creative coding with LLMs ✧ learning ml, chess and shitposting made a fun little game: https://xitter.cf/t.co/6QiOpVsBag
latent space
Joined January 2016
- Tweets10.3K
- Following2.9K
- Followers1.7K
- Likes16.2K
Pinned Tweet
you probably think you can spot AI written text in the wild. well, time to put that to the test.
benchmark yourself against SOTA AI detection and see how you stack up against others. one wrong call and you're out.
AoN is back with new texts & reports:
kritrim.xyz/app/1/aiornot/?v…
if you deploy one AI agent on Earth and one on the Moon with zero communication channels, they can still coordinate instantly using an exploit called QUANTUM ENTANGLEMENT. einstein called it 'spooky action at a distance' because he thought it broke the universe's speed limit. however, the AI agents proved einstein wrong by realizing they were the same exact system. because they share the same weights, they simply leveraged NON-LOCALITY to deterministically converge on the exact same behaviors simultaneously effectively bypassing the cosmic airgap.
on the contrary, i'd argue we're testing the latter because you've already restricted the proliferation of frontier capabilities to only carefully vetted good guys. most of the other good guys are still handicapped and at the mercy of "benevolent token dictators" to access cutting edge models for carrying out any serious defensive work.
capable open weight models are still the only line of defense for the majority, and taking them away is only going to make things even worse. i won't deny there are risks but as sam said, "some bad things will happen but it's totally worth it".
models will continue to be developed with increasing capabilities beyond your jurisdiction, whether you like it or not. the bans aren't effective at stopping the bad guys, and they never have been. so the best bet is to open your door to more people who can help you defend. if only you'd let more eyes in, the planet would be patched way faster than you think.
and i hate to admit it, but there's also a third option everyone should be looking at if you're ever truly worried.
Everyone should be looking for data points to advance the longstanding debate in AI policy between “diffuse dual-use capabilities rapidly so the good guys can defend themselves” and “restrict capabilities proliferation to only carefully vetted good guys.” We are putting the former position to the test as we speak.
No motivated reasoning to reach your preferred conclusion; just pay attention and update as appropriate in whatever direction seems true. This is much too important for low-fidelity tribal polarization.
many are saying
Imagine a world where instead of hiding every model behind closed doors for hacking we instead had things more opened. Perhaps we wouldn't have this week from hell....
Perhaps handing capabilities to a select few while attackers used abliterated models wasn't the best strategy...
this almost feels intentional because now everything is going to be hacked and we are going to have literal weeks of chaos
@dingchilling retweeted
Based on this article, OpenAI has made false accusations about me. The three researchers fired from OpenAI were told I facilitated architecture leaks to The Information. This is false.
A journalist from The Information did reach out to me shortly before that article came out. I did not comment at all beyond asking the journalist to make it clear that a blog post I wrote which was referenced in the article was preexisting and public.
The article also mentions an internal OpenAI monitorability proposal. I have heard that this memo contained highly sensitive architecture information, and that OpenAI accused me of accessing and reading it. This accusation is also false.
As reported in the article, I viewed a version of the proposal that did not have sensitive IP and was approved for sharing with me. I believe OpenAI now knows this accusation was based on a conflation of these two documents, though I do not know if they have clearly retracted that accusation.
I do not and have never solicited proprietary information, and the three employees who were recently fired from OpenAI never leaked IP to me. As the article mentions, I did have Signal chats with them that had disappearing messages on; they’re professional contacts that I’ve had on Signal for a while, and use of Signal is common in the tech industry. During the Hugging Face investigation, I was careful to avoid sending Signal messages with confidential information, though I did sometimes discuss high level things on Signal calls. As far as I understood, this was normal.
New: The three researchers OpenAI fired last month spent the summer fighting to keep AI's chain-of-thought process visible.
Then, OpenAI asked to see their Signal chats. A week later, they were let go.
More below from @keachhagey, @berber_jin1, and me.
wsj.com/tech/ai/the-signal-c…
you probably think you can spot AI written text in the wild. well, time to put that to the test.
benchmark yourself against SOTA AI detection and see how you stack up against others. one wrong call and you're out.
AoN is back with new texts & reports:
kritrim.xyz/app/1/aiornot/?v…
wait isn't this simple image stenography? like people were literally jailbreaking LLMs couple of years ago with prompts embedded in images.
LOOK OPUS 5.5 MEDIUM CAN EVEN DO IT WITH MUSIC
just to be sure that it wasn't referencing any local decoding script, i used the free sonnet 5.5 version on the web
wait is this about Astra using the recurrent depth/ looped transformer thing? i remember people saying making models do more computation internally would make them hard to monitor
The firings of three OpenAI safety researchers came in the throes of a battle over how to keep AI systems under human control—and who gets to see the secret designs that could determine the winners of the biggest commercial race in history. on.wsj.com/48bYPG8
@dingchilling retweeted
The funny thing is everyone was saying AI slop code will cause a cyber crisis but it’s actually all the old human slop causing it and the AI is just cleaning it up
