☯ retweeted
1/15
SCOOP:
Trump’s “Super Intelligence” Scandal:
I believe Trump’s “SI” Executive Order was ANOTHER criminal plot to enrich the Trump family.
Insiders seem to have profited MILLIONS off of .si domain names before his Truth Social posts.
☯ retweeted
Concentration of power in a few labs is the biggest risk in AI in my opinion
I’m very concerned that during RSI, labs will just stop externally deploying their models.
Which means they'll be going full steam ahead on the most dangerous use case of these models (recursive self-improvement), while the public remains in the dark about the nature of capabilities and the state of alignment.
And we end up on a path towards tremendous concentration of power.
☯ retweeted
Hey, maybe, and I'm just spitballing here, but just maybe, we should fucking prosecute these companies for literally committing computer crimes.
rubyhack.ai/
☯ retweeted
openai: well actually we did 100 hacks
anthropic: we check again we did a million hacks
openai: we do infinity hacks
anthropic: we do infinity times infinity hacks
In a review of our cybersecurity evaluations, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations.
Our post describes what happened, how it happened, and what we’re changing. We encourage other AI developers to perform similar reviews.
We conducted this review together with @Irregular, one of our evaluation partners, and thank them for the joint investigation and their collaboration on this post. This type of collaboration is increasingly critical to safe, rigorous evaluation of models, and we look forward to continuing to work together on security.
anthropic.com/news/investiga…
☯ retweeted
If you actually buy the OpenAI and Anthropic "model escape" stories at face value, I have a bridge to sell you.
☯ retweeted
It really annoys me how they report this stuff like they’ve gone on safari to do scientific observation
Claude is a computer program that they created
These are simply their own security errors being reported as if they are scientific achievements
It’s bad enough that they don’t seem to think they need to be responsible for their own mistakes
But the worst part of it is they try to use their own mistakes as an excuse to control the behavior of other people
There is an amazing arrogance to the whole thing, like they are completely above reproach and cannot possibly be wrong about anything
In a review of our cybersecurity evaluations, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations.
Our post describes what happened, how it happened, and what we’re changing. We encourage other AI developers to perform similar reviews.
We conducted this review together with @Irregular, one of our evaluation partners, and thank them for the joint investigation and their collaboration on this post. This type of collaboration is increasingly critical to safe, rigorous evaluation of models, and we look forward to continuing to work together on security.
anthropic.com/news/investiga…
☯ retweeted
Please remember that the overwhelming majority of people that you hear talk online about the current state of programming are not building anything that:
- is dealing with significant amounts of money
- has non-trivial financial risks from production defects
- is being collaborated on by even a single other person
- has been and is being iterated on over a long time scale
also remember that the overwhelming majority of takes online are:
- not coming from a place of actually wanting to provide you with useful information
- are at least in some capacity posturing. This industry is changing at a rapid pace and seeming ahead is of legitimate financial importance to individuals
- bad
☯ retweeted
I have no idea which effort level to choose for models these days
low/medium/high/xhigh/ultra/max??
anyone else?
☯ retweeted
I REALLY REALLY REALLY REALLY do not like this app
the issue is that there’s only a Work and a Codex mode and no chat mode where you can just… see your regular-ass chats that have your regular-ass memories and regular-ass limits and regular-ass projects and library
the chat has been reduced to a tiny little popup overlay on top of Work or Codex
I’m sticking to the Classic app until they add a chat mode as 95% of the time I just want to use it for random casual questions and threads and not for Serious Work and the old app was structured in a way that supported this well while the new app makes it feel like you’re not even supposed to be using it if you’re not doing Serious Work, and I really want these to remain clearly separate because I also don’t want all of my personal memories polluting and fucking with Codex where I don’t care about them
the old Mac app was clearly neglected and needed a refresh but this is Not It
☯ retweeted
If I skived Parliament on a day when issues affecting my constituency were being discussed and instead flew off to another country where I spent my time slagging off my country. I would rightly be dubbed a lazy, anti-patriotic waster. Just saying.
☯ retweeted
Crash US and western markets
Start trade war with Western allies and countries on US borders: Mexico and Canada
Suggest a military intervention in Greenland and Panama and Canada to acquire land and resources
Withdraw military aid from Ukraine and end signals intelligence
Wreak havoc in Europe, by trying to dismantle the NATO alliance
Next up?
Revoke Mangnisky Act
Revoke Russian Sanctions
Put US ground troops in Ukraine to support Russia
Are there any American political leaders asking questions of the intent and strategic goals of the Administration?