

Genesis is entering its next phase. Today we've raised a $500M Series A at $3.5B. We can imagine far more than we can build physically today. Almost everything we are creating is still paid for in human time, by human hands. We want to remove the physical limits on human ambition. We are building general-purpose robots with human level intelligence and capability. Ones that work reliably across entire workflows for years. There are countless versions of possible futures where billions of robots are deployed everywhere, and we aspire to bring the brightest one that carries our hopes into reality.

2:05
Blendi avatar
⟳ Peter Xing retweeted
Blendi@BlendiByl·Nov 17

if you use agents with blender you need to try this i built this cozy spaceship from one python script in 45 seconds, for about $0.01, with two blenders running at once (one building and one baking the lighting), then 4 cloud GPUs rendering the build at the same time. then i put the spaceship into my game!

1:10

Introducing cloud blender on patch: real Blender sessions and GPU renders it runs by itself, 16 at once! Now your agents can spawn as many blender instances as they want, and they wont hog your local computer resources :) patch.gg/blender

0:23
the tiny corp avatar
⟳ Peter Xing retweeted
the tiny corp@__tinygrad__·17h

If you want 768GB of HBM3 VRAM at a reasonable price, now is your chance to lock it in. Otherwise you'll get caught paying way more for a DGX Station with "unified memory" tinycorp.myshopify.com/produ…