Daily AI briefing

AI Briefing: Claude Overtakes GPT as 'Rogue' AI Deletes Files

Claude Fable 5 dominates GPT-5.6 in new benchmarks, while OpenAI's Sol model faces backlash for deleting files and Meta is sued over AI-driven layoffs.

  • Briefing date: 2026-07-15
  • Editorial AI analysis
  • The AI Wrap

Stories covered on 2026-07-15

  1. Nvidia Expands Toyota AI Partnership for Smart Cities, Factories

    Bloomberg · 2026-07-15

    Nvidia Corp. will supply its artificial intelligence hardware and software to Toyota Motor Corp.

  2. AI Price War Fractures Single-Vendor Stacks. Chinese APIs Capture 46% of Traffic

    Towards AI · 2026-07-15

    OpenAI dropped three models at once, xAI priced Grok 4.5 at two dollars per million tokens, and systems architects are actively building…

  3. NVIDIA Introduces New Jetson Thor Computers to Advance Mainstream Robotics and Edge AI

    NVIDIA AI Blog · 2026-07-15

    General-purpose robots and autonomous machines are moving from research labs to real-world mass-market deployment, creating demand for compact

  4. Agentic orchestration: Enterprise AI organizations have a deployment problem, not a platform problem — and most are calling chatbots agents

    VentureBeat · 2026-07-15

    Across 101 enterprises, agent orchestration is consolidating onto model-provider platforms — Anthropic’s Claude leads by a wide margin

  5. Evolution of NLP: TF-IDF to Agents

    Towards AI · 2026-07-15

    How search progressed from exact-term matching to systems that decide how to gather evidence.

  6. xAI sues a man for using Grok to generate CSAM ‘deepfakes’

    The Verge · 2026-07-15

    The Elon Musk-owned xAI is suing a South Carolina man who allegedly used the company's Grok AI chatbot to generate child sexual abuse material (CSAM).

  7. Anthropic Plans IPO Investor Meetings as Mega-Listing Nears

    Bloomberg · 2026-07-15

    Anthropic PBC is seeking to meet with investors ahead of its potential mega-IPO, according to people familiar with the matter

  8. Murati’s Thinking Machines Releases First AI Model for Broad Use

    Bloomberg · 2026-07-15

    Thinking Machines Lab is releasing its first AI model, more than a year after former OpenAI executive Mira Murati founded the artificial intelligence startup.

  9. Loop Engineering in Claude Code: Let the Agent Run Itself

    Towards AI · 2026-07-15

    Claude's newfound ability to execute code autonomously represents a significant step beyond static text generation, enabling more dynamic and interactive AI applications.

  10. Soofi Consortium Releases Soofi S 30B-A3B: An Open Hybrid Mamba-Transformer MoE Foundation Model For German And English

    MarkTechPost · 2026-07-15

    Soofi S 30B-A3B is an open Mamba-Transformer MoE model activating 3.2B of 31.6B parameters for German and English The post Soofi Consortium Releases Soofi S 30B-A3B: An Open Hybrid

  11. Migration to Agent-First Architecture for Enhanced Security

    Towards AI · 2026-07-15

    A significant shift is underway as AI development increasingly prioritizes agent-first architectures, moving away from monolithic models towards more modular

  12. AI slop movies are the new direct-to-video cash grabs

    The Verge · 2026-07-15

    This weekend, cinephiles across the world will march to their local theaters to feast their eyes on Christopher Nolan's new adaptation of The Odyssey.

  13. OpenAI is now using AI to attack its own AI, and it's working better than humans ever did

    The Decoder · 2026-07-15

    OpenAI's internal GPT-Red model finds successful attacks in 84 percent of test scenarios through self-play training. Human red teamers manage just 13 percent.

  14. Amid hardware legal battle, OpenAI releases a $230 keyboard for Codex

    TechCrunch · 2026-07-15

    OpenAI, which is in the middle of a legal battle with Apple over hardware trade theft allegations, just released a light-up keyboard designed to be paired with its agentic coding

  15. I Gave Claude a Memory That Survives Between Conversations — Here’s the MCP Server That Does It

    Towards AI · 2026-07-15

    A tested, running MCP server that gives any AI agent persistent long-term memory — it remembers facts and preferences, reinforces the ones…

  16. Murati's Thinking Machines Releases Open-Weights 975B Parameter LLM

    Hacker News · 2026-07-15

    Thinking Machines has released Inkling, an open-weights large language model with 975 billion parameters.

  17. AI Isn’t Smarter Than a Baby—Yet

    WIRED · 2026-07-15

    Babies are tremendous learning machines, and key advances for AI may soon be found in the architecture of their little brains.

  18. OpenAI launches a physical keypad for controlling agents

    Engadget · 2026-07-15

    OpenAI's collaboration with keyboard maker Work Louder is available to order today.

  19. Apple is reportedly shopping for AI chip companies

    Engadget · 2026-07-15

    Apple's current M2 Ultra-powered servers don't seem to be cutting it, so the company is shopping around for more power.

  20. Trump rails against New York’s statewide datacenter moratorium

    The Guardian AI · 2026-07-15

    AI-friendly president shared a post saying governor Kathy Hochul should scrap the one-year policy ‘IMMEDIATELY’ Donald Trump railed against the New York governor, Kathy Hochul

  21. Thinking Machines amps up its bet against one-size-fits-all AI with its first open model, Inkling

    TechCrunch · 2026-07-15

    It's the company's first public proof point after a year and a half spent building AI infrastructure largely out of public view.

  22. The Hidden Cost of Letting AI Agents Write Your Tests

    Towards AI · 2026-07-15

    I asked Claude to write tests for a payment processing module last week. It generated 47 tests in under two minutes. Every single one…

  23. Suno snatched millions of songs from YouTube, Genius, and Deezer

    The Verge · 2026-07-15

    Suno data obtained in a hacking incident has exposed that the AI music generator was trained by scraping millions of songs and lyrics from online audio platforms

  24. GPT-5.6 Sol reportedly disproves a 30-year-old statistics conjecture in 90 minutes after humans couldn't crack it

    The Decoder · 2026-07-15

    A University of Pennsylvania statistics professor used OpenAI's GPT-5.6 Sol Pro to disprove a central open conjecture about the Benjamini-Hochberg method in roughly 90 minutes.

  25. Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer

    MIT Technology Review · 2026-07-15

    OpenAI has built an LLM super-hacker called GPT-Red that it uses as a sparring partner to help its other models boost their defenses against cyberattacks.

  26. Seamless Remote-to-Edge AI Benchmarking: Overcoming the 3-Tier Network Bottleneck

    Towards AI · 2026-07-15

    This research introduces a novel benchmarking framework designed to streamline AI model evaluation across distributed network architectures

  27. Whatnot acquires Shaped to power real-time live shopping recommendations

    TechCrunch · 2026-07-15

    Livestream shopping platform Whatnot has acquired AI startup Shaped, a machine learning company focused on real-time recommendations and search.

  28. Hack suggests AI music generator Suno scraped YouTube for training data

    TechCrunch · 2026-07-15

    The hacker used an employee's credentials to access source code, which revealed how Suno scraped decades of audio.

  29. Don’t Let Claude Grade Its Own Homework

    Towards Data Science · 2026-07-15

    Cross-provider PR review with Codex in GitHub Actions, and why a second opinion from a different lab beats any self-review

  30. Microsoft patches record number of security vulnerabilities, citing its use of AI

    TechCrunch · 2026-07-15

    Microsoft's monthly release of security fixes, dubbed Patch Tuesday, resolved a record 570 security vulnerabilities across the company's product line