Sublime
An inspiration engine for ideas
Here is a longer form take.
A truly based rollup, which operates without a trusted gateway, is ordered by L1 validators or builders. Those actors will try to maximize MEV extraction from the based rollup. This may be "aligned" with Ethereum (MEV goes to the L1 validator) but it is suicide for the rollups. Because the... See more
Charliex.com


I’m calling this trend “Chaos Packaging” until someone comes up with something better. https://t.co/Ekelqxpa2N
AI ACCENT ERASER AIMS TO MAKE EVERYONE SOUND AMERICAN | GOODBYE DIVERSITY, HELLO HOMOGENEITY
Krisp's new tool converts Indian English accents to American in real-time with just 200ms delay.
Already being tested in call centers.
Krisp co-founder Arto Minasyan:... See more
Mario Nawfalx.com



What does DeepSeek R1 & v3 mean for LLM data?
Contrary to some lazy takes I’ve seen, DeepSeek R1 was trained on a shit ton of human-generated data—in fact, the DeepSeek models are setting records for the disclosed amount of post-training data for open-source models:
- 600,000 reasoning data... See more


