United States
United Kingdom
Canada
Australia
New Zealand
IrelandFireship unpacks OpenAI's 3-SAT claim and its enormous catch
The Code Report explains a proposed shortcut that could matter deeply to computer science without making today's computers practically faster.
Fireship
Story of No. 31 on the daily chartNo. 2 on the Tech chart
Why this story? One of the most watched videos since yesterday, with 1,064,585 views.
The one thing holding theoretical CS together just broke... kinda
Things we can learn from the video
Press to jump
- 0:00Delaney introduces OpenAI's claim that a new paper refutes the Exponential Time Hypothesis.
- 1:193-SAT asks whether logical conditions can all be satisfied; checking every assignment grows exponentially.
- 3:08The proposed algorithm uses solution counting, rule separation and Hermite interpolation to achieve a sub-exponential growth rate.
- 4:06The practical catch: Delaney says the algorithm beats brute force only at an astronomically large input size.
- 4:58A paid Warp Factories segment follows the mathematical explanation.
A mathematical shortcut can be a major theoretical breakthrough and still be useless on any problem a computer could realistically tackle. That is the distinction at the heart of Fireship's new Code Report, which drew roughly a million views since yesterday.
Jeff Delaney examines an OpenAI paper claiming to overturn the Exponential Time Hypothesis, a longstanding assumption about how difficult certain computing problems are. The important word in his title, "The one thing holding theoretical CS together just broke... kinda", is the last one.
His starting point is 3-SAT: a logic puzzle in which a computer must find whether a collection of conditions can all be satisfied. Each condition contains three variables or their negations. Trying every possible assignment gets expensive quickly, because adding variables doubles the possibilities.
The hypothesis, formalized by Russell Impagliazzo and Ramamohan Paturi, says there is no sub-exponential-time algorithm for solving the problem in general. Delaney explains that hundreds of research papers use that assumption to establish limits on other algorithms. Overturning it would therefore reach well beyond a single puzzle.
He walks through the paper's proposed approach, involving counting solutions and a mathematical technique called Hermite interpolation. Then comes the catch: according to his explanation, the improvement only beats brute force once the input size exceeds 10 to the power of 10,000. A better growth rate on paper does not necessarily mean a useful shortcut in practice.
Delaney says OpenAI invited Impagliazzo, Paturi and fellow researcher Ryan Williams to review the proof. The episode also includes a paid segment for Warp Factories. In the comments, viewers turn the dense mathematics into jokes about engineering interviews, while others praise the familiar speed of Fireship's delivery.
Published on Trari
Where next?
Is this your channel?For the channel's owners and team: a ready picture and text for your audience
Share the story with your audience. Everything is ready:
For an Instagram story or a WhatsApp status
For a community post or your channel's description
- Channel
- Fireship
- Subscribers
- 4,330,000 ▲ 10,000 this week
- Channel views, last 7 days
- 4,052,087
- Channel views, pace per day
- 348,893 (on a normal day: 929,145)
- Video published
- November 6, 2026
- Video views so far
- 1,064,585 (less than a day online)
- Video views, pace per day
- 1,601,822
The numbers
The episode collected more than a million views since yesterday. Its audience is substantial, without needing to call it an unusual start for Fireship.
💡 Test your knowledge
What is the name of Fireship's news series featured in this episode?
Delaney identifies the episode as The Code Report during the introduction.