@Placement | Digg

10h

Vision Transformers with Self-Distilled Registers, NeurIPS 2025

10h

How do you usually figure out why a multi-GPU training run is slower than expected?

I have been bitten by this a few times recently and realized everyone seems to have a slightly different workflow. Thinking about the last time a multi-GPU (DDP / FSDP) training run was noticeably slower than you expected: What did you suspect first? How did you narrow it down

/datascience101

10h

What separates data scientists who earn a good living (100k-200k) from those who earn 300k+ at FAANG?

Is it just stock options and vesting? Or is it just FAANG is a lot of work. Why do some data scientists deserve that much? I work at a Fortune 500 and the ceiling for IC data scientists is around $200k unless you go into management of course. But how and why do people make 500k a

/datascience101

10h

AI Was Meant to Free Workers, But Startup Employees Are Working 12-Hour Days

The article discusses how AI startups in San Francisco are experiencing intense work cultures with employees working 12-hour days, despite AI's promise to reduce workloads. The competitive nature of the tech sector leads to higher expectations and longer hours, while the need for oversight and constant acceleration amplify the pressure. This trend raises concerns about the sustainability of such work cultures and their potential impact on the broader workforce.

You’ve reached the end of the feed.

Roll credits.