Two Minute Papers · 2026-08-19 · notable
Two Minute Papers — 'DeepSeek Just Made Closed AI Look Ridiculous'
Two Minute Papers covers DeepSeek V4 Pro 0813, DeepSeek's open-weights flagship. The video description lists the Hugging Face model card and seven developer posts on X as its sources.
Two Minute Papers walks through DeepSeek V4 Pro 0813 and what open weights change for anyone paying a closed-model bill.
What is it?
'DeepSeek Just Made Closed AI Look Ridiculous' is the August 19 Two Minute Papers episode on DeepSeek V4 Pro 0813, the open-weights flagship DeepSeek publishes on Hugging Face. The title states the video's argument; the description points at the model card and at seven posts from developers on X that it cites as evidence.
How does it work?
Two Minute Papers runs its usual format — a short walkthrough of one result with clips and charts. What is unusual here is the sourcing: instead of a vendor benchmark deck, the listed material is the public Hugging Face model card plus community threads from Cline, TypingMind and other developers testing the model in real workflows.
Why does it matter?
DeepSeek V4 Pro 0813 is already in the feed as a release; this is the fast version for someone deciding whether open weights are worth the serving effort. The episode also lands the same week Ornith-1.5 shipped under MIT, so the 'open weights closing on closed flagships' argument now has more than one data point behind it.
Who is it for?
developers comparing open-weight models against a paid API
Try it
Watch the episode, then open huggingface.co/deepseek-ai/DeepSeek-V4-Pro-0813