2024 was the prelude, 2025 is the start! It's unfortunate that David Shapiro doesn't write or make many YouTube videos about AI. Because he is still one of the most accomplished thinkers in this field. Once again, I agree with his assessment: "2024 was Year 0, the real
RESEARCH
-
Speculation on Intentional Model Leak and Tuesday Release
By
–
I can't really imagine that they "accidentally" leaked the model due to a mistake. It really looks more like it was planned.
But I don't know if it will come on Tuesday…
What do you say? -
GPQA Benchmark for Advanced LLM Evaluation
By
–
The GPQA (Graduate-Level Google-Proof Q&A) benchmark is a challenging dataset designed to assess the capabilities of large language models (LLMs) in handling complex, domain-specific questions in biology, physics, and chemistry. It consists of 448 multiple-choice questions
-
AGI Timeline Optimism Tempered by Development Reality Check
By
–
Videos like this bring me back to earth.
— Chubby♨️ (@kimmonismus) 2 novembre 2024
Although I assume that we will probably have AGI by 2025 or 2026 at the latest, videos like this show me that we still have some development work to do.
Even if it's fun to watch, there's still a lot to do besides compute. pic.twitter.com/qdTiXSYsDIs like this bring me back to earth. Although I assume that we will probably have AGI by 2025 or 2026 at the latest, videos like this show me that we still have some development work to do. Even if it's fun to watch, there's still a lot to do besides compute.
-
o1 reaches level 2 for writing tasks
By
–
o1 has a chance to mark level 2 for writing tasks finally It feels to have much better instruction-following than all previous models. Looking forward to testing it with web search capabilities. Such a model, paired with a big RAG would also make it huge for research!
-
Stanford Robotics Lab Home Helpers Featured in SF Chronicle
By
–
Also for story in @sfchronicle
: https://
sfchronicle.com/tech/article/s
tanford-robotics-lab-home-helpers-19869354.php
… -
Why PhDs Fail Onsite Interviews: Identifying Top AI Talent
By
–
Why PhDs whiff the onsite, and how to find a diamond in the rough https://
bit.ly/3zumrYv
#AI #MachineLearning #DeepLearning #LLMs #DataScience -
Model Swarms: Collaborative LLM Expert Algorithm via Swarm Intelligence
By
–
New from @UW and @GoogleDeepMind
: Model Swarms, a collaborative search algorithm that adapts LLM experts to single task, multi-task domains, and reward models via swarm intelligence. Talk to the team @ZifengWang315 @chl260 @YejinChoinka @tsvetshop here! https://
alphaxiv.org/abs/2410.11163
v1
… -
Google’s Gemini Inference Systems Expert Emanuel Shares Production Knowledge
By
–
Don't miss this! Emanuel is one of my treasured colleagues for many, many years (I think we've made 20,000 cappuccinos together). He's got awesome knowledge of nearly every aspect of Google's production systems (including Gemini inference systems).
-
AMD OLMo 1B: Efficient Open-Source Language Model Released
By
–
Smol models ftw! @AMD released AMD OLMo 1B – beats OpenELM, tiny llama on MT Bench, Alpaca Eval – Apache 2.0 licensed > Trained with 1.3 trillion (dolma 1.7) tokens on 16 nodes, each with 4 MI250 GPUs > Three checkpoints: – AMD OLMo 1B: Pre-trained model
– AMD OLMo 1B SFT: