it's a dangerous myth that AI safety and AI capability are orthogonal vectors. also a myth that we can "avoid ruin" without careful iteration and contact with reality. things have not gone as our best experts have predicted, and that will continue to be the case.
SAFETY
-
Tabula-Rasa Model Oxymoron: NFL Theorem and Fundamental Assumptions
By
–
"Tabula-rasa model" is an oxymoron, by the NFL theorem. The question is what are the fundamental assumptions that need to be built in. Causality is a candidate. The laws of physics are another.
-
Testing Stock Prediction Models for Security Flaws
By
–
Testing stock prediction models for security flaws 📈 pic.twitter.com/XO9AQCglr1
— IBM Data, AI & Automation (@IBMData) 16 novembre 2022Testing stock prediction models for security flaws
-
Self-examination in AI science models for error detection
By
–
what would be cool with science models like Galactica would be to add self examination of the inputs like the coming Codex model, where the model can inspect it’s theorem proofs or essays and spot errors
-
Concerns about GPT alignment with human values and safety
By
–
"Finally, we are very concerned that this GPT could be unaligned with humans. This would be bad. We want this to be a nice GPT that deeply loves all humans and is always considerate and helpful. Thanks"
-
Mastodon’s Safety-First Design Built by Underrepresented Communities
By
–
Mastodon was built by sensible, thoughtful people, many from underrepresented communities who had very clear needs around safety and moderation – they've put a lot of careful work into that, and the community there continues to invest in it I'm optimistic!
-
OpenAI Emphasizes Safety as Core Pride and Track Record
By
–
we do take safety extremely seriously, and it's what i'm most proud of openai for. i am sure we will make some very wrong decisions along the way, but i think our track record as a whole will really hold up.
-
Tesla Autopilot Safety Data Transparency Concerns
By
–
https://
nytimes.com/2022/06/08/tec
hnology/tesla-autopilot-safety-data.html
… -
Tesla FSD Still Falls Short of Required Reliability Standards
By
–
Thorough story on FSD in a neighborhood where Tesla's CEO has sent engineers to solve previous failures noted by the car owner. Appears to be at about one 9 to my eye. Needs to be at about six 9s to be trusted. Each 9 is ten times harder than the previous. https://
nytimes.com/interactive/20
22/11/14/technology/tesla-self-driving-flaws.html
…