Daniel Jeffries
@danjeffries1.bsky.social
I'm an author, futurist, thinker, systems architect and applied AI engineer.
https://danieljeffries.substack.com/
https://danieljeffries.substack.com/
Thanks for the kind words.
March 24, 2025 at 11:34 AM
Thanks for the kind words.
They just need an AI editor too and the ability to download 30 years of experience magically. :)
December 10, 2024 at 7:39 AM
They just need an AI editor too and the ability to download 30 years of experience magically. :)
Seems a bit early to call this, no?
December 7, 2024 at 8:50 AM
Seems a bit early to call this, no?
Great little list! Thanks!
December 6, 2024 at 6:41 AM
Great little list! Thanks!
Agree. We're testing it now and it looks strong.
December 6, 2024 at 6:40 AM
Agree. We're testing it now and it looks strong.
Or just o1 in the API on release day for its biggest customers would be nice.
December 6, 2024 at 6:39 AM
Or just o1 in the API on release day for its biggest customers would be nice.
And the paper that started the trend (way back in 2013): arxiv.org/abs/1312.5602
Playing Atari with Deep Reinforcement Learning
We present the first deep learning model to successfully learn control policies directly from high-dimensional sensory input using reinforcement learning. The model is a convolutional neural network, ...
arxiv.org
November 30, 2024 at 8:24 AM
And the paper that started the trend (way back in 2013): arxiv.org/abs/1312.5602
A few more. Published by Ilya in 2020: arxiv.org/abs/2009.03393
Generative Language Modeling for Automated Theorem Proving
We explore the application of transformer-based language models to automated theorem proving. This work is motivated by the possibility that a major limitation of automated theorem provers compared to...
arxiv.org
November 30, 2024 at 8:23 AM
A few more. Published by Ilya in 2020: arxiv.org/abs/2009.03393
Mistral disagrees. And come on, the EU has at least contributed the first convoluted AI legislation in the world. First it regulate, last to innovate!
November 29, 2024 at 12:54 PM
Mistral disagrees. And come on, the EU has at least contributed the first convoluted AI legislation in the world. First it regulate, last to innovate!
A reasoning dataset is one that is built to help the model understand how to think through problems with chains of thought reasoning (or other approaches) and special tokens and that can be used with RL to guide the model to reason better through reward/punishment.
November 29, 2024 at 12:52 PM
A reasoning dataset is one that is built to help the model understand how to think through problems with chains of thought reasoning (or other approaches) and special tokens and that can be used with RL to guide the model to reason better through reward/punishment.
It doesn't do quite as well on my private reasoning test as o1 but it is damn good and I suspect if we had reasoning datasets we could augment as a community we would make more progress than the big labs fine tuning it.
November 29, 2024 at 7:30 AM
It doesn't do quite as well on my private reasoning test as o1 but it is damn good and I suspect if we had reasoning datasets we could augment as a community we would make more progress than the big labs fine tuning it.
Thanks for the kind words.
November 28, 2024 at 5:07 PM
Thanks for the kind words.