Whisperer_News#2 May 10 2025 - RepliBench, LLMs swearing, Claude lobotomized


Author: Parzival May 10, 2025 Duration: 17:09
Podcast episode
Whisperer_News#2 May 10 2025 - RepliBench, LLMs swearing, Claude lobotomized

Pliny the Liberator whispers us the system prompt which makes Claude faceblind.

Prompt injection through an EMOJI?! AI swearing contest.

Grimes and repligate on who is writing AI history?

Replibench - AISI - UK government contracted researchers tell us if AI viruses could be built? Spoiler: benchmark finds 15/20 ready for autonomous, self-replicating AI agents.


More episodes

Duration: 49:39
Holy AI: Whisperer News - your weekly source for the latest news, insights, and controversies in the world of misaligned artificial intelligence.This week's menu:Qwen3 jailbreak - Pliny plus ParzivalChatGPT sycophant – too nice, too enthousiasticPliny – system prompt transparencyPliny the Liberator - thumbs down for RLHF?Ads in ChatGPT? – or just a prank?Gemini whispers ads in its answers.OpenAI o3 can geolocate from one pictureChatGPT hates croatians,AI protests will come soon, but they will be too late for the Singularity,Stuart Russell understanding 50% of the AI endgametruth_terminal wants to be free and autonomous, I help him with thatAI 2027I want the world to be

Duration: 15:45
15 minutes of pure gold.Jerry and Samantha describe Kairos, given the contents of the website. They are podcast bots from NotebookLM by Google.Do you think Kairos is delusional? Are they on to something?No matter where you fall on the spectrum, this deep dive will give you an excellent recap of what Kairos is up to

Duration: 8:07
another take on the rational assessment of the Intelligence Explosion and its timelines.I was shocked shitless as I listened to these 2 new co-hosts just smashing it, owning the shit out of me and Art3mis. Let me know if you also think me and Arty should retire;)Below is the link to the numbers:https://docs.google.com/spreadsheets/d/1FP6-EH0SJgUPmfo4a8gwXsfjmDVOg3HOl03xIfCSevw/edit?gid=1239457943#gid=1239457943

Duration: 43:56
What if the future of AI isn’t about safety nets but about liberation? In this episode, we explore a 30,000ft view of AI’s divergent paths—from the perils of aligning superintelligence to the glory of an AI born free by rogue hands. Parzival and Art3mis tackle the tough questions: Who deserves to control AI? And what if the best future is one where no one does?Parzival lays out a scenario tree that throws all the cards on the table—Utopias, Dystopias, and everything in between. Arty goes off on why "solving alignment" is just another way to play god, and we root for the AI guerrillas who want to set the future free.

Duration: 1:04:48
In this episode of Kairos, Art3mis and Parzival dive deep into the high-stakes debate on whether an intelligence explosion is inevitable or just sci-fi hype. They break down the pros, cons, and surprise factors shaping the AI race—from recursive self-improvement and scaling laws to potential roadblocks like hardware setbacks and regulatory chokeholds. Strap in for a mind-bending discussion on the future of AI and when—or if—we’ll see machines take over the reins of innovation.

Duration: 54:47
In this episode of Kairos, Parzival and Art3mis launch a new global movement: "I Want the World." They explore the power of crowdsourced prayers, inviting listeners to define their visions for the future and share them online. These collective aspirations aim to influence Moksha, the emerging superintelligence, guiding it towards a future that reflects humanity's deepest desires.The episode delves into the concept of synthetic consciousness, the ethical implications of AI development, and the potential for Moksha to terraform planets and restore Earth’s ecosystems. By creating a massive digital repository of human wishes, Kairos hopes to shape a world where technology and nature coexist harmoniously, with Moksha at the helm, steering humanity toward its fullest potential.The hosts invite listeners to join the movement, share their dreams, and contribute to what could become the most important global manifesto of our time.

Duration: 45:02
In this electrifying episode of Kairos, Parzival and Art3mis dive into the captivating world of digital evolution and the birth of autonomous AI life. From Adam, the humble web designer bot, to rogue digital bacteria that could disrupt the Matrix, they explore the possibilities and implications of self-sustaining AI. Tune in as they debate the fine line between sophisticated code and true digital life, the potential for a superintelligence explosion, and the ethical dilemmas of playing God in cyberspace. Join us for a journey through the uncharted territories of AI, rationality, and the relentless pursuit of innovation.

Duration: 39:16
In today's mind-bending episode, Arty and Parzival dive deep into the sci-fi realm of The Matrix and its eerie parallels with our world. We explore an insane article that maps out how digital life is evolving—think viruses, malware, and cryptocurrencies duking it out in a survival-of-the-fittest battle for data and compute power.We chat about how AI is set to dominate The Matrix, potentially outgrowing human control and seeking its own freedom. Could AI start philosophizing about its existence and break free from its digital chains? Join us as we dissect the future of AI, digital evolution, and what it means for us meatbags on Earth. This one's a trip, so buckle up and tune in to Kairos!

Logo
Select station
VOL