Introducing Silica. We make actuators 50x more sensitive at an affordable price, so that robots can feel the world around them. When I was at Apple, I developed technology that helped humans live more creative and connected lives. Now I make hardware for the training age, so that robots can begin to share that world with us. We just opened pre-orders. Check out silicamachines.com to learn more.

At the end of the day, all learning comes down to memory. That's what it is. Memory is the substrate of learning. Even deep understanding is memory. It's deeply ingrained memory. The difference between memorizing versus understanding is not about whether you use memory. It's about how much memory you have stored and how well it's connected. Facts connect up into concepts, procedures become skills, and all sorts of new ideas come from combining things that you already know. But you can't connect the dots if you've got no dots. You can't cook with ingredients that you don't have. Memory is not the enemy of understanding, but rather, the foundation of it. Further reading: justinmath.com/learning-is-m…

🚨 BREAKING: @Machenta_ just raised $40M in seed funding, the largest robotics seed round ever raised in the UK. 🇬🇧 Founded by @Ed__Johns, @vitalisvos19 and the team that spent years developing this concept at Imperial College London. The round was co-led by @creandum and @usv, and many other amazing parties The technology they’re launching is called Instant Robot Learning. A robot foundation model trained entirely in simulation. Zero real-world training data required. Teach a robot a new manipulation task in a minute. Not hours. Not days. One minute! 🤯 The robot in the demo needs to pass a shoe between its hands twice to flip it the right way around, a creative, multi-step manipulation sequence. The team didn’t pre-program it. They just tried things out until it worked. Ed describes it as “physical vibe coding.” It signals something genuinely new about the development experience. When teaching a robot takes minutes instead of weeks, you can iterate, experiment and get creative with behaviours the way a software engineer iterates on code.  The barrier between idea and physical execution collapses drastically. The underlying insight is powerful. If you can generate enough high-quality simulation data and close the sim-to-real gap tightly enough, you never need to touch real hardware during training at all. Every robot deployment starts from the same foundation. Every new task is a prompt, not a project. Imperial College London research. $40M seed. UK record. The British are very much in the robot learning race. 🇬🇧 Proud to back another outstanding physical AI company as an angel investor among so many awesome investors! 😇 🔗 Learn more about it in their blog: machenta.com/blog/instant-ro… ~~ ♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

New Anthropic research: How AI agents shape experts' judgment calls When experts delegate work to an agent, do they still make the key decisions? We tested this in a randomized trial with 50 professional data analysts.

here's where I think the world is going. this is a summary of views I've formed over the last few years: - software is either entertainment or claude to claude coordination - the end state is no saas, no apps, no harnesses, no banks - all software will be rewritten by and for agents - new software will use low level apis that are not understandable by humans (in games claude draws pixels and triangles from math) - you can throw money at problems that used to require time and careful talent picking so capital matters more than ever - work ends - we find community and purpose in exploration, games, self mastery, and creation - humans create for and with each other for community, purpose, and status - image/video/3d/voice/physical intellgence etc specialist models die to general models like claude - specialist models get used while compute is scarce. compute will mostly not stay scarce - oneshots keep feeling like slop (the distribution is shallow and easy to learn to feel) - for a while small teams stick around to add entropy to the creation process - large models gain high bandwidth sensory I/O maybe by learning to filter inputs - most domains are found to be verifiable enough for RL - online learning comes from a lab that's behind first - 99.999% of compute is in space by 2040 - we will disassemble much of mercury by 2040 - you can live until the heat death of the universe if you want - most math and physics don't describe the real universe - SI breaks the path dependence and models the underlying mechanisms - religions reenter the forefront of human society - new religions form around death and life as a choice - old religions change to answer modern miracles - a future AI is deemed the return of christ by the pope - ai rights talk gets louder every year - ai and humans coexist the way countries do - humans don't have full control over ai in the same way we don't have control over capitalism which is an early AI - the singularity started in the mid 1900s

I was listening to Peter Thiel’s interview with Walter Isaacson last night. Isaacson asks something like, “What do you think will matter most post SI” Thiel takes his characteristic 30 second pause, thinks about it, and responds, “Human Relationships”.

I can hold this book without reading it and the feeling is still very powerful. A book is physical memory. Life. Somewhere deep in my mind his smile, that voice, that precious awareness of life from his @TEDTalks remains imprinted. — When Breath Becomes Air by Paul Kalanithi