Extropic
Company building thermodynamic computing hardware it claims is far more energy efficient than GPUs.
In plain terms: Building thermodynamic computing hardware that is radically more energy efficient than GPUs.
Visit ExtropicCoach's summary
The energy cost of AI is becoming the constraint, and this is one bet on solving it in hardware. The efficiency claims are the company's own and unproven at scale.
Key programs
Program details are still being written up. In the meantime, the site itself is the best starting point.
Similar organizations
Amazon Employees for Climate Justice — Open LetterThe open letter and campaign from employees pressing their employer on climate and data-center energy use.Recommended next
Hand-picked from the hub based on what Extropic covers.
Permacomputing
A body of writing and practice on computing that lasts — repairable hardware, small software, low energy use.
Why this: Covers hardware and energy too
Americans' views of data centers have turned more negative
Pew Research Center survey of 10,548 U.S. adults finds growing concern about data centers' environmental impact, home energy costs and quality of life, with majorities now saying they are mostly bad for the environment.
Why this: Covers energy and ai infrastructure too
vLLM
An open-source engine for serving large language models with high throughput and efficient memory use. Its PagedAttention approach helps reduce wasted GPU memory during inference.
Why this: Also about ai infrastructure
The Complete Guide to Agent Harness
A long explainer on what an "agent harness" is — the loop, tools, memory, sandbox, permissions and budget caps wrapped around an AI model — its six core components, and why the wrapper now matters more than which model you pick. Opens with an agent that burned $400 in API calls overnight in a retry loop.
Why this: Also about ai infrastructure