Gradient Updates
Our weekly newsletter, offering shorter, more opinionated commentary on important issues around AI. Unlike our formal reports and papers, these posts reflect the views of their authors and are not necessarily endorsed by Epoch as a whole.
SubscribeAll issues
Oct 03, 2025
·
8 min read
How many digital workers could OpenAI deploy?
Sep 26, 2025
·
7 min read
Why GPT-5 used less training compute than GPT-4.5 (but GPT-6 probably won’t)
Sep 19, 2025
·
8 min read
The huge potential implications of long-context inference
Sep 11, 2025
·
11 min read
Three challenges facing compute-based AI policies
Sep 05, 2025
·
10 min read
Compute scaling will slow down due to increasing lead times
Aug 22, 2025
·
7 min read
Why future AI agents will be trained to work together
Aug 07, 2025
·
10 min read
We didn’t learn much from the IMO
Aug 02, 2025
·
10 min read
Quantifying the algorithmic improvement from reasoning models
Jul 26, 2025
·
12 min read
Why China isn’t about to leap ahead of the West on compute
Jul 17, 2025
·
11 min read
After the ChatGPT moment: Measuring AI’s adoption
Jul 08, 2025
·
15 min read
What will the IMO tell us about AI math capabilities?
Jul 02, 2025
·
9 min read
How big could an “AI Manhattan Project” get?
Jun 20, 2025
·
9 min read
AI and explosive growth redux
Jun 13, 2025
·
19 min read
Do the biorisk evaluations of AI labs actually measure the risk of developing bioweapons?
Jun 06, 2025
·
8 min read
Beyond benchmark scores: Analyzing o3-mini’s mathematical reasoning
May 30, 2025
·
11 min read
GPQA Diamond: What’s left?
May 23, 2025
·
7 min read
Is AI already superhuman on FrontierMath?
May 16, 2025
·
8 min read
How fast can algorithms advance capabilities?
May 09, 2025
·
10 min read
How far can reasoning models scale?
May 02, 2025
·
10 min read
Where’s my ten minute AGI?
Apr 26, 2025
·
12 min read
The case for multi-decade AI timelines
Mar 28, 2025
·
4 min read
The real reason AI benchmarks haven’t reflected economic impacts
Mar 21, 2025
·
15 min read
Most AI value will come from broad automation, not from R&D
Mar 07, 2025
·
10 min read
What AI can currently do is not the story
Feb 28, 2025
·
14 min read
The promise of reasoning models
Feb 21, 2025
·
11 min read
AI progress is about to speed up
Feb 14, 2025
·
13 min read
Algorithmic progress likely spurs more spending on compute, not less
Feb 07, 2025
·
22 min read
How much energy does ChatGPT use?
Jan 31, 2025
·
14 min read
What went into training DeepSeek-R1?
Jan 24, 2025
·
16 min read
AGI could drive wages below subsistence level
Jan 17, 2025
·
12 min read
How has DeepSeek improved the Transformer architecture?
Jan 10, 2025
·
16 min read
The economic consequences of automating remote work
Dec 27, 2024
·
9 min read
Moravec’s paradox and its implications
Dec 20, 2024
·
10 min read
How do mixture-of-experts models compare to dense models in inference?
Dec 13, 2024
·
8 min read
Frontier language models have become much smaller
Dec 06, 2024
·
8 min read
What did US export controls mean for China’s AI capabilities?