TLDRocket
Sign in
Latest OpenAI says Apple’s own security practices undermine its trade secrets... — TechCrunch Anthropic recommends a git worktree per agent. Your runtime infra make... — The New Stack Amid legal battles, Suno says it will start watermarking songs — TechCrunch Exclusive: Mirendil inks $100M+ Google Cloud deal to scale self-improv... — TechCrunch Ex-Spotify employees raise $10M to bring the AI behind its recommendat... — TechCrunch Google Maps adds agentic features, including food ordering and hotel b... — TechCrunch Meta Launches Muse Code to Rival Claude Code and OpenAI’s Codex — Trending Topics Meta greift mit Muse Code Anthropic und Codex von OpenAI an — Trending Topics

The AI intelligence platform

Every AI story that matters and the intelligence behind it.

TLDRocket reads all relevant sources, removes duplicate coverage, and publishes a short neutral summary of every story, linking back to the original. Free, no spam, unsubscribe anytime.

Add to Slack

Every story also updates live profiles event timelines weekly rankings the AI Market Index

Tuesday, 27 January 2026

ATLAS: Practical scaling laws for multilingual models

Google Research 6 months ago 11

Researchers published ATLAS, a set of scaling laws for training multilingual AI models across 400+ languages, based on 774 training runs spanning 10M–8B parameters. The study quantifies cross-lingual transfer effects between 1,400 language pairs and shows that supporting twice as many languages requires increasing model size by 1.18x and training data by 1.66x. The work provides practical guidance for developers building models for non-English speakers on optimal language mixtures, model sizes, and whether to pre-train from scratch or fine-tune from multilingual checkpoints.

Management as AI superpower

One Useful Thing 6 months ago 18

A University of Pennsylvania professor ran a four-day startup creation competition where MBA students used AI tools like Claude and ChatGPT to build prototypes, achieving roughly an order of magnitude more progress than students typically accomplish over a full semester. The key finding was that success depended on the students' ability to effectively specify requirements and evaluate outputs—a management skill—rather than AI expertise, with GPT-5.2 achieving 72% parity or superiority to human experts on complex tasks. Effective delegation to AI agents follows an equation balancing human baseline time, probability of AI success, and evaluation overhead, meaning management skills and domain expertise become the bottleneck rather than AI capability.

Terminally online Mistral Vibe.

Mistral AI 6 months ago 5

Mistral released Vibe 2.0, a terminal-based coding agent powered by its Devstral 2 model, adding custom subagents, multi-choice clarifications, and slash-command skills for automated code tasks. Devstral 2 API pricing starts at $0.40 per million input tokens and $2.00 per million output tokens, with Vibe available on Le Chat Pro and Team plans plus pay-as-you-go options. Developers can now build specialized agents for targeted tasks like deployment, testing, and code review without manually switching between tools.

PVH reimagines the future of fashion with OpenAI

OpenAI 6 months ago 34

PVH Corp., the parent company of Calvin Klein and Tommy Hilfiger, is adopting ChatGPT Enterprise across its fashion design, supply chain, and consumer engagement operations. The company will deploy the tool across multiple business functions including product development, inventory management, and customer interactions, though no specific timeline or investment amount was disclosed. This integration aims to streamline creative processes and operational efficiency, though the concrete impact on design output or business metrics remains unspecified.

Introducing Prism

OpenAI 6 months ago 24

Prism is a free LaTeX workspace that integrates GPT-5.2 to help researchers write, collaborate, and organize their work in a single environment. The tool runs natively within LaTeX, eliminating the need for separate applications. This consolidation allows researchers to draft documents, share them with collaborators, and use AI assistance without switching between platforms.

TRUSTBANK uses AI agents to personalize Furusato Nozei gifts

OpenAI 6 months ago 38

TRUSTBANK deployed AI agents built with Recursive and OpenAI models to provide personalized gift recommendations for Furusato Nozei, a Japanese tax incentive program for regional donations. The system uses conversational AI to guide users through gift selection based on their preferences rather than requiring manual browsing. Users can now discover regional gifts more efficiently through natural language interaction instead of traditional category-based searches.

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.