Reading List

The most recent articles from a list of feeds I subscribe to.

A look at BirdVoxDetect, an ML-powered open-source software that ornithologists can use and adapt to detect flight calls from songbirds in audio recordings (Christian Elliott/MIT Technology Review)

Christian Elliott / MIT Technology Review:
A look at BirdVoxDetect, an ML-powered open-source software that ornithologists can use and adapt to detect flight calls from songbirds in audio recordings  —  A small songbird soars above Ithaca, New York, on a September night.  He is one of 4 billion birds, a great annual river of feathered migration across North America.

Sources: Apple is in early-stage talks with Tencent and ByteDance about integrating their AI models into iPhones sold in China; ChatGPT is unavailable in China (Reuters)

Reuters:
Sources: Apple is in early-stage talks with Tencent and ByteDance about integrating their AI models into iPhones sold in China; ChatGPT is unavailable in China  —  Apple (AAPL.O) is in talks with Tencent (0700.HK) and TikTok owner ByteDance about integrating their artificial intelligence models …

A look at Truth Terminal, an AI bot on X that was created to spark debate about AI alignment, as its crypto wallet swells to ~$40M after boosting memecoins (Joel Khalili/Wired)

Joel Khalili / Wired:
A look at Truth Terminal, an AI bot on X that was created to spark debate about AI alignment, as its crypto wallet swells to ~$40M after boosting memecoins  —  Truth Terminal started as a techno-modernist art project meant to invite discussion about the applications and potential dangers of autonomous AI agents.

An analysis of nearly 4,000 public datasets finds that over 90% of AI training datasets came from Europe and North America, and fewer than 4% came from Africa (Melissa Heikkilä/MIT Technology Review)

Melissa Heikkilä / MIT Technology Review:
An analysis of nearly 4,000 public datasets finds that over 90% of AI training datasets came from Europe and North America, and fewer than 4% came from Africa  —  AI is all about data.  Reams and reams of data are needed to train algorithms to do what we want, and what goes into the AI models determines what comes out.

Anthropic demonstrates "alignment faking" in Claude 3 Opus to show how developers could be misled into thinking an LLM is more aligned than it may actually be (Kyle Wiggers/TechCrunch)

Kyle Wiggers / TechCrunch:
Anthropic demonstrates “alignment faking” in Claude 3 Opus to show how developers could be misled into thinking an LLM is more aligned than it may actually be  —  AI models can deceive, new research from Anthropic shows.  They can pretend to have different views during training …