Daily AI briefing

AI Briefing: OpenAI Shakes Up the Agent Race

OpenAI's GPT-5.6 Sol Ultra solves a 50-year-old math problem, while developers shift toward stateful agents and open-source embedding models.

  • Briefing date: 2026-07-12
  • Editorial AI analysis
  • The AI Wrap

Stories covered on 2026-07-12

  1. AI’s Biggest Bottleneck Isn’t GPUs Anymore

    Towards AI · 2026-07-12

    The recent analysis from Towards AI suggests that the primary constraint on AI development has shifted from hardware availability, specifically NVIDIA's H100 GPUs

  2. How to Make Your Website Usable by AI Agents with WebMCP

    Towards AI · 2026-07-12

    I built a hotel booking demo to show how a LLM can discover website tools, call application actions, and update the interface through…

  3. DPO Fine-Tuning from First Principles in Python

    Towards AI · 2026-07-12

    Fine-tuning a Large Language Model (LLM) with human preferences used to require Reinforcement Learning from Human Feedback (RLHF): collect…

  4. Unisound U2: The 266B Model That Scored 87.9% on PhD-Level Science for $0.15/M

    Towards AI · 2026-07-12

    On 7 , a Chinese speech-AI company that most people outside China have never heard of released a large language model that scored…

  5. One Line, Any Model: Multi-Provider Agents in Google ADK via LiteLLM

    Towards AI · 2026-07-12

    All Code Examples:: GitHub RepositoryContinue reading on Towards AI »

  6. Apple Is Suing OpenAI. The Real Story Is What OpenAI Can’t Build.

    Towards AI · 2026-07-12

    Apple's 2026 lawsuit against OpenAI reveals the real AI moat: not the model but the decades-deep hardware supply chain you can't buy or…

  7. What are your plans for AI Appreciation Day?

    Engadget · 2026-07-12

    The best way to celebrate AI Appreciation Day is to not.

  8. Lorde says Ray-Ban Meta AI glasses are ‘not sexy’

    The Verge · 2026-07-12

    Lorde was performing at the Real Cool Festival in Madrid on Thursday and took some time during her set to speak out against AI glasses.

  9. Guide to Loop Engineering: How ‘autoresearch’ and ‘Bilevel Autoresearch’ Turn AI Agents Into Autonomous Machine Learning ML Research Loops

    MarkTechPost · 2026-07-12

    Most people still use AI like a 2015 search box. You type, you read, you type again. A newer pattern replaces that manual back-and-forth with a loop.

  10. AI Agent Production Deployment Best Practices

    Towards AI · 2026-07-12

    A new guide outlines practical strategies for deploying AI agents into production environments, focusing on reliability, scalability, and security.

  11. GitHub Spec Kit vs Kiro vs Claude Code — SDD Workflows

    Towards AI · 2026-07-12

    Developers comparing Spec-Driven Development setups in 2026 are usually not asking which model is smartest. They are asking which workflow…

  12. I Put a Free Local AI Model on My Mac to Cut My API Bill. Here’s the Honest Verdict.

    Towards AI · 2026-07-12

    It beat my expectations on some tasks and fell short on others — and where the line falls is the actually useful part.

  13. Stop Using Claude Projects Like a Folder

    Towards AI · 2026-07-12

    Anthropic's Claude is now offering a new project-based organization system for users, moving beyond simple folder structures.

  14. Prompt Caching Is the Cheapest Way to Cut Your AI Bill, and Most People Still Do Not Use It

    Towards AI · 2026-07-12

    A recent piece highlights prompt caching as an underutilized technique for reducing inference costs in large language models, akin to memoization in traditional programming.

  15. AI Boosts Research Careers but Flattens Scientific Discovery

    Hacker News · 2026-07-12

    AI tools are accelerating the pace at which researchers can publish papers, but this efficiency might be leading to a decrease in truly novel scientific breakthroughs.

  16. Apple’s failed self-driving car program left a legacy of powerful AI chips

    The Verge · 2026-07-12

    Apple's self-driving car program never really got off the ground, but it may have been what made the company's chips the powerful AI performers they are.

  17. Agent Protocols for Building Enterprise AI Assistants

    Towards AI · 2026-07-12

    AI Engineer Interview PreparationContinue reading on Towards AI »

  18. LinkedIn is the undisputed king of long-form AI slop, according to a study spanning five platforms

    The Decoder · 2026-07-12

    One in four longer social media posts is entirely AI-generated, according to a Pangram analysis. LinkedIn leads with 41 percent of long-form posts flagged as AI-written.

  19. I Gave Three AI Agents One Shared Toolbox — Here’s What MCP Actually Fixes

    Towards AI · 2026-07-12

    Building a working multi-agent research assistant with the Model Context Protocol — tested, running, and ready to forkContinue reading on Towards AI »

  20. Building a Multi-Agent Support Ticket Triage With LangGraph

    Towards AI · 2026-07-12

    Route tickets and draft replies locally with Ollama or

  21. Claude Code now has a built-in browser that lets the AI read, click, and type on external websites

    The Decoder · 2026-07-12

    Claude Code now has a built-in browser that lets the AI open, read, and interact with web pages directly inside the development environment.

  22. RAG vs Fine-Tuning Explained: What They Actually Do and When to Use Each

    Towards Data Science · 2026-07-12

    Two techniques, two different problems, and why the question is not really "which one wins"

  23. HTTP Has a New Method Now (After 16 Years!), and it lands where AI agents were bleeding

    Towards AI · 2026-07-12

    The Hypertext Transfer Protocol (HTTP) has gained a new method, `SEARCH`, for the first time since 2007.

  24. Yes, Your Browser Can Run Its Own Tiny AI. Here’s How

    Towards AI · 2026-07-12

    A new browser extension allows users to run smaller, on-device AI models directly within their web browser, eliminating the need for cloud processing for certain tasks.

  25. I Built an AI Agent Team That Fixes Its Own Mistakes — Here’s the Full, Tested Code

    Towards AI · 2026-07-12

    Most agent tutorials show you the happy path. This one shows you what happens when an agent fails — and how to make the system heal itself…

  26. No-code SaaS is the Highest-Leverage Passive Income AI Business of the Decade

    Towards AI · 2026-07-12

    The rise of no-code platforms has enabled individuals to build and monetize AI-powered Software-as-a-Service (SaaS) applications without extensive programming knowledge

  27. ‘These are some of the most complex structures ever created’: how tech reporting moved into the physical world

    The Guardian AI · 2026-07-12

    The Guardian’s global tech reporting team are investigating the impact of the vast datacentres being built to power the AI revolution.

  28. OpenAI, Meta, SpaceXAI Compete for More Cost-Efficient AI Models

    Bloomberg · 2026-07-12

    Three prominent artificial intelligence developers released new models over the past week.

  29. Apple’s M6, M7 and M8 Chips Show How AI Is Reshaping the Company

    Bloomberg · 2026-07-12

    Apple's latest M-series chip roadmap, codenamed M6, M7, and M8, indicates a significant internal prioritization of on-device AI processing power.

  30. Paper Walkthrough — Geometrically-Constrained Agent for Spatial Reasoning

    Towards AI · 2026-07-12

    A new paper introduces a geometrically-constrained agent capable of performing complex spatial reasoning tasks