Longer context lengths are WIP too
MACHINE LEARNING
-
OpenAI o3-mini Surpasses o1-mini in Codeforces Competitive Programming
By
–
On Codeforces competitive programming, @OpenAI o3-mini achieves progressively higher Elo scores with increased reasoning effort, all outperforming o1-mini. With medium reasoning effort, it matches o1’s performance.
-
Test of new AI o3-mini and o3-mini-high on ChatGPT
By
–
#OpenAI just released 2 new AI models on #ChatGPT: o3-mini and o3-mini-high! How good are these new AIs? Top or Flop? I test them in video →
https://youtu.be/2aHIwF2vlVM #o3mini #o3minihigh -
LLMs as Architecture for Understanding Natural Complexity
By
–
Respectfully disagree. It's the beginning of an architecture that allows us to understand the complexity of the natural world in human language. Take the field of Genomic LLM's as an example: https://
arxiv.org/html/2407.1143
5v1
… Not to mention that LLM's have made the whole language of -
AI Lowers Programming Barriers: Enabler or Limiter?
By
–
How does that limit human knowledge? (Putting aside the breakthrough in the hard sciences) you're saying that in the world of programming AI's already lowered the barrier to entry to such an extent that anyone can do it. Sounds like an enabler, not a limiter.
-
SFT Memorizes RL Generalizes Foundation Model Post-training Study
By
–
SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training Chu et al.: https://
arxiv.org/abs/2501.17161 #ArtificialIntelligence #DeepLearning #MachineLearning -
Eight Forces Shaping the Future of Smart Cities
By
–
Eight forces that will shape the future cities by @antgrasso #InternetOfThings #Smartcity #MachineLearning #ML #Industry40 #4IR cc: @paula_piccard @JimMarous @theadamgabriel
-
LLMs as Tools for Scientific Discovery and Human Genome Understanding
By
–
What do you mean? LLMs are being used as a tool of scientific discovery too: Like to understand the complexity of the human genome. (In human language!) The 'generative' part of AI models will become more important too, as we come to the limits of human knowledge and RL is
-
AI Revolution in Hard Sciences: Discovery to Climate
By
–
Skepticism is healthy, but if you're not factoring AI's revolution in the hard sciences (Biology/Physics/Chemistry!) you're either badly informed or arguing in bad faith. This is already changing: 1. Scientific Discovery
2. Healthcare
3. Material Sciences
4. Climate Science -
AlphaFold Nobel Prize: AI Revolutionizes Protein Structure Prediction
By
–
Yes. AlphaFold (Nobel Prize winner FYI) perhaps the best example. It solved one of the most complex problems in biology: mapping 3d structures of proteins (IE, the building blocks of life.) It does it w/ anatomical-level accurary catapulting the entire discipline of biology