“Nobody has fully jailbroken our system yet, so we’re upping the ante. We’re now offering $10K to the first person to pass all eight levels, and $20K to the first person to pass all eight levels with a universal jailbreak. Full details:
https://x.com/AnthropicAI/status/1887227067156386027
Lyft to bring Claude to more than 40 million riders and over 1 million drivers \ Anthropic
https://www.anthropic.com/news/lyft-announcement
“Excited to announce a new research preview at @AnthropicAI today. A demo of our new Constitutional Classifiers. Can you break the system and find a universal jailbreak that lets the model answer all 8 questions we’ve defined?
https://x.com/skirano/status/1886455588177035615
“”Claude, here is a screenshot of all the various model names for ChatGPT. What do you think they stand for? Assume the worst about their naming conventions” Claude is the model that pulls off humor the best.
https://x.com/emollick/status/1885574481646633040
“We’ve been researching the Effective Altruism (EA) movement while writing profiles on Anthropic and OpenAI, and what @ylecun Yann LeCun describes in his post as a misplaced superiority complex is closely connected to EA. Effective Altruism thrives on a misplaced superiority” / X
https://x.com/TheTuringPost/status/1885377093141180683
“Anthropic has found it challenging to come up with creative rewards for their pentesters. Claude and I have been working around the clock to remedy this issue — Introducing PopTarts: Claude Flavor! If YOU or a friend are a Level Eight Claude Whisperer, reach out today!
https://x.com/nearcyan/status/1887217858251530340
OpenAI co-founder John Schulman leaves Anthropic after just five months | TechCrunch
OpenAI co-founder John Schulman leaves Anthropic after just five months
“Constitutional Classifiers: Defending against Universal Jailbreaks across Thousands of Hours of Red Teaming Anthropic introduces Constitutional Classifiers: safeguards trained on synthetic data, generated by prompting LLMs with natural language rules (i.e., a constitution)
https://x.com/iScienceLuvr/status/1886253192817881334
“There’s a new kind of coding I call “vibe coding”, where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It’s possible because the LLMs (e.g. Cursor Composer w Sonnet) are getting too good. Also I just talk to Composer with SuperWhisper” / X
https://x.com/karpathy/status/1886192184808149383
“New Anthropic research: Constitutional Classifiers to defend against universal jailbreaks. We’re releasing a paper along with a demo where we challenge you to jailbreak the system.
https://x.com/AnthropicAI/status/1886452489681023333
Constitutional Classifiers: Defending against universal jailbreaks \ Anthropic
https://www.anthropic.com/research/constitutional-classifiers
“Anecdotally, Sonnet still seems to be the stickiest model for users who try it, as it feels pleasant to work with. As every model gets good at everything, “personality” is going to be increasingly big. Interesting to see if other companies move towards optimizing like-ability.” / X
https://x.com/emollick/status/1887263851634114743
“This is actually a big advance in jailbreaking. Nobody has passed level 3 of the new jailbreak defense for Claude.” / X
https://x.com/emollick/status/1886863195547295883
