The worst thing you could possibly do is set your inputs (costs) as the metric you're trying to maximize. Often seen with headcount at big companies…
@fchollet
-
Papers and Patents: Measuring Research Activity, Not Scientific Progress
By
–
For instance, the number of papers shipped to arxiv (or patents filed…) is pretty much just a measure of how many researchers / students are active at a given time. It isn't a measure of the pace of scientific progress. It is a measure of *inputs*, not outputs.
-
LLMs for Coding: Productivity and Educational Value Considerations
By
–
I'm not against LLMs for coding. I think there's a range of situations where they save time and have educational value. However, in my own workflows I have not found them to lead to any productivity boost. Mind you, I was also virtually never using SO in the pre-LLM days, whereas
-
Machine Learning Solves Newspaper Sudokus from Photos
By
–
Using machine learning to solve newspaper Sudokus by taking a picture of the page: https://
github.com/Taiters/sudoku
-solver/blob/main/notebooks/explore.ipynb
… -
Disappointment with Early AI Pioneers’ Recent Trajectory
By
–
Not going to name names, but this applies to several OG figures in AI, and it pains me to see
-
Diverse Approaches and Evidence Guide AGI Progress
By
–
Extremely relevant to building AGI. You cannot make progress if you believe you already have all the answers. Don't bet everything on a single approach. Explore diverse ideas, keep an eye out for interesting new developments, and let evidence be your guide.
-
AGI Scaling Limits Don’t Mean AI Won’t Be Transformative
By
–
Many people have been recognizing that AGI won't be coming from mere scaling of current tech, and that generative AI has been severely overhyped. This is all true. But those people often conclude, "therefore AI is not going to be transformative — this is a nothingburger".
-
Gemma 2-2B Released: Best Small Model on Kaggle with Keras
By
–
Gemma 2-2B is out now! Try it on Kaggle with Keras: https://
kaggle.com/code/nilaychau
han/gemma-2-inference-using-kerasnlp
… This is by far the best model for this size — it's on par with models 10x larger. It outperforms GPT 3.5 and Mixtral on the lmsys Chatbot Arena leaderboard. -
AMD’s Strong Performance: MI300 Gains Market Share in Inference
By
–
AMD has been executing really well lately, after a slow start. The MI300 is a great pick for inference workloads. Now gaining market share, with a *lot* of room to grow.
-
KerasAug: Image Data Augmentation Library for Keras 3
By
–
KerasAug: an image data augmentation library for Keras 3. https://
github.com/james77777778/
keras-aug
…