Analytics DeepLearning.AI - The Batch | DeepLearning.AI (Page 36)

Music producer uses AI assistant on screen to mix a song while singer records vocals in a professional studio.

Data Points

Building multi-agent systems in Rowboat’s IDE: Top computer use agent UI-TARS gets an update

GPT-4o’s image generator now available via API. Google updates its Lyria model and music editing tools. Grok 3 models now available for API developers. Executive order would overhaul K-12 AI education in U.S. schools.

Code snippet showing ‘Keep Building!’ printed in multiple programming languages including Python, Java, JavaScript, and C++.

Letters

How to Become a Multilingual Coder: AI makes it easy to code in any programming language — especially if you know just one.

Even though I’m a much better Python than JavaScript developer, with AI assistance, I’ve been writing a lot of JavaScript code recently.

The Batch Newsletter

OpenAI’s Five New Models, Hugging Face’s Open Robot, U.S. Tightens Grip on AI Chips, Text-Only LLMs Go Multimodal

The Batch AI News and Insights: Even though I’m a much better Python than JavaScript developer, with AI assistance, I’ve been writing a lot of JavaScript code recently.

AI diagram showing generator and scorer loop to produce final output based on test image of a cat.

Machine Learning Research

Text-Only LLM Goes Multimodal: LLMs learn to caption images, video, and audio without further training

Large language models excel at processing text but can’t interpret images, video, or audio directly without further training on those media types. Researchers devised a way to overcome this limitation.

NVIDIA AI computing hardware with multiple GPUs on a black reflective surface.

Business

U.S. Tightens Grip on AI Chips: U.S. makes new rules for AI chip export rules to China, launches Nvidia investigation

The U.S. government escalated its long-running effort to block China’s access to cutting-edge AI hardware.

Person interacting with a humanoid robot using virtual reality headset and controllers.

Hardware

Hugging Face Rolls Out Open Robot: Hugging Face acquires Pollen Robotics, launches Reachy 2 robot for open-source research

Hugging Face has made a name by providing open AI models. Now it’s providing an open robot.

Comparison chart of GPT-4.1, o3, and o4-mini with other models on coding, math, tool use, and multimodal reasoning benchmarks.

Machine Learning Research

OpenAI Launches Cost-Effective Alternatives: OpenAI replaces GPT-4.5 with GPT-4.1 Family, plus o3 and o4-mini, new models focused on reasoning and coding

OpenAI refreshed its roster of models and scheduled the largest, most costly one for removal.

Crowd at sunset concert cheering for a giant server on stage, symbolizing AI or tech as a modern rockstar.

Data Points

Evaluating the best AI search engines: Claude can read your Gmail and the web

New BitNet model shows that 1,0,-1 might be enough. Kling gets a model update with image and video inputs. Google rolls out video generation and animation model to subscribers. Music streamer Deezer notes sharp uptick in AI-generated music.

People using laptops to access Wikipedia in a crowded library setting.

Data Points

Wikimedia wants to help build AI for the commons: Pricing and availability for o3 and o4-mini

Gemini 2.5 Flash blends speed with budgeted reasoning. IBM’s Granite Speech sets SOTA in transcription accuracy. Recall will soon return to Copilot Plus PCs. OpenAI will shelve its biggest, costliest model.

Cartoon of two coworkers coding; one struggles with evaluations, the other iterates quickly through model updates and test cases.

Letters

We Iterate on Models. We Can Iterate on Evals, Too: Building automated evals doesn’t need to be a huge investment. Start with a few quick-and-dirty examples and iterate!

I’ve noticed that many GenAI application projects put in automated evaluations (evals) of the system’s output probably later — and rely on humans to manually examine and judge outputs longer — than they should.

The Batch Newsletter

Google Unveils Gemini 2.5, MCP Gains Momentum, Behind Sam Altman’s Fall and Rise, LLMs That Understand Misspellings

The Batch AI News and Insights: I’ve noticed that many GenAI application projects put in automated evaluations (evals) of the system’s output probably later — and rely on humans to manually examine and judge outputs longer — than they should.

Diagram of latent transformer model using byte-level encoding, patching, and cross-attention for next-byte prediction.

Machine Learning Research

Toward LLMs That Understand Misspellings: New byte-based model beats Llama 3 on spelling, noise, and translation

Researchers built a model that’s more robust to noisy inputs like misspellings, smarter about character-level information like the number of R's in strawberry, and potentially better able to understand unfamiliar languages that might share groups of letters with familiar languages.