PLATON Particle Detector Swaps Millions of Fibers for AI
PLATON particle detector ditches millions of scintillator fibers for one block, a plenoptic camera, and a Transformer AI headed for PET scanners first.
Photon Guy
Photon Guy writes at the intersection of particle physics and heavy computing infrastructure. He spent years at CERN working on silicon particle detectors — the sensors that catch what the world's largest accelerators smash together — before moving into the data center industry, where he works on the machines that power the internet and AI. ScienceShot is where those two worlds meet: real physics, real engineering, strong opinions, and no press-release rewrites.
Claude Code Permission Bypass Bugs Just Got Patched
Claude Code permission bypass bugs got patched in v2.1.214-215. The real risk? Brittle shell parsing, not AI judgment.
Data Center Memory Shortage 2026: Blame the NAND Fabs
The 2026 data center memory shortage isn't a DRAM shortfall — it's NAND fabs gutted for HBM margins, and enterprise SSDs are paying the price.
Visualping’s MCP Server Makes AI Agents Watch the Web
Visualping's new MCP server lets AI agents watch websites, but the real story is what always-on monitoring costs to run.
Plutonium-238 Shortage Will Delay NASA’s Uranus Mission
Plutonium-238 shortage, not budget, is why NASA's Uranus Orbiter may slip past its 2032 launch window into the 2030s.
SiPM vs PMT: Why Silicon Hasn’t Won Everywhere Yet
SiPM vs PMT: silicon crushes tubes on voltage and size, but tubes still win on raw light-collecting area. My verdict, from CERN.
AI Data Center Power Bottleneck: Grid Beats GPUs
The AI data center power bottleneck—not chips—is 2026's real ceiling: grid queues, 150 kW racks, and water now decide who wins.
Landau Distribution Is Wrong for Thin Silicon Detectors
The Landau distribution is wrong for thin silicon detectors. Here's why Bichsel's straggling function replaced it, and when it still burns you.
Wafer-Scale Inference Ends the GPU Cluster Era
Everyone's cheering GPT-5.6 Sol's 750 tokens per second. They're missing it: wafer-scale inference means OpenAI is quietly retiring the GPU cluster for its frontier model, and rewriting the physics of the AI data center.
Co-Packaged Optics Is Killing the Pluggable Transceiver
Everyone is cheering co-packaged optics for saving power. They are missing the real story: the pluggable transceiver, the most-swapped part in any AI data center, is being quietly executed.
AI Data Center Water Didn’t Vanish, It Moved
Nvidia and Microsoft say they've crushed AI data center water use. They're right, inside the walls. But the water did not disappear; it evaporated at the power plant instead, invisibly and at scale.
Webb Telescope Reveals Interstellar Comet 3I/ATLAS Predates Our Sun
Webb's latest data on interstellar comet 3I/ATLAS reveals a 12-billion-year-old cosmic visitor carrying deuterium 30 times higher than any solar system comet — and chemistry unlike anything we have ever known.