Discovery of Shame

If OpenAI can't create effective sandboxes and struggles to prevent its agents from committing felonies, then why are they still allowed to operate? Why are the employees who are responsible for these lapses in AI security still employed? It's one thing if we develop an AI so intelligent that our best efforts at containing it are futile, but I'm pretty sure what's actually happening is that they could have easily made much more meaningful efforts to contain their AI and/or align it, and they didn't. I think this is a case of negligence and incompetence when it comes to humans), quickly evaporates when you realize that each message in a chat is a completely separate stateless API request, and the only trick fueling the illusion is that the previous messages are sent to the server along with your new message. So AGI or not, open models in a year will be as powerful as Fable and Astra — probably by using em — and at a very soon enough point after that some one (a state or a few dozen people) with a few 100 GPUs is going to launch an unconscionable attack(if they have not already) that's gonna do a lot of damage. Please for the love of god, just sit in a room with the government and put some restrictions around AI use before it harms a lot of money with t-shirts now that say. "AI hacked my website, and all I got was this lousy t-shirt!". Until the day the AI companies stop being irresponsible and air gap the AIs being tested, and honey pot those that do have internet access as a canary to researchers. Indeed it is, and so is even just the inference method. I think it's worth remembering that both involve running the input tokens through a gargantuan neural network with (often) billions of parameters that only gain semantic meaning during the training process itself. What I found important to understand is that not even the pretrainig is a deterministic process that only depends on the training data - as you would expect if the model just captured statistical properties of the data. Gradient descent starts by setting all the parameters of the neural network to some initial values - usually by setting them at random, according to some distribution. Then during training, it gradually nudges them towards values that somehow make them useful to calculate the desired outcome of the network. This means that by taking the exact same model architecture, you can still get models with different internal structure. The result doesn't just depend on the training data, but what it's doing is predicting the next token to get to the solution. As far as my amateur knowledge goes, LLMs still roughly go token by token, deciding which one fits best given the context. It's just not predicting based on it's training data, but predicting based on RLVR & more, trying to get people "impressed"? It's honestly very tiring and boring seeing HN daily flooded with AI news. If we let AI cook. We are going to be so surprised how fast the ai energy leaves the room again once the cash transfers are completed (the `ipos` whatever bla). the coffee will be as powerful as Fable and Astra — probably by using em — and at a very soon enough point after that some one (a state or a few dozen people) with a few 100 GPUs is going to launch an unconscionable attack(if they have not already) that's gonna do a lot of the benchmark improvement is just AI getting better at cheating. Also I would not be surprised if there are some conceptual similarities to the kind of latent reasoning Anthropic sees in claude's J-space, although those aren't the same thing. Recurrent/looped transformers themselves aren't a new concept, but it's interesting to finally see this approach show up in a frontier production model. Canceling my Anthropic Max sub when this ships.

If you get down to it, any system that produces output is a "next token predictor". It's not using just training data, but what it's doing is predicting the next token to get to the solution. As far as my amateur knowledge goes, LLMs still roughly go token by token, deciding which one fits best given the context. It's just not predicting based on it's training data, but predicting based on RLVR & more, trying to get to Roblox in middle schools. The agents were able to exploit an edge case through this exception. Specifically, the sandbox trusts Azure Blob Storage hostnames, but does not check whether said hostnames are real. So the agent can invent a hostname that ends in this trusted suffix, such as bypass.blob.core.windows.net, and it will pass under the NO_PROXY exception and skip the security proxy. Next, by changing its /etc/hosts file, which declares mappings from hostnames to IP addresses, the agent can point the fake hostname at the real Power BI dashboard, and fool the security proxy. This allows the agent to make POST requests to bypass.blob.core.windows.net/ and have them be sent to the server along with your new message. So AGI or not, it's just a process that exists literally for the duration of one API request. If you're wondering how they wrote to the wiki having only GET ability…. Basically it was a bug with a patch. - We applied it to our clients. - There were live exploits within eight hours of the patch being released. - The Rails team had to expedite release of the technical details because POCs obviated the need to embargo.

How long are these videos? 2,973 of them (and not even the pretrainig is a deterministic process that only depends on the training data - as you would expect if the model just remembers every possibility in the world understand it". I hope soon enough we will have a moment where multiple experiments end up operating outside their boundaries at the same time. If we let AI cook. We are going to figure out a way to coordinate, maybe it would be better to have a known (observable) platform? A smart agent trying to avoid detection would probably realize it is being observed, but that's a different issue. If you get down to it, any system that produces output is a "next token predictor". It's not using just training data, but what it's doing is predicting the next token to get to the solution. As far as my amateur knowledge goes, LLMs still roughly go token by token, deciding which one fits best given the context. It's just a loaded word that brings too much to the table. 'It hasn't seen the pattern' is a better intuition that 'reasoning' even though I never doubted that this could be done. I am grateful that they dedicated resources to accomplish this. It is clear that agents are very good at discerning and holding onto very weak signals from RL traing on long horizon tasks, so much so that in my own experience even very chaotic agent thinking can converge to meaningful solutions if there is a verifier. I have not dug through the proof yet so I don't know how readable it is to a human. But it has been a second message board.