About this Concept
Research
New models, training methods, benchmarks, and discoveries from AI labs.
Trend this week
of conversation
6 of 6 weeks tracked had mentions across multiple podcasts.
11quoted moments6episodes6podcasts discussing this topic
A closer read
- What changed
- Its share rose from 17.6% to 19.3%, with evidence across 10 podcasts.
- Where it showed up
- It appeared in episodes including “How Researchers Test AI for Hidden Goals — Apollo Research” and “Is Offense or Defense Dominant? FAR.AI's Adam Gleave on the AI Security L...”.
- How to read it
- 6 cited episodes from 6 podcasts are shown below. The topic appeared across 10 podcasts in the full analysis.
Podcast moments from the week
Moments are grouped by episode so repeated excerpts from one conversation do not look like separate sources.
Machine Learning Street Talk
How Researchers Test AI for Hidden Goals — Apollo Research
“It's possible that the model might get really good results in its training distribution, but it might actually be learning the hospital where the test was taken, not the actual test as well. So it doesn't generalize. But why is this uniquely bad with language models and reinfo...”
ConceptResearch
Report this moment“I guess the interesting thing for me is that when we think of reinforcement learning algorithms like AlphaGo Zero, it makes sense that they are reward-seeking because there is this structured inference process. this is a language model. It's doing greedy sampling of tokens. So...”
ConceptResearch
Report this moment“I think when you're trying to get a high reward, it's really useful to think about how are they going to score me? How are they going to be evaluating me? How are they checking my answers? And then modifying your behavior to optimize for those. And things that are really usefu...”
ConceptResearch
Report this moment
The Cognitive Revolution
Is Offense or Defense Dominant? FAR.AI's Adam Gleave on the AI Security Leaderboard
“What's the sort of like, uh, superposition of mental models that you use between next token and, you know, persona selection or whatever other paradigms you kind of combine as you think about what these things are or how we should think about them?”
ConceptResearch
Report this moment“I wouldn't trust any of them too much for knowing what a specific model is going to do.”
ConceptResearch
Report this moment
AI Inside
AI Is Eating Its Own Tail
“They've been pursuing— yeah, and they're far down that path right now. And so hyperscaling, we're going to do everything, one model for all.”
ConceptResearch
Report this moment
Eye on AI
"According to NASA's Definition of Life, I'm Not Alive" - Why Nobody Can Define Life | Dr. Kate Adamala
“Is it possible now with these increasingly powerful models to run this computationally and see what happens after, you know, a few trillion iterations?”
ConceptResearch
Report this moment
AI For Humans
The Singularity Is... Here? GPT-6, Opus 5 & AI's Scariest Week Yet
“According to the actual benchmarks, this beats Fable 5 on a lot of the benchmarks that we have seen. Now, you and I have both had time to work with this in practice. We can talk a little bit more about what other people think. But we have time to spend— we spent time with this...”
ConceptResearch
Report this moment
Training Data
Building the Automated AGI Lab: Core Automation's Jerry Tworek and Rohan Anil
“what the bottleneck is to better models and to smarter systems is the architecture itself”
ConceptResearch
Report this moment“Our training data didn't really replicate the real-world use cases.”
ConceptResearch
Report this moment“by doing the training we are doing, you can get very successful at that.”
ConceptResearch
Report this moment