Community: Eval for LLMs are broken! Academic benchmarks are not representative of real world performance! . We need better evals! Also the same community: Lets make definitive rankings & leaderboards based on just four zero-shot "LM harness" tasks! Not wanting to single
ETHICS
-
AI Technology Recreates Deceased Individuals Through Advanced Algorithms
By
–
AI Brings Deceased Loved Ones Back to Life: A groundbreaking development in artificial intelligence (AI) has stirred conversations around the world as it demonstrates the power of technology in bridging the gap between life and death. A young man in… https://
analyticsvidhya.com/blog/2023/04/a
i-brings-deceased-loved-ones-back-to-life/?utm_source=dlvr.it&utm_medium=twitter
… -
Recent Activity and Friday Musings on AI Governance
By
–
There has been a lot of recently activity on #governing #ai from academics, to industrialists to governments. Some Friday musings on this hot topic. Governance of algorithms is not a new idea, but given the rapid availability…
https://
lnkd.in/g-_XxbRD https://
lnkd.in/g6scxpMX -
AGI Technical Alignment and Societal Integration Increasingly Tractable
By
–
Increasingly optimistic about the technical alignment problem given developments of past few years, and given those of the past six months that coordination & positive societal integration may be tractable too. All critical to getting AGI right:
-
Why AI safety concerns suddenly matter to the public
By
–
2019: OpenAI thought GPT-2 presented serious long term risks. Others in the community didn’t care. Public didn’t care. 2023: OpenAI releases GPT-4 with lots of safety tests. Loud parts of the community goes nuts. Public cares enough that governments rush to legislate. Explain?
-
AI Ethics Debate Will Surpass Content Moderation Challenges
By
–
The #AI #Ethics War Will Make the Content Moderation Debate Look Like a Picnic https://
bit.ly/3DudHkg -
Differential Privacy in Machine Learning: Techniques and Challenges
By
–
Today, we discuss the current state of differentially private ML (DP-ML) research with an overview of common techniques for obtaining DP-ML models, engineering challenges, mitigation techniques and current open questions. Learn more ↓
-
Tech-Wise Over Tech-Savvy: Wisdom in Technology Use
By
–
𝗧𝗵𝗲 𝗶𝗺𝗽𝗼𝗿𝘁𝗮𝗻𝗰𝗲 𝗼𝗳 𝗯𝗲𝗶𝗻𝗴 "𝘁𝗲𝗰𝗵-𝘄𝗶𝘀𝗲" 𝗼𝘃𝗲𝗿 "𝘁𝗲𝗰𝗵-𝘀𝗮𝘃𝘃𝘆" Instead of just knowing how to use technology, understanding when not to use it, and having the wisdom to use it effectively and responsibly, will become essential skills in the future
-
Spanish lawyers challenge police EncroChat hacking as human rights violation
By
–
Spanish lawyers claim police hacking of EncroChat cryptophones breaches human rights law #Technology #Tech #FutureTech https://
computerweekly.com/news/366537937
/Spanish-lawyers-claim-police-hacking-of-EncroChat-cryptophones-breaches-human-rights-law?utm_source=dlvr.it&utm_medium=twitter
… -
Manhattan Project-Scale Investment Needed for AGI Safety Research
By
–
its good to hear so many people starting to get serious about agi safety. we need to be very ambitious. in ww2 the Manhattan Project cost 0.4% of U.S. GDP. Imagine what an equivalent programme for safety could achieve today.