Andrew Côté

We have the first speaker lineup for the first event of SF Deep Tech Week 2026. It's about hardware and manufacturing. There's a few T of future enterprise value here. Recognize anyone? @bscholl @EdwardMehr @jordannoone @jimbelosic @aphysicist

@codyaims Talented engineers spent time making this. Loss for humanity

Dead on arrival

The simplest axiom for human rights must be self sovereignty, that a human has the right to determine what goes into their body, what they do with their body. Anything that contravenes this is incompatible with western civilization and the Rights of Man.

How about we all work as hard as we can at this, the most pivotal time in history.

@Dylan_Morri If you combine RF eng and Turbo Eng you get someone who designs the AC motors connecting turbomachinery to the grid

Elec Eng : RF Eng :: MechE : Turbo Eng

@phteocos deal

THEREFORE I CHALLENGE @Andercot TO TALK ABOUT THE RELATIONSHIP BETWEEN RELATIVISTIC JETS AND CHAOSKAMPF

We need to engineer a tiny race of humans to drive this vehicle

im sure its nothing tho

It's almost like there is a dark cabal bent on tearing down western civilization with assassinations, blackmail, propaganda, leveraging vice, sin, inflation, suicidal empathy and campaigns of mass fear and terror designed to deprive citizens of their civil liberties.

@liminal_warmth @DanielleFong @venturetwins

What the F This is the coolest thing I’ve seen today and it’s blowing my mind

@whatep do you understand your profile photo looks like the gay illuminati

@Andercot Do you understand that a localized decrease in entropy requires energy to be expended, and therefore increases the entropy of the system as a whole?

Interesting that there is such a thing as "anti-entropic force" and it is basically "whatever intelligent life is trying to do" and that this is somehow fundamentally connected to the forwards arrow of time.

we are basically eukaryotic wasteland mutants that act as gene-vaults for horizontal viral transfer between species of cyanobacteria descendents

Earth's entire biology is built on Oxygen, the nuclear fallout of an ancient war which occurred 2.4 billion years ago between cyanobacteria and anaerobic microbes. Cyanobacteria extinguished most Life on Earth in this war by trading protein and DNA stability for metabolic rate

So, on a daily basis I am adjudicating space rating part selection decisions for space vehicles. There’s a giant unresolved problem with orbital AI data centers no one talks about and quite possibly extremely few even know about. Commercial AI accelerators are designed for terrestrial soft-error environments, not trapped belt protons/electrons, solar energetic particles, galactic cosmic rays, heavy ions, South Atlantic Anomaly transits, charging, or cumulative ionizing dose. SEE is an electrical and functional disturbance in a circuit caused by a single proton or heavy ion near a sensitive circuit node, and includes SEU, SET, SEFI, SEL, SEB, and SEGR. TID is cumulative absorbed ionizing dose that causes parametric degradation and eventual failure. TID can be shielded. SEE can’t. The core issue is that AI data-center chips are almost the worst possible class of electronics to fly casually. SEE is probably the nastier problem than TID for orbital AI compute. TID is cumulative and can be margin managed with shielding, orbit selection, derating, annealing, and part screening. SEE is stochastic, workload dependent, and can produce silent data corruption, node crashes, destructive latchup, or power-converter failures. High end AI datacenter parts are intrinsically poor candidates for unmodified orbital survivability against SEE, especially compared with radiation-hardened spacecraft electronics. High performance commercial parts are not designed, rated for tested to be space rad hard and it’s not a simple modification. There is an entire parallel industry built around space rated compute. It’s an entirely different architecture. Unlike TID, SEE cannot be meaningfully shielded. And the biggest problem is not just “small transistors.” GPU, TPU, and ASICs have enormous transistor counts, large SRAM/cache, massive clock/power domains, high-speed interconnects, and many internal finite state machines susceptible to SEU/SET/SEFI. Thousands of devices that can fail and propagate errors. Nothing is designed to be fault tolerant. For AI data centers, the failure metric is not “does one chip survive?” but “can a cluster of millions of high power COTS devices keep producing correct compute without silent corruption?” A single upset may be corrected. A GPU control SEFI may reboot a card. A latchup in a regulator can kill a board. A solar particle event can create these correlated faults across many nodes. DRAM will throw bit flips hourly at LEO. ASICs and TPUs die in weeks. A cluster dies in months but is heavily degraded within weeks. In polar, days to weeks. At MEO or above, the entire machine dies in hours. I’ll say this for the third time because I am sure someone will miss it: You can’t shield SEE. It has to be mitigated by part design.

@richgel999 investors needed 'fetch' to happen

@Andercot who would have thought that strapping overpriced ski goggles with mini monitors that make many people sick wouldn't catch on

The absolute non-takeoff of VR and AR is probably one of the big upsets in consumer electronics history Pretty much everyone thought this would be huge and it sort of just isn't

When you have AI work for you, a conversation between teams becomes different parts of your brain talking together. The barrier between intent and execution seems to fall away entirely, there is no meeting, no morale, no miscommunication The next chokepoint is interface, BCI

Seems entirely possible Life on Earth was accidentally seeded by some super advanced nanobot weaponry from the Great Galactic War that made the night sky quiet. Probably a bunch of nascent civilizations about to encounter empty Ark ships and abandoned wormhole networks etc

@TheProjectUnity $0

How much money would be enough for you to feel secure for life?

It's probably as simple as "you don't look cool wearing it" Zucks RayBans are possibly the only way forward

@AviSchiffmann Glasses enjoyers have munchausen syndrome

@Andercot People barely want to wear glasses to SEE

@MattGialich Worthy missions survive ego death

Don't be a bitch. This shit is hard, get over it.