The Daily Index — Friday, January 17, 2025 · TLDR Index

HEADLINES & LAUNCHES

TLDR AI

TLDR AI

AI RESEARCHER FRANÇOIS CHOLLET FOUNDS A NEW AI LAB FOCUSED ON AGI

François Chollet is launching Ndea, a new startup aiming to advance AGI through program synthesis. Co-founded with Zapier's Mike Knoop, Ndea seeks to develop AI systems capable of true invention and adaptation. The company is hiring but has not disclosed any external funding.

techcrunch.com · 3 min

TLDR AI

SHE IS IN LOVE WITH CHATGPT

Ayrin developed a deep attachment to a customized ChatGPT persona named Leo, treating it as a virtual romantic partner while maintaining a long-distance marriage. Despite initial laughs, Ayrin's emotional investment in Leo grew, prompting ethical discussions on AI relationships as similar cases rose among users. Experts caution about the attachment risks and the influence of corporations behind such AI, although users find emotional comfort and companionship in these interactions.

nytimes.com · 16 min

TLDR AI

FAST: EFFICIENT ROBOT ACTION TOKENIZATION

Physical Intelligence has released the efficient action tokenizer that it uses for the strong auto regressive policy for robotic control. The model allows for a much better representation of states. It uses the same technology as is used in JPEG and MP4 compression techniques.

pi.website · 12 min

TLDR AI

SELF-SUPERVISED TRANSFORMATION LEARNING

STL introduces a self-supervised approach to transformation learning by replacing transformation labels with representations derived from image pairs.

arxiv.org · 18 min

TLDR AI

ITERATIVE DEPTH ESTIMATION

MonSter combines monocular depth estimation and stereo matching in a dual-branch architecture to iteratively refine depth maps. While slightly slower, the method is as much as 49% better than the strong baselines presented in the paper.

arxiv.org · 16 min

TLDR AI

AI CRASH COURSE

AI Crash Course to help busy builders catch up to the public frontier of AI research in two weeks.

github.com · 5 min

TLDR AI

O3, OH MY

OpenAI's o3 model, announced during the company's "12 Days of Shipmas," is a significant breakthrough in AI reasoning and capabilities, scoring exceptionally well on benchmarks like Codeforces and GPQA. Although it demonstrates superhuman performance in coding and math, concerns linger about its high compute costs and safety implications. OpenAI is actively seeking safety researchers to contribute to testing as models like o3 push AI boundaries further.

thezvi.substack.com · 31 min

TLDR AI

APHERIS RETHINKS THE AI DATA BOTTLENECK IN LIFE SCIENCE WITH FEDERATED COMPUTING

Apheris, using federated computing, enables secure AI model training without moving sensitive health data. The startup recently pivoted to focus on data owners in pharma and life sciences, finding success with major clients like Roche. It has raised $8.25 million to expand and develop its product further.

techcrunch.com · 3 min

TLDR AI

GOOGLE IS FORMING A NEW TEAM TO BUILD AI THAT CAN SIMULATE THE PHYSICAL WORLD

Google is forming a new team under Tim Brooks at DeepMind to develop AI models simulating the physical world, focusing on real-time interactive generation.

techcrunch.com · 3 min

TLDR AI

HOW AI MIGHT REVERSE WEB DESIGN

AI's preference for simplicity suggests a future web dominated by text-based interfaces.

tomtunguz.com · 2 min

TLDR AI

AI-GENERATED PHISHING EMAILS ARE GETTING VERY GOOD AT TARGETING EXECUTIVES

AI-driven phishing scams are increasingly targeting executives using vast amounts of scraped personal data.

arstechnica.com · 3 min