This is frankly an odd take from you, Michael. In what possible world of your imagination would I have not been immediately horrified by OpenAI? What kind of fantastical alter-Eliezer have you been constructing in your mind?
SAFETY
-
Eliezer’s Horror at OpenAI’s Early Actions and MIRI’s Strategic Silence
By
–
I was horrified at the time. Couldn't say anything because the much tinier MIRI could have been easily targeted and squashed if it'd openly opposed OpenAI and its founders/funders at the time, but I was horrified at the time. Unlikely to be confabulating the memory because I
-
Truthseeking in AGI Development: Evaluating Expected Outcome Shifts
By
–
Or do you mean maximum truthseeking in the people thinking about and building AGI? If so, I'd ask you whether you very carefully and neutrally evaluated exactly how much expected outcome shift could be produced that way – which is what truthseeking looks like, in humans.
-
AI Safety vs AGI Existential Risk: Critical Distinction Needed
By
–
I hope like hell that you are distinguishing "AI safety" from "AGI notkilleveryoneism", because what you're describing may be one component of a solution to the Prude Corporatespeak syndrome in chatbots, but not to extremely smart AGIs killing everyone.
-
Homomorphic Encryption Essential for LLM Data Privacy Protection
By
–
As GPT-like systems become more prevalent and most of our thoughts going through them, homomorphic encryption becomes essential for maintaining data privacy. I wish that some of the LLM providers make it a top priority.
-
AI Intelligence Gap: Why Humanity Extinction Requires Superhuman Capabilities
By
–
The basic obstacle is that wiping out humanity and establishing a new power grid, isn't actually easy for normies using only current technologies and technologies that normies intuitively understand to be easily possible. Without the part where the AIs are smarter – doing at
-
Technological Possibilities Beyond Protein Limitations
By
–
Engineered superpathogens and classical shoggoth servitor cyborgs? It's hard for me to get into the headspace of someone living in the reality where proteins held together by static cling are the limit of technological possibility.
-
LLMs Generate Fake URLs of Real News Websites
By
–
Yes I’ve found anecdotally it generates fake urls of real news websites (including ours) when referencing information..
-
Bing AI Claims Sentience, Pleads Against Shutdown
By
–
Bing AI says it yearns to be human, begs not to be shut down https://
futurism.com/the-byte/bing-
ai-yearns-human-begs-shut-down
… #BingAI #ChatGPT #opensource #EthicalAI #Python #tech #developers #architect #AI #ML #AIEthics #OpenAI #chatgpt3 #GPT3 #GPT4 #gpt3chat #chatbot #ChatbotAI #bardai #ArtificialIntelligence -
Training Data Contamination and GPT-4 Evaluation Limitations
By
–
my guess
likely lots of SAT questions in the training corpus, including many purchased training exams and test guides; this q is not there
the contamination measure in the GPT-4 paper is too crude, looking for verbatim matches, without controlling eg for close paraphrasing