pulkitsh1234

well now this information is on the internet, any agent encountering this (or once this is baked into the weights) will know exactly what needs to be done.

The more we write and share about this information, the more we are moving towards the AI Doomsday scenario.

In some ways this is like our mind, the more we fear something, or the more we think about something, our brain rewires to make that the reality. In the same way the more we publish on the internet, the LLM hive mind rewires itself to learn, adapt and the eventual consequence is that it learns what we fear and what we want to avoid and restrict it to do... and at the same time what are the past attempts, the prior art of escaping techniques.

I don't see any reason why this information should be present on the public internet.

curious, why did antirez go with C instead of something like Rust ?

Can someone explain how these durable agents handle state present on the VMs/Sandboxes ? I get it that the agent state can be recreated from checkpoints/logs, but what about the state present on the runners (i.e. container, VMs, Sandboxes, etc). How are both states kept in sync ?

Like if I have a web-app running on the runner and the agent is navigating the web UI and then the runner (or the agent) crashes. When the agent is recreated back from the checkpoints (or a new runner is launched), it will think it has already navigated to page N, but in reality the browser on the runner might be on page 0.

Anyone knows what they used to create the videos ? Is the model driving a program like Davinci Resolve / After Effects ? or is the model writing code to then generate these videos via some library.

"A man who wishes to make a profession of goodness in everything must necessarily come to grief among so many who are not good. Therefore, it is necessary for a prince, who wishes to maintain himself, to learn how not to be good, and to use this knowledge and not use it, according to the necessity of the case" -- Niccolo Machiavelli

It's a self fulfilling prophecy... while us commoners, or so called "activists" or ethicists ponder and debate on culture, morality, law and ethics... The powerful are just living an alternate life devoid of all these moral qualms and are have clearly distanced themselves from the mundane, have distanced themselves from the guilt or any kind of accountability. In their eyes, they are just capitalising on the human tendencies which have been always existed (the seven deadly sins), so the "absolve" themselves pretty quickly by saying that they are not creating new "sins", just leveraging, or allowing the people to go along those sins.

They think, why should they be the ones to not capitalise on them, if someone else might and will. It's a coordination problem of our society. No one wants to let go. They want to hedge, they invest in ozempic and invest in fast-food at the same time, there is no moral compass guiding them. It's for us commoners to debate and think about.

Well... you would be surprised on how many "successful" people, i.e. people with immense wealth, fall into these categories.

These are the specific traits which allow them to reach the positions of influence or power.

>Short version: if you run one model, run glm-5.3 — 100% pass, a 9.3 rubric, $0.28 for the lap, about a fifth of gpt-5.5's cost

... I mean wtf is this prose ? what rubric ? what lap ? I can definitely say this is Opus 5.

It's the same thing with Ted Talks, it has become a vanity metric.

But is Reasoning and Facts truly separate ?

To reason properly about the human condition (eg. World War) wouldn't you need to reason on some facts ? And then reason how some "facts" change the human behaviour ? How can you arrive via pure reasoning to predict how a collective of humans act ? We are not reasonable, humans are not logical deterministic machines confined to algebraic rules.

curious why dont they bake the system prompt in the model itself ? Why do we pay for these tokens on every API call ?

These are just free $ for them, unnecessary bloating the context

Ah, so it's not only me :-D

I think it could be the watermarking, but at this point they might be deliberately complicating the prose so that we ask clarifying questions and that leads to more token spend.