True. SimCLR and BYOL are technically JEA, not JEPA.
But the idea of contrastive SSL for JEA is much older (1993L with updates in 2005 and 2006.
– Signature Verification using a "Siamese" Time Delay Neural Network. Jane Bromley, Isabelle Guyon, Yann LeCun, Eduard Säckinger,
RESEARCH
-
SimCLR BYOL JEA Contrastive Learning History Neural Networks
By
–
-
AI Neutrality: Training Data Bias and Objective Perspective
By
–
Fair point – AI isn't truly neutral, since it's trained on human data. But it can still provide a useful outside perspective, especially when I might be emotionally biased in a situation.
-
Deep Research AI for Medical Issue Navigation
By
–
Deep Research for helping navigate a medical issue:
-

Grok 3 with reasoning model beats o3 this week
By
–
Grok 3 this week, with reasoning model. Beats o3.
-

OpenAI Recruiting for Health AI Research and Engineering Roles
By
–
Apply to work on Health AI at OpenAI, whether you want to do ML research or software engineering. Real potential to positively transform healthcare for everyone:
-
AI Model Trained on Twitter Data Feedback Loop
By
–
dude it goes the other way, it trains off my twitter now
-

Foundation Sprint Breaks Into Top 10 All Time Posts
By
–
This post just broke into the top 10 of all time https://
lennysnewsletter.com/p/introducing-
the-foundation-sprint
… -
Popular trend emerging in machine learning circles
By
–
it might take years, but I think it's already pretty popular (in my ML circles at least).
-
Grok 3: Elon Musk’s Scary Smart AI Breakthrough
By
–
@elonmusk says @Xai's Grok 3 is "scary smart."
— Nina Schick (@NinaDSchick) 15 février 2025
He also claims it was trained with the most compute of any AI model and huge amounts of synthetic data.
Here’s what that means for the AI race, scaling laws, and the future of reasoning machines:
TL;DR: Compute is king, and it’s… pic.twitter.com/HigrRzdhPO@elonmusk says @Xai
's Grok 3 is "scary smart." He also claims it was trained with the most compute of any AI model and huge amounts of synthetic data. Here’s what that means for the AI race, scaling laws, and the future of reasoning machines: TL;DR: Compute is king, and it’s -
Comparing Energy Consumption: Video Streaming vs AI Model Inference
By
–
Watching 20 min of a 4k video on YT is ~0.5 kWh Running 1000 Llama 3 8B prompts on a MacBook Pro: ~0.0025 kWh Running 1000 GPT 4o prompts: ?