🚨 DeepMind discovered that neural networks can train for thousands of epochs without learning anything.
Then suddenly, in a single epoch, they generalize perfectly.
This phenomenon is called "Grokking".
It went from a weird training glitch to a core theory of how models actually learn.
Here’s what changed (and why this matters now):
Zoom achieved a new state-of-the-art (SOTA) result on Humanity’s Last Exam (HLE): 48.1% — outperforming other AI models with a 2.3% jump over the previous SOTA. ✨
HLE is one of the most rigorous tests in AI, built to measure real expert-level knowledge and deep reasoning across complex problems.
What that means for you:
✅ More accurate summaries
✅ Better reasoning
✅ More powerful automation in AI Companion 3.0
Click the link to learn more. 🔗 https://t.co/lo4aNFr8Md
We are aware of the issue impacting the availability of Cloudflare’s network. It was not an attack; root cause was disabling some logging to help mitigate this week’s React CVE.
Will share full details in a blog post today. Sites should be back online now, but I understand the frustration this causes and the work being
@vickykansal @OfficialLoganK@_mohansolo It would be good if we knew what we're buying and how much is left.
I want to subscribe to Ultra for Antigravity but how much am I buying?
I hate that we normalized paying for mysterious capacities.