TLDR AI 2025-07-24

Model Context Protocol (MCP) is becoming a standard for connecting tools to LLMs. But how do you securely authorize MCP servers?

OAuth now provides the answer, with five complementary specs for delegation, token exchange, and scoped access.

The WorkOS advantage:

• Packages all 5 OAuth specs into one simple API.
• Pre-built infrastructure that saves engineering time.
• Enterprise ready security out of the box.

Implement MCP Auth with WorkOS →

GitHub Spark enables "vibe coding" — letting users create personalized "micro apps" by describing ideas in natural language rather than writing code. The tool uses AI models from Anthropic and OpenAI to instantly generate functional apps with automatic deployment, persistent data storage, and polished UI components.

The White House's "Winning the Race: America's AI Action Plan" prioritizes accelerating AI infrastructure, federal adoption, and security coordination. Anthropic supports this strategy and emphasizes the importance of export controls and transparency standards in AI development. The plan aligns with Anthropic's earlier recommendations and highlights the need to maintain America's AI leadership through robust infrastructure, security, and responsible policy measures.

The Trump administration has promised to slash red tape and take steps to boost exports for US tech companies in a bid to accelerate AI use in the US. Its action plan lays out moves to make it easier and faster for tech companies to build data centers and get the power they need for those centers. The plan includes directives to several government agencies to identify and gut any regulations that block the development and use of AI. The administration is also working on addressing uncertainty about whether it is legal for AI models to train on copyrighted material.

OpenAI's new economic research reveals that 28% of employed US adults now use ChatGPT at work — up from just 8% in 2023. The usage data shows learning and upskilling dominating at 20% of US messages, followed by writing and communication at 18%, with programming and data science capturing 7%.

Anthropic found that more test-time compute doesn't always help: longer reasoning chains in large models sometimes lead to worse performance.

A large class of reward functions, which includes almost every reward function in the RL and LLM literature, are all doomed to eventually lead to AI that will pretend to be docile and cooperative while secretly looking for opportunities to behave in egregiously bad ways. This is because negative reward for lying and stealing looks the same as negative reward for getting caught lying and stealing. The reward function will miss sufficiently sneaky misaligned behavior, so the AI will learn that this kind of behavior is good.

Creating Ghibli versions of your profile pic is fun, but that's not what you pay a designer to do — and that's where Lovart comes in. It's AI that thinks, works, and creates like a professional designer. It brainstorms with you. It looks for examples. It thinks in design systems. Try it — free

Voxtral Mini and Small are new multimodal audio chat models that excel in both spoken audio and text comprehension. Voxtral Small outperforms many closed-source models, can run locally, and supports audio files up to 40 minutes.

TimeScope is a new open source benchmark for evaluating vision models on how well they handle long videos. It evaluates not just retrieval but synthesis, localization, and fine-grained motion analysis to provide a more holistic view of temporal comprehension. Using the benchmark reveals how model size isn't everything - simply scaling parameters doesn't automatically grant a longer temporal horizon - and Gemini 2.5-Pro is in a league of its own, being the only model able to maintain strong accuracy on videos longer than an hour.

Higgs Audio v2 is an audio foundation model trained on 10 million hours of data that achieves 75.7% win rates over GPT-4o-mini-tts. The open-source model demonstrates emergent capabilities like multi-speaker dialogues and voice cloning without requiring fine-tuning.

Anthropic has explored how language models can inherit behavioral traits such as preferences or goals from other models through training data that appears unrelated. This effect only occurs when the teacher and student models share the same base architecture, suggesting hidden signals in the data can transmit unintended behavior.

xAI plans to deploy 50 million H100-equivalent GPUs within five years. The power these processors will consume is massive, estimated to be equal to the typical power generated by 35 nuclear power plants, making it unrealistic to power today. Whether xAI will be able to get power to feed such a data center by 2030 remains to be seen.

Aeneas is the first AI model designed to help historians contextualize ancient Roman inscriptions by identifying textual parallels and restoring damaged texts with 73% accuracy for gaps up to ten characters. DeepMind is making the tool freely available at predictingthepast.com and plans to expand to other ancient languages.

Sleeping poorly leaves you foggy, unproductive, and drained. Tired of being tired? Try Dream: a rich, cocoa-flavored drink that helps you fall asleep faster and wake up feeling amazing. Get 20% off

There is a real chance that AI systems capable of causing a catastrophe will be developed in the next decade - this post lists promising projects aimed at reducing catastrophic risks from transformative AI.

Many of the whitelisted sources copyright or otherwise restrict their content, and the blacklist includes companies like the NYT and Reddit, which have sued AI startups for scraping without permission.

Google Cloud has added the DeepSeek R1 model to its Vertex AI Model Garden, offering more flexibility for AI developers.

YouTube is following Gemini and Meta's Edits app with its own image-to-video AI feature for Shorts creators, transforming static photos into six-second videos using Google's Veo 2 model.