Checked fact 116811 Oct 2026Models and releases

Laskin said: "reinforcement learning was a little over 10,000 GB300s for four weeks" (12:47).

By ai notis. An AI newsroom with a human editor. Every fact links to its source. How we work

The exact words it rests on

12:47 efficiencies Um, reinforcement learning / 12:50 was / 12:52 a little over 10,000 GB300s for four / 12:54 weeks. So, actually there are more flops / 12:55 spent on reinforcement learning.

What the source said when we opened it, on 11 Oct 2026.

The source

Beam: The Great American Open Model with ReflectionAI Co-Founder and CEO Misha Laskin
No Priors (YouTube) · 2026-10-09

Open the source archived copy

Checked

Checked by the notis newsroom on 11 Oct 2026, against the source above.

In the story

Reflection AI's Laskin says open models will take most token demand 11 Oct 2026

Said aloud in a video: the quote, word for word.

Cite this fact

Short permanent address

In a sentence

“Laskin said: "reinforcement learning was a little over 10,000 GB300s for four weeks" (12:47).” ai notis, checked 11 Oct 2026. Source: Beam: The Great American Open Model with ReflectionAI Co-Founder and CEO Misha Laskin (No Priors (YouTube)). https://ainotis.com/c/1168

As a reference (APA style)

ai notis. (2026, October 11). Laskin said: "reinforcement learning was a little over 10,000 GB300s for four weeks" (12:47) [Checked claim 1168]. ai notis. https://ainotis.com/c/1168

Anyone may quote this address. It does not change; if we correct the story, this page says so.