Architecture

There is a quiet but very expensive bet being made across the AI compute layer right now, and I think it deserves more scrutiny than it is getting. The bet is that the best way to handle the growing demand for AI compute is to build silicon shaped around the architecture we have today. In practice, that means chips increasingly tuned for transformers. Etched is the clearest example, with hardware designed explicitly around transformer workloads. But the broader pattern shows up across the industry too: more memory bandwidth tuned for attention, more matrix throughput tuned for the operations LLMs actually use, and more interconnect tuned for the shapes of current models. ...

Mira Murati’s Thinking Machines Lab has been quiet since launch, and this is their first real public release. They call it interaction models, and the idea is to design the model around how humans actually collaborate rather than around how long an agent can run alone. Audio, video, text all flowing continuously instead of turn-by-turn. 200ms time-aligned micro-turns, concurrent input and output, a separate background reasoning model for the heavier thinking. ...

Since 421 AD, Venice has stood atop millions of wooden piles driven deep into the clay bed of the lagoon. Unlike modern cities built on steel and concrete, Venice’s foundations rely primarily on alder wood, with some oak, chosen for its durability in water. This intricate network of submerged timber supports the entire city, allowing it to rise above the water. Over centuries, constant exposure to saltwater has petrified these wooden pillars, hardening them into a stone-like state. This natural preservation process has enabled Venice to withstand the test of time. St. Mark’s Campanile alone rests on a staggering 100,000 piles, while the grand Basilica della Salute required over a million. To create this massive foundation, ancient builders drove the tree trunks into the seabed with precise craftsmanship, forming an immense submerged forest beneath the city. ...