Good morning, welcome to this week’s Lore Brief, your 3-minute brief of the most important moves in AI and tech.
This issue is brought to you by Factory, creators of Droid, the frontier software development agent. Available to anyone, with any model, in any IDE/terminal.

Anthropic releases Claude Fable 5, its first public Mythos-class model → The new model delivers major gains in coding, long-horizon tasks, and scientific work, and quickly took the top spot on several Arena leaderboards. However, Anthropic added heavy safeguards that automatically downgrade risky queries (especially in cybersecurity and biology) to the weaker Opus 4.8, while keeping the unrestricted version (Mythos 5) limited to trusted partners. The decision sparked strong backlash from users who feel the public model was intentionally nerfed. Read more here →
Tim Cook unveils a significantly smarter Siri at WWDC → At Apple’s annual developer conference, the company previewed major upgrades to Apple Intelligence, headlined by a much more capable Siri that better understands context across apps and devices, powered by Google’s Gemini. The update is one of the biggest improvements to Siri in years. Read more here →
Dario Amodei calls for stronger frontier AI regulation → In a new essay, Anthropic’s CEO argues that AI is advancing too quickly for current policy frameworks and pushes for mandatory third-party testing on cyber, bio, and autonomy risks. The piece has drawn heavy criticism from the AI community, with many accusing it of regulatory capture aimed at slowing down competitors. Read more here →

Several examples of how good Fable is:
Fable for game design:
“Branding matters” post gone viral, left image generated with Grok:

OpenAI files confidential S-1 for potential IPO → The company submitted paperwork to go public but said it hasn’t decided on timing yet and may stay private for a while to complete certain plans more easily. OpenAI noted it expects the filing to leak, which is why it made the announcement now. Read more here →
River AI launches to build personal, user-owned AI → Ex-xAI researcher Igor Babuschkin announced a new company focused on creating AI systems that are fully owned and shaped by individual users rather than big corporations. The startup aims to build a personal AI stack that aligns with each user’s values and operates on their own terms. Read more here →
Google releases DiffusionGemma → Sundar Pichai announced an experimental open diffusion-based text model built on Gemma 4 that generates entire blocks of text in parallel instead of one token at a time. It delivers up to 4x faster inference while remaining fully open under Apache 2.0. Read more here →
Thariq shows how he used Fable 5 to edit its own launch video → The Anthropic engineer shared a video explaining how he used Claude Fable 5 to handle the entire video editing workflow — including transcription, color grading, and rendering — without using traditional video editing software. See more here →
Google brings Gemini into Apple’s developer tools → Gemini models are now available to Apple developers through Apple’s Foundation Models framework and directly inside Xcode. Developers can switch between local and cloud inference using a shared API and get agentic coding assistance from Gemini inside Xcode. Read more here →
Hiten Shah on why fixing AI output matters more than generating it → The founder emphasized that the real value in using AI often comes from the correction loop rather than the initial output. He’s running a session on how to identify the right tasks and build practical skills with tools like Claude and ChatGPT. Read more here →

