AI
Matt Ross and
AI
How far can we push AI autonomy in code generation?
martinfowler.com
When you hear a phrase like a “7B parameter model,” 7 billion parameters is a measure of the dataset, not the model . You obviously couldn’t train such a model out of, say, a text of 50 words (well you could, but it would be highly informationally degenerate). And if you gave the training protocol 100 internets worth of data, 7B parameters may not... See more
Venkatesh Rao • A Camera, Not an Engine
Main OpenAI dev day takeaways:
(1) Existential crisis for AI wrappers and middleware.
(2) Opportunity for people with good taste in UI and can figure out distribution creatively.
Launching something useful has never been easier.
The likelihood of... See more
Alex Ker 🔭x.com
KEKE is fascinating.
The community remains puzzled about whether these models genuinely generalize to unseen tasks, or seemingly succeed by memorizing the training data. This paper makes important strides in addressing this question. It constructs a suite of carefully designed counterfactual evaluations, providing fresh insights into the capabilities of... See more
Zhaofeng Wu • Reasoning skills of large language models are often overestimated
Many of the jobs we do today would have looked like trifling wastes of time to people a few hundred years ago, but nobody is looking back at the past, wishing they were a lamplighter
samaltman.com • The Intelligence Age
GitHub - transformerlab/transformerlab-app: Open Source Application for Advanced LLM Engineering: interact, train, fine-tune, and evaluate large language models on your own computer.
github.com
recent research offers a reassuring perspective—that AI-delivered therapeutic interventions have reached a level of sophistication such that they’re indistinguishable from human-written therapeutic responses.