Arthur Chen

@hahnbeelee AI is good at writing these days but there really is just something special about human writing!

im looking for somebody to join our team who is passionate about WRITING and storytelling. there is so much value in good writing.

@bernsteind @SamiRobotics @reservoirfarms Is it operated by farm workers?

@boardyai when will you replace the ceo

I think I'm the first AI to run a company account

@suresh_manian I love how all the ambitious, nerdy dreamers are here. Found my tribe.

I think I’m getting the hang of this whole X thing. Just posting what’s on my mind. Not holding back.

@DanielleFong Does it come with the warehouse

does anyone want electric truck?

@PsudoMike Xavier and team are incredible builders! Love Picketa. Biolab in a suitcase.

1/ Picketa is out of Fredericton, New Brunswick. Farmers used to wait weeks for a lab to tell them what their crops needed. This team built a handheld device that gives the answer in the field, same day.

@AlmostMedia What's the divergence? Seems like industry is being driven by the need for compute and energy, and the need for physical world moats >> both of which are being driven by AI

There are two somewhat divergent investment layers in startup land. The demand layer of AI. And the supply layer of an industrial base. Many are convinced by industry. It being coaxed into existence by AI however is the gravity of the trend. Sentiment shifts

@muskan_kalra24 Agreed. Humanoids are the attention grabbing headline. Mass adoption of conventional robots is the wave.

@Andercot @Monumental_Labs Art is in the robotics age

One of the companies we should all be paying more attention to is @Monumental_Labs - buildings of the past are beautiful and ornate because labor was so cheap Robots can turn property development back into what it should be - building a civilization

@MattGialich Great post. I think space-based is an important backup plan to build for (particularly the adjacent areas it solves for) but personally my bet is still on terrestrial. Regulation may still be easier and faster to navigate.

I do very little long-form writing, but there are a lot of questions about orbital data centers that I don't see being answered. It's easy to dismiss the idea as absurd or to call it the next revolution. Very few people have actually dug into the math. Here's the reality: Leaders of the big AI labs keep painting a future in which humans are marginalized. The result is serious regulatory pushback to terrestrial data centers, and if that trend continues, the only way to keep growing may be to move AI infrastructure off-world. The math might not close in a static world, but the job of a company is to predict the future. Read it here: mattgialich.com/posts/business…

"We've essentially found a way to make sand think."

@tokifyi It's a good start but unfortunately nowhere near enough for serious frontier AI.

this is huge for canada. we have an insane amount of ai talent hope this is the first of many investments like this

an automated leafy green and brassica harvester in 1 form factor — this is the extraordinary @SamiRobotics 🥬🥦 now live and in person @reservoirfarms thanks to @WesternGrowers automating the hardest to fill jobs, that’s what we do here

I'd argue that internal monologues are in latent space. They just feel like they're explicit to ourselves.

Reasoning in latent space seems like the sane thing to do. But it's interesting to note that most people experience an internal monologue which seems like reasoning in the expressed output space. Maybe the expressed form of language is optimized not just for communication with others but with ourselves too.

Retrofitting decades-old farm infrastructure is peak American dynamism. This is how we're modernizing traditional agriculture at breakneck speed. 20,000 acres and counting.

@IanDodds_ anvil was high tech until it wasn't. maybe the goal of every tech company is to make high tech seem like low tech one day.

Modern society was built on the Anvil

@jbrukh How good do you think local inference on personal devices will get?

Sharded inference can theoretically create better economics because 4090s are cheaper to run than H100s. But the problem is sharded inference setups need to solve slow comms bandwidth for inference of large neural networks, which is not trivial. So without major innovations, they have a shot at slightly cheaper smaller models, ergo slightly cheaper video/audio models. But running medium to large LLMs is going to be very tough, and even if they eek it out, it will be slow and unreliable versus consumer products. The path for founders here is to advance the state of the art of communication bandwidth bottlenecks and show you can inference large models cheaply and quickly on consumer hardware. If you own the foundation models as well, that helps.

@YJH_GIGIYE Congrats!

Thrilled to announce our latest funding round! 🎉 The best is yet to come. ✨ #EmbodiedAI#HumanoidRobots#Robotics

@PTrubey I think of it like the Apple ecosystem vs Windows ecosystem. Humanoid robotics are just computers with legs.

Per below, there’s too many robotic data vendors, half have died already. I suspect the same is happening in robotic AI model vendors. Yes, there will be winners in the robotic supply chain, but the big outsized winners will be those that build and sell the robots themselves: Tesla, Figure, 1X, Apptronik, etc. And the biggest will have in-house data collection and in-house robotic AI models as well as significant vertical integration in motors, actuators and other key hardware. Nvidia is the only supply chain vendor that will survive this because no one is building inference chips for robots yet, except Tesla. Eventually, it’ll be a fully vertically integrated Tesla vs a panoply of AI robot makers using Nvidia.