I like these policies as code because they are interpretable and suitable for human feedback. https://
flip.it/WiuOD3
AI
-
Policies as Code for AI Interpretability and Human Feedback
By
–
-
NeurIPS 2022 Research Asset on Blockchain
By
–
NeurIPS.Crypto Asset: https://
opensea.io/assets/matic/0
xa9a6a3626993d487d2dbda3173cf58ca1a9d9e9f/85236298060416470676056917210060248218208711908924985904219618869194675364809
… #NeurIPS #NeurIPS2022 -
Robotics Research Conference in New Zealand December 2022
By
–
Join us in sunny New Zealand 14-18 December for 200 papers, 15 workshops, and 10 Keynote talks on the cutting edge of robotics research. @ToyotaResearch @GoogleAI @DeepMind @amazon @BoschGlobal @IntuitiveSurg
-
Unifying Data Analytics and AI with Lakehouse Architecture
By
–
People who don’t unify their #data, analytics and #AI with #Lakehouse are a bit sus
-
Building AI Projects with Hugging Face Spaces
By
–
Would be so cool to build a http://
hf.co/spaces with it! -
Prompt Engineering for AI Agents: Step-by-Step Operation Sequencing
By
–
The game is “least to most” — have it write out the most mechanical, reliably performed operations first. Then gradually move up to more interpretive and analytical steps, so it’s never asked to do anything multi-step in its head. The prompt should feel like an income tax form.
-
Global Interconnection Index Report on Digital Transformation Trends
By
–
Download! #news The Global Interconnection Index #GXIreport @Equinix gives valuable view on macro #trends & benchmarks shaping #DigitalTransformation from #Sustainability to #Skills https://
eqix.it/3ToVTMW #ESG #tech #innovation #SDGs @Shi4Tech @Chels_LA @Analytics_699 -
Twitter’s Risk Calculus Shifts After Going Private
By
–
I would also add that the risk calculus for Twitter has changed. At a large public company, risk is avoided (which can be a good & safer thing, but slows change). Private companies, and especially entrepreneurial ones, make bigger bets with higher risk of failure or big returns.
-
GPT-3 Tokenization Challenges and Model Limitations
By
–
+1 for this approach. GPT-3’s tokenization is a big gotcha. Model doesn’t see text as sequences of characters unless you add spacing/hyphens. It can’t even reliably tell you the last letter of a word without first rewriting it that way. Similar for counting length:
-
Mastodon Privacy Settings Superior to Twitter Scraping Capabilities
By
–
Masrodon has more fine-grained privacy settings than Twitter. Can‘t scrape all.