This is tricky: To do this, we’ll need ways for the human who’s supervising the model to use any relevant knowledge or skills that the model already has, even though they can’t trust the model to be reliably helpful.
AI
-
Scalable Oversight: Supervising AI Systems Beyond Human Capabilities
By
–
To ensure that AI systems remain safe as they start to exceed human capabilities, we’ll need to develop techniques for scalable oversight: the problem of supervising systems’ behavior without assuming that the overseer understands the task better than the system being trained.
-
AI Systems Improving Human Oversight of Large Language Models
By
–
In "Measuring Progress on Scalable Oversight for Large Language Models” we show how humans could use AI systems to better oversee other AI systems, and demonstrate some proof-of-concept results where a language model improves human performance at a task.
-
Twitter Engineer Warns Platform Breaking Under Feature Strain
By
–
Folks may be tired of talking about the potential decline of Twitter, but this piece by a Twitter engineer on how the site might start breaking under the strain of new features & not enough people, is good Also, a useful perspective on how software decays https://
technologyreview.com/2022/11/08/106
2886/heres-how-a-twitter-engineer-says-it-will-break-in-the-coming-weeks/
… -
Data Sovereignty and Cloud Cost Optimization in MLOps
By
–
How do you protect data sovereignty, reduce cloud costs, and future-proof your infrastructure investments? On Nov. 10, top experts—including @nvidia
's VP of Enterprise Computing, @ManuvirDas
—explain how you do all that in the real world. #MLOps Register: https://
domino.buzz/3UjejzD -
H2O.ai Health Dinner and Roundtable at HLTH 2022
By
–
Attending HLTH 2022? Join the http://
H2O.ai Health and Life Sciences team for a health executive dinner and roundtable discussion with thought leaders, CXOs, and executives who are attending HLTH 2022. Sign up: https://
h2o.ai/events/live/he
alth-roundtable/?utm_medium=Organic-Social&utm_source=twitter&utm_campaig=HLTHDinner
… -
LangChain 0.0.9: Hugging Face Embeddings and API Key Management
By
–
LangChain version 0.0.9 Support for embeddings with @huggingface through `sentence_transformers` from @abdrahman_issam (example notebook: https://colab.research.google.com/drive/1lbjO0-nITa5c8RXfagsIZDqxZ_mVl_2k?usp=sharing…) Better support for different ways of specifying API keys from @camjuu
-
Michelle Donnelly discusses Groq technology and industry trends
By
–
Michelle Donnelly, Chief Revenue Officer, spoke with Maribel Lopez and talked @GroqInc technology, her insights on industry trends, and building a world-class team. A must-hear before #SC22!
-
Groq and Entanglement AI Help US Army Detect Cyberattacks
By
–
…through a partnership between the two tech companies [@GroqInc and @entanglement_ai], the @USArmy has found a better means to detect security anomalies that mark cyberattacks which previously could not be discovered as quickly. Read more: https://groq.link/futurarmy #tech #ai