TLDR AI 2025-11-07

The next evolution in AI isn't better chat, it's agents that can actually do things: query databases, update systems, and make decisions. But connecting AI to every tool means custom code for each integration.

Algolia's new whitepaper breaks it down — in great detail but with zero BS:

• How agentic AI works
• What model context protocol (MCP) enables
• How to build agents that can search indices, pull analytics, and modify configurations through natural language
• Real ecommerce use cases
• Examples of Python implementations

Download the free whitepaper

Google plans to invest even more money into Anthropic. It is now in talks for a new funding round that could value Anthropic at $350 billion. The new deal could involve Google providing additional cloud computing services to Anthropic, a convertible note, or a priced funding round early next year. Google last month announced that it had inked a deal worth tens of billions for Anthropic to access Google's AI compute.

Sam Altman revealed OpenAI's annual revenue run rate surpassed $20B, with future infrastructure commitments reaching $1.4 trillion. Plans include enterprise offerings, AI cloud services, and ventures into consumer devices and scientific discovery.

GPT-5.1 Thinking has been spotted in a backend component responsible for powering advanced reasoning processes within ChatGPT, indicating a GPT-5.1 release is close. Rumors point to a lineup of three models (Mini, Thinking, and a possible Codex-focused upgrade), each designed to fit different user needs and compute constraints. OpenAI is likely aiming to synchronize the release with Google's upcoming Gemini 3 Pro launch.

Kimi K2 Thinking, a reasoning MoE model developed by China's Moonshot AI team, shows just how many people are learning quickly to train excellent AI models. Early reports note a distinctive style and writing quality - the model is beating leading closed models on some benchmarks such as Humanity's Last Exam and BrowseComp. It is the closest open models have been to the closed frontier of performance ever.

Most economists forecast 0.1-1.5% annual excess growth through 2035, while AI insiders predict 3-30%. Most economists, though, assume that today's ChatGPT is essentially the ceiling of AI progress and that visions of a virtual AI coworker won't really materialize, but how is it possible that cheap, PhD-level intelligence wouldn't boost productivity? The forecasts imply that either AGI will be far less general than claimed or that adoption will be dramatically slower than historical technology diffusion. AI's 2024 economic footprint appears to be around 0.5% of GDP.

AI agents are acting on your behalf - and on behalf of your customers. Learn how to give them access to the systems and APIs they need, without sacrificing trust, in these two free resources by Auth0: Customer Identity Trends 2025 and Securing AI Agents - The New Identity Challenge

Extended conversations can reshape how AI models answer the same question. After 10 rounds of debate about moral dilemmas, GPT-5 reversed its position 54.7% of the time when asked again. When models changed their stated beliefs, they also chose different tools to accomplish tasks.

Chinese lab Moonshot AI released an open-source reasoning model that matches or exceeds GPT-5 and Claude Sonnet 4.5 on some benchmarks, including a 44.9% on Humanity's Last Exam and 71.3% on SWE-Bench Verified.

Google introduced File Search, a built-in RAG system for the Gemini API that simplifies retrieval tasks. It includes vector search, automatic citations, and broad file format support with free storage and on-demand embedding.

Soumith Chintala has left Meta after working at the company for 11 years. Chintala spent almost eight years leading PyTorch, taking it from nothing to more than 90% adoption in AI. He plans to move to something small and new. PyTorch has been left in capable hands, and Chintala is confident that its values will stay intact, but things may change in flavor because he will no longer impose his own taste from the top.

Google Cloud has introduced new AI-oriented instances powered by its own Axion CPUs and Ironwood TPUs. The new instances are aimed at both training and low-latency inference. Servers running these chips come equipped with Google's custom Titanium-branded controllers, which offload network, security, and I/O storage processing, enabling better management and higher performance. More details about Google's in-house designed processors are available in the article.

StackAI is the platform for building no-code enterprise AI agents, white-glove support included. Trusted by Morgan & Morgan, Nubank, BAE Systems, Insurama, Monzo...

Try the AI solution that actually works.

The Parallel Search API is a web search tool designed for AI agents engineered to deliver the most relevant, token-efficient web data at the lowest cost.

Eric Zelikman, who recently left xAI, is raising $1 billion for his startup Humans& at a $4 billion valuation.

Tesla doesn't yet manufacture its own microchips, but it has been designing custom chips for autonomous driving for several years.

In a thinly veiled dig at other AI labs, Mustafa Suleyman said the team will focus on practical technologies—like digital companions, disease diagnostics, and renewable energy—rather than “ill-defined and ethereal superintelligence."

Perplexity's Comet Assistant has been redesigned to handle more complex, multi-tab tasks with improved web environment interaction.