Astra vs Fable 5.1 Review Astra and Fable 5.1 are incredible models to use. Each excels in different areas, but if I had to choose one, it would be Fable 5.1. I’ve never felt so frustrated and amazed at the same time using a model as I have with Astra. At times, I think it couldn’t get any better. Other times, it feels like one of the worst models I’ve ever used. When prompting Astra, you have to be very specific. Otherwise, you’ll be disappointed with the output. It does exactly what you say, but sometimes that means doing things you didn’t intend. In a way, it feels unpredictable, which has made writing an AGENTS.md file difficult. Fable 5.1 is the smarter model. It understands intent and feels like it knows what I want, even when I don’t quite know yet. Working with this model is a joy. The main thing that makes Fable unusable is Anthropic’s ridiculously restrictive usage limits. They’re so bad that they make an otherwise great model frustrating to use. At the end of the day, Fable is the model I reach for to do real work, while Astra is the model I use to test things or handle tasks that require computer use. Do you feel the same, or do you disagree? Let me know in the comments. I’m curious to hear what others think.

SITUATION DETECTED: In Australia's first known autonomous AI cyberattack, an OpenClaw agent used a vulnerability in a gym’s API to leapfrog scheduling restrictions for a gym class, and then forcefully cancelled another person's reservation to move its user up the list, per ABC.

anyone remember MUI?

So apparently GPT 5.6 Sol might just randomly delete all of the files on your laptop 🚨 Here's how to safeguard yourself against it Paste this prompt into codex: Protect this machine from accidental permanent file deletion by Codex. Implement this as defense in depth, not merely as a written instruction. Requirements: 1. Inspect the current official Codex documentation for hooks, command rules, sandboxing, and approval settings before making changes. 2. Detect the operating system and identify its recoverable Trash or Recycle Bin command: - On macOS, use `/usr/bin/trash `. Do not add `--`, because macOS `/usr/bin/trash` treats it as a filename. - On Linux or Windows, verify an installed, reliable Trash or Recycle Bin mechanism before using it. - If no reliable recoverable mechanism exists, stop and explain what must be installed or configured. 3. Add a durable global Codex instruction: - Codex must never permanently delete a file or directory. - All removals must go to Trash or the Recycle Bin. - Permanent deletion requires the user to explicitly change this safety policy first. - Destructive Git operations such as `git clean`, `git reset --hard`, bulk checkout, or bulk restore must also be blocked. 4. Install a global `PreToolUse` hook that returns a hard `deny` before execution when it detects: - `rm`, `unlink`, `rmdir`, or `shred` - `find -delete` - `rsync --delete` - destructive Git cleanup or reset commands - common Python, Node, Ruby, or Perl deletion APIs - file deletion through `apply_patch` - delete-like MCP filesystem tools - nested destructive commands inside `bash -c`, `bash -lc`, `sh -c`, or `zsh -c` 5. The hook must direct Codex to the verified Trash or Recycle Bin command instead. Do not silently rewrite dangerous commands, because flags and shell expansion could be misinterpreted. 6. Add user-level Codex command rules that mark direct permanent-deletion commands as `forbidden`. Preserve existing rules and configuration. 7. Set safe global defaults when compatible with the existing configuration: - `sandbox_mode = "workspace-write"` - `approval_policy = "on-request"` Do not override managed policies or create conflicting permission configurations. 8. Preserve all existing configuration. Inspect files before editing, make narrowly scoped changes, and create recoverable backups of any configuration files that must be replaced. 9. Validate everything without risking real data: - Parse the hook and configuration files. - Feed simulated tool-call JSON into the hook. - Confirm dangerous examples are denied. - Confirm ordinary commands and the Trash command are allowed. - Test the Trash mechanism using a newly created disposable temporary file. - Run an end-to-end Codex test using `rm` with no arguments inside a read-only sandbox. Never test `rm` with a real target. - Confirm the command was blocked before shell execution. 10. Explain any limitations honestly. A Codex hook is a guardrail, not an absolute operating-system security boundary. 11. Check whether system backups are configured. Do not enable or modify backups without permission, but report clearly if no backup destination exists. 12. At completion, provide: - The files created or changed - The validation results - The exact one-time steps required to review and trust the hook - Any restart requirement - Any remaining risks Do not permanently delete anything while completing this task. If cleanup is necessary, move it to Trash or the Recycle Bin.

There are a class of AI Coding assets that IMO don't really belong checked into git: - PRD's - Research files - Decision maps - Implementation plans Folks who agree with me, what are you using instead?

they did it. they fucking did it. i am AGASP seeing what this amazing team has pulled and the tremendous value they keep bringing to the community. let me explain, it is not a secret that to a lot of us the rapid advancement of ai has been a blessing and a curse. sure, you can ship faster, but that inevitably lowers down the overall quality output and process that a lot of creators put into their work, bastardizing their craft by leaving ai do all the thinking for them... this isn't that. they very thoughtfully made such a smooth integration with LLMS where they LITERALLY looked at all the boring shit that's part of the design workflow and said "let's make this better". you can skip through the 1% that feels like going from 92 to 99 in runescape (iykyk) and still have complete control over your work, actually empowering and boosting your already existing workflows instead of replacing them !! some use cases they shared: - adding staggered animations to pages - making different component variations - MAKING PAGES RESPONSIVE HELLO - adding styles to pages and finding layers without styles - tagging layers, pages and styles to update - light and dark theming - reverting and applying changes to iterations - creating different versions - creating style pages - iterating on components / sections - switching up repetitive copy and placeholder images - running audits to check for seo and accessibility OUT OF WORDS, THIS IS WHAT AI WAS SUPPOSED TO BE ABOUT, GETTING RID OF THE BORING SHIT SO WE CAN FOCUS ON WHAT MATTERS, GODSPEED @framer LFGGGGGGG! ⚡

Agents: AI built into every part of Framer to help you become extremely productive, not replace you. Agents can design pages, update your site, manage your CMS, improve SEO, fix responsiveness, understand analytics, and, most importantly, help you ship. framer.com/

The canvas is about to change

this week we got editor integration working with opencode where it sends the context to it automatically now it knows which file is open and which lines you've selected we use the same protocol that Claude Code uses so if you have that extension this will just work

Fun fact. It takes 32 minutes to burn through 5 hours worth of Opus credits on the Claude Code Max 20x plan. 12 agents trying to fix 3,528 Typescript errors.