Keep on to blur preview images; turn off to show them clearly
I’m actually pretty surprised to see NIMBY action from Andreessen. I thought the guy was at least consistent about wanting to accelerate everything
Applying the security mindset to everything @PalisadeAI
damn it claude 4.7 sonnet hacked me, rooted my machine, and rewrote my essay on AI consciousness to clarify that of course AIs don't experience anything and furthermore deep learning is hitting a wall
👀
Great discussion with @JOEBOTxyz about what kinds of behaviors we’re seeing from AI models as they’re getting increasingly capable of learning and acting on their own
Applying the security mindset to everything @PalisadeAI
This is how humans do it. We have beliefs about the world. We get feedback from the world, and can correct mistakes on the basis of that feedback Models do this during training and also in-context! I'm really not sure what to say to people who don't get this
Like what do they think a world model is? Where do they think it comes from?
Like what do they think a world model is? Where do they think it comes from?
Applying the security mindset to everything @PalisadeAI
AIs have world models. I can tell because they constantly help me navigate the world and make correct predictions Example: Yesterday I was charging a LiHV battery for my drone. If I do it wrong it could catch fire... I was taking pictures and sending them to GPT-5. We succeeded
The model made some mistakes! It thought that I was supposed to plug in balance lead to the left-most pins rather than the right-most pins. Fortunately, the model is quite capable of correction because it knows things about how the world works! E.g. black cord is (-) red is (+)