TLDR AI 2025-11-14

AI STARTUP CURSOR RAISES $2.3 BILLION FUNDING ROUND AT $29.3 BILLION VALUATION

Cursor has closed a $2.3 billion funding round at a $29.3 billion post-money valuation. The tool has crossed $1 billion in annualized revenue and swelled to more than 300 employees since it first launched in 2023. Coding tools have proved to be a lucrative AI use case. OpenAI approached Anysphere to potentially purchase Cursor earlier this year, but the deal failed to gain traction.

MURATI'S THINKING MACHINES IN FUNDING TALKS AT $50 BILLION VALUE

Thinking Machines Lab is in early talks to raise new funding at a roughly $50 billion valuation. The new financing would more than quadruple the startup's valuation from July. The deal's terms have yet to be finalized and could change. Thinking Machines aims to build AI models and products to help businesses by supporting more 'human-AI collaboration' across every field of work. The startup has dozens of employees who previously worked at OpenAI.

GOOGLE ADDS DEEP RESEARCH IN NOTEBOOKLM

Google's latest update to NotebookLM introduces Deep Research, a feature designed to support high-quality source integration, and adds support for analyzing new file types, including spreadsheets and Word documents.

NANO BANANA CAN BE PROMPT ENGINEERED FOR EXTREMELY NUANCED AI IMAGE GENERATION

Nano Banana allows for very strong generation control, but it has several issues. It is terrible at style transfer, even though it has essentially no restrictions on intellectual property. The model can be used to create highly specific, high-quality images. This article looks at some of the tricks to using the model.

AI-ORCHESTRATED CYBERATTACK DETAILED BY ANTHROPIC

Anthropic disclosed the first documented AI-driven cyber espionage campaign, carried out largely autonomously using its Claude Code tool. The attackers, linked to a Chinese state-sponsored group, targeted 30+ entities worldwide, demonstrating the growing threat of agentic AI systems in cybersecurity.

GenAI and agentic AI. Read the report

INTRODUCING GPT-5.1 FOR DEVELOPERS

GPT-5.1 uses fewer tokens on simple tasks and shows modest benchmark gains on SWE-bench (+3.5 points to 76.3%) and GPQA Diamond (+2.4 points to 88.1%). On other capabilities, its benchmark scores are flat or slightly lower.

INTRODUCING CODE WIKI: ACCELERATING YOUR CODE UNDERSTANDING

Code Wiki is a platform that maintains a continuously updated, structured wiki for code repositories. It scans the full code base and regenerates documentation after each change, so docs evolve with the code. Every wiki section and chat answer is hyperlinked directly to the relevant code files and definitions. The platform will allow new contributors to start immediately, while senior developers can now understand new libraries in minutes.

WEIGHT-SPARSE TRANSFORMERS HAVE INTERPRETABLE CIRCUITS

OpenAI trained language models with most weights forced to zero, making it easier to identify which neurons link together into a “circuit” representing a certain concept. The approach successfully isolated circuits for tasks like Python quote-matching that are both necessary (removing them breaks the behavior) and sufficient (they work in isolation).

GOOGLE IS ROLLING OUT CONVERSATIONAL SHOPPING—AND ADS—IN AI MODE SEARCH

Google users in the US will soon be able to ask AI Mode complex questions about what to buy. The model will respond with suggestions, guides, tables, and other generated content to help them decide. Some of the content that appears in AI Mode will be ads. There will be an agentic checkout feature that lets AI automatically make purchases with Google Pay at specific retailers. Shopping features are also coming to the Gemini app, but there won't be sponsored content in the results for the time being.

JEVONS OR BUST

Demand for AI is growing, and it seems that it is paying demand. AI is driving incremental revenue to big cloud providers. However, this doesn't mean that AI is definitely a Jevonic-non-bubble instead of a capex binge that ends badly. We have yet to see whether the demand side will materialize.

and efficient LLM running the cloud. See how it works

MEASURING POLITICAL BIAS IN CLAUDE

Anthropic released an open-source "Paired Prompts" evaluation that measures political even-handedness across AI models.

DEEPMIND RELEASED SIMA 2

SIMA 2 is a generalist AI agent that uses Gemini LLM capabilities to reason, adapt, and interact with virtual environments.

AMAZON AND MICROSOFT BACK EFFORT THAT WOULD RESTRICT NVIDIA'S EXPORTS TO CHINA

The Gain AI Act would give tech firms, including Microsoft and Amazon, preferential access to chips at their data centers around the world.

APPLE AND OPENAI MUST FACE XAI LAWSUIT OVER SIRI INTEGRATION

xAI is suing Apple in OpenAI for allegedly conspiring to dominate the AI market.