Patrick Mooney

So now that everyone has seen the disappointment of attempts to squeeze out a bit more federal power and the risk that it might be used against you... can we all agree we should give federalism a shot and limit the federal government to just its specifically enumerated powers?

Battlestar Galactica cinematographer validated

California’s last attempt at online age verification (CAADCA) was enjoined by a federal court for violating the First Amendment. So they wrote a new law that does the same thing with a different chokepoint. AB-1043 shifts the burden from app developers to operating system providers. Instead of every app asking your age, your OS sends a “signal” to apps telling them whether you’re under 13, 13-16, 16-18, or 18+. Four age brackets, transmitted via API every time you launch an app. The theory is clever. Courts struck down CAADCA because requiring every business to assess content harm to children was a content-based speech regulation that couldn’t survive strict scrutiny. AB-1043 sidesteps this by saying “we’re not regulating content, we’re just making the OS collect a birthday.” The practice is absurd. The law defines “operating system provider” as anyone who “develops, licenses, or controls” an OS. That includes Linus Torvalds. It includes the FreeBSD foundation. It includes every hobbyist who maintains a Linux distribution. The penalty is $2,500 per affected child for negligent violations, $7,500 per intentional violation. Linux distributions don’t have “account setup.” They don’t have app stores with handshake APIs. Most don’t even have mandatory user accounts. The law assumes every computing device works like an iPhone, with a centralized identity system and a curated marketplace. Colorado just introduced an identical bill (SB 26-051) two days ago. This is a coordinated legislative strategy across states, not a one-off California experiment. They’re trying to establish age-gating at the OS layer as a national norm before the legal challenges catch up. The same trade group (NetChoice) that killed CAADCA will almost certainly challenge AB-1043. The First Amendment problem didn’t disappear because you moved the compliance obligation from the app layer to the kernel layer. You just added a step.

We’re building an LLM chip that delivers much higher throughput than any other chip while also achieving the lowest latency. We call it the MatX One. The MatX One chip is based on a splittable systolic array, which has the energy and area efficiency that large systolic arrays are famous for, while also getting high utilization on smaller matrices with flexible shapes. The chip combines the low latency of SRAM-first designs with the long-context support of HBM. These elements, plus a fresh take on numerics, deliver higher throughput on LLMs than any announced system, while simultaneously matching the latency of SRAM-first designs. Higher throughput and lower latency give you smarter and faster models for your subscription dollar. We’ve raised a $500M Series B to wrap up development and quickly scale manufacturing, with tapeout in under a year. The round was led by Jane Street, one of the most tech-savvy Wall Street firms, and Situational Awareness LP, whose founder @leopoldasch wrote the definitive memo on AGI. Participants include @sparkcapital, @danielgross and @natfriedman’s fund, @patrickc and @collision, @TriatomicCap, @HarpoonVentures, @karpathy, @dwarkesh_sp, and others. We’re also welcoming investors across the supply chain, including Marvell and Alchip. @MikeGunter_ and I started MatX because we felt that the best chip for LLMs should be designed from first principles with a deep understanding of what LLMs need and how they will evolve. We are willing to give up on small-model performance, low-volume workloads, and even ease of programming to deliver on such a chip. We’re now a 100-person team with people who think about everything from learning rate schedules, to Swing Modulo Scheduling, to guard/round/sticky bits, to blind-mated connections—all in the same building. If you’d like to help us architect, design, and deploy many generations of chips in large volume, consider joining us.

Thrilled to share that we have closed our $200M Series C. Thank you to Thomas Tull's @USITfund for leading, along with @EclipseVentures, Riot Ventures, @JaneStreetGroup, @intelcapital, @counterpartvc, @friendsfamcap and all of our additional investors. buff.ly/ozCrZUE

Okay so there is magic sauce and part of it is paying your warehouse workers $100k+

@planefag Nuclear tipped SM-3 will remind them who was made in God's image

i assume what's going to happen is minnesota is going to try as hard as it can to prosecute knowing all this, and trump is going to block everything, and it will all drag on interminably and public faith left and right in institutions will decline further enjoy arbys

it's underappreciated that the "guy with a glock" (gender neutral) is actually a key part of the sgx security model

End this godforsaken doctrine, an absolute poison for the people's faith in government. (via @billybinion) reason.com/2025/10/26/pol…

one of my most productive work days ever when I decided I was sick of dealing with an intractable perf problem, took an uber to get ramen, had three beers with lunch, decided to walk the two or three miles back, and realized a redesign that would trivially parallelize 95% of it

I've often made this mistake, and regretted it later when writing up the bug.

@theodorvaryag *bought Mellanox

...nvidia makes NICs?

inverse designed grating coupler with ~ 1 dB coupling loss over ~25 nm bandwidth around c band for both polarizations. why do they cut off two corners of their SEM?

The comparative advantage is almost magic. Free trade will always beat protectionism in the long run. And tariffs are taxes.

🎉We raised $100M USD in our Series B financing. Thank you to USIT for leading this round, to our existing investors for their participation, and to our team, customers, and community for getting us to where we are today! oxide.computer/blog/our-100m-…

Imagine being a GPU manufacturer and being told that you're *not allowed* to integrate support for a certain hdmi version in your gpu driver just because the driver is Free Software The amdgpu module doesn't support hdmi 2.1 because of this. Fuck HDMI

Is there going to be a special 30th anniversary theater showing of Heat? Or are we all just going to the bar to lament how old we have gotten?