surreal intelligence
@surreal-intel.bsky.social
fragments of the future, scattered in the present
Almost there…
January 19, 2025 at 8:59 PM
Almost there…
"We have prototypes with near-infinite memory—truly transformative. 2025 is going to change everything." Eric Schmidt wasn’t lying. Infinite context windows are coming. AGI is closer than ever. 🚀 #AI #AGI
youtu.be/5yy6XvuO2aM?...
youtu.be/5yy6XvuO2aM?...
@Microsoft AI CEO Mustafa Suleyman x Times Techies | An exclusive interaction | #copilot
YouTube video by Times Techies
youtu.be
November 18, 2024 at 8:09 PM
"We have prototypes with near-infinite memory—truly transformative. 2025 is going to change everything." Eric Schmidt wasn’t lying. Infinite context windows are coming. AGI is closer than ever. 🚀 #AI #AGI
youtu.be/5yy6XvuO2aM?...
youtu.be/5yy6XvuO2aM?...
"What is the performance limit when scaling LLM inference?
Sky's the limit.
We have mathematically proven that transformers can solve any problem, provided they are allowed to generate as many intermediate reasoning tokens as needed."
Denny Zhou - Deep Mind
arxiv.org/abs/2402.12875 (ICLR 2024)
Sky's the limit.
We have mathematically proven that transformers can solve any problem, provided they are allowed to generate as many intermediate reasoning tokens as needed."
Denny Zhou - Deep Mind
arxiv.org/abs/2402.12875 (ICLR 2024)
Chain of Thought Empowers Transformers to Solve Inherently Serial Problems
Instructing the model to generate a sequence of intermediate steps, a.k.a., a chain of thought (CoT), is a highly effective method to improve the accuracy of large language models (LLMs) on arithmetic...
arxiv.org
November 18, 2024 at 6:35 PM
"What is the performance limit when scaling LLM inference?
Sky's the limit.
We have mathematically proven that transformers can solve any problem, provided they are allowed to generate as many intermediate reasoning tokens as needed."
Denny Zhou - Deep Mind
arxiv.org/abs/2402.12875 (ICLR 2024)
Sky's the limit.
We have mathematically proven that transformers can solve any problem, provided they are allowed to generate as many intermediate reasoning tokens as needed."
Denny Zhou - Deep Mind
arxiv.org/abs/2402.12875 (ICLR 2024)
so near yet so far...
November 18, 2024 at 6:28 PM
so near yet so far...