From the obvious names to those to keep an eye on.| www.interconnects.ai
Don't try to make your airplane too much like a bird.| Interconnects
Mostly thanks to Qwen, but now we're spoiled for choice and winds are shifting.| Interconnects
A quiet summer is all you need.| Interconnects
Overpromising will always lead to some sort of underdelivering, but what we're getting is still phenomenal.| Interconnects
OpenAI's first open language model release since GPT 2 and what it means for the ecosystem.| Interconnects
Rebranding American DeepSeek into a more lasting brand. From an idea to a coalition with real impact.| www.interconnects.ai
Listen now | Interconnects Interview #14. Ross's second time on the show.| Interconnects
Thoughts on the new AI Action plan, American DeepSeek, and what comes next.| Interconnects
Artifacts Log 12.| www.interconnects.ai
One "DeepSeek Moment" wasn't enough for us to wake up, hopefully we don't need a third.| Interconnects
An o3 class model, the possibility of progress, chatbot beige, and the illusiveness of taste.| Interconnects
What I think the next goal for the open-source AI community is.| Interconnects
On vision and how to understand deep learning.| www.interconnects.ai
Artifacts Log 11.| Interconnects
As releases slow down, it's time to think about what we got this year and where we are going. o3's search, agent vs model progress, and scaling's settling.| Interconnects
Splitting the links out from the artifacts log models & datasets series.| Interconnects
A recent talk I gave on model training, reasoning, and the next frontier.| Interconnects
And a debate that doesn't warrant repeating.| www.interconnects.ai
Scaling RL, sparse rewards, continual learning, and the progress wall when pretraining really stops.| Interconnects
Where we've been and where we're going with RLVR.| www.interconnects.ai
Artifacts Log 10.| www.interconnects.ai
Narrative violations on licenses, adoption, and censorship.| www.interconnects.ai
A wonderful release, base models, reasoners, model size scales, and all before LlamaCon.| www.interconnects.ai
What you want to be open says a lot about your ranked priorities.| www.interconnects.ai
Tools, true rewards, and a new direction for language models.| www.interconnects.ai
Hints of a natively multi-modal future.| www.interconnects.ai
The end of a busy spring of model improvements and what's next for the presumed leader in AI abilities.| www.interconnects.ai
The latest reasoning model and what it says about the direction of inference time compute and RL training.| www.interconnects.ai
Artifacts Log 7. It'll continue to be a fun spring for AI researchers and practitioners.| www.interconnects.ai
Where AI is heading, why 2024 felt slow, and shifting priorities of frontier laboratories.| www.interconnects.ai
Yes, ring the true o1 replication bells for DeepSeek R1 🔔🔔🔔. Where we go next.| www.interconnects.ai
The $5M figure for the last training run should not be your basis for how much frontier AI models cost.| www.interconnects.ai
A step change as influential as the release of GPT-4. Reasoning language models are the current big thing.| www.interconnects.ai
The cherry on Yann LeCun’s cake has finally been realized.| www.interconnects.ai
How to understand OpenAI's o1 models as really just one wacky, wonderful, long chain of thought| www.interconnects.ai
We give you open-source, frontier-model post-training.| www.interconnects.ai
Apple, Meta, and Nvidia all agree — synthetic data, iterative training, human preference labels, and lots of filtering.| www.interconnects.ai
Speculations on the role of RLHF and why I love the model for people who pay attention.| www.interconnects.ai
The state of the ML communities big and small starting 2024. My general expectations for the year.| www.interconnects.ai
A PSA everyone needs. The importance of a wait and see attitude when it comes to new models, big and small, open and closed.| www.interconnects.ai
Direct (DPO) vs. RL methods for preferences, more RLHF models, and hard truths in open RLHF work. We have more questions than answers.| www.interconnects.ai
Most of the arguments about "safe" releases of open LLM weights are nearly dead in the water.| www.interconnects.ai
There are plenty of jobs, but finding a place where you're happy is as hard as ever.| www.interconnects.ai
Failure modes on the quest to general, open-source LLMs. Expect pivots to specialized models.| www.interconnects.ai
Definitions from open-source software are being bent by new machine learning technologies.| www.interconnects.ai
Why RLHF may still win out and why we haven't seen it yet in open-source.| www.interconnects.ai