ANTHROPIC π₯: Mythos class models are expected to become generally available after getting stronger safeguards, according to the latest Project Glasswing update.
> And in the near future, once weβve developed the far stronger safeguards we need, we look forward to making Mythos-class models available through a general release.
Soon? π
> And in the near future, once weβve developed the far stronger safeguards we need, we look forward to making Mythos-class models available through a general release.
Soon? π
β€10π₯5 1
Perplexity open-sources Bumblebee security scanner
Perplexity released Bumblebee, an open-source, read-only scanner for macOS and Linux that inventories risky packages, extensions, and AI tool configs on developer machines to help security teams spot supply-chain exposure fast.
π #perplexity @testingcatalog
Perplexity released Bumblebee, an open-source, read-only scanner for macOS and Linux that inventories risky packages, extensions, and AI tool configs on developer machines to help security teams spot supply-chain exposure fast.
π #perplexity @testingcatalog
TestingCatalog AI News
Perplexity open-sources Bumblebee security scanner
Perplexity has released Bumblebee, an open-source security scanner for developer machines that helps detect risky packages and extensions.
β€3π₯3 2
Anthropic is launching a new "Gift a week of Cowork" referral program. Users will get 3 invite codes to share with others.
> Send a friend a free week of Cowork. If they love it and subscribe, you'll get β¬10 in usage credits.
> They'll also get Claude in: Excel, PowerPoint, Chrome, Claude Code
Who wants to give Cowork a test? π
> Send a friend a free week of Cowork. If they love it and subscribe, you'll get β¬10 in usage credits.
> They'll also get Claude in: Excel, PowerPoint, Chrome, Claude Code
Who wants to give Cowork a test? π
π₯5β€3π1
Anthropic prepares Mythos 1 for Claude Code and Claude Security
Anthropic signals broader access for Claude Mythos, shifting from prior restrictions. Project Glasswing now covers wider organizations, while Mythos 1 preview and new Claude Security tools point to an upcoming guarded release.
h/t @goodiers
π #claude @testingcatalog
Anthropic signals broader access for Claude Mythos, shifting from prior restrictions. Project Glasswing now covers wider organizations, while Mythos 1 preview and new Claude Security tools point to an upcoming guarded release.
h/t @goodiers
π #claude @testingcatalog
TestingCatalog AI News
Anthropic prepares Mythos 1 for Claude Code and Claude Security
Anthropicβs Claude Mythos is moving closer to public release as Mythos 1, with new leaked evidence and an updated Claude Security dashboard for Enterprise customers.
Grok Build CLI is now available to SuperGrok and X Premium+ users! It is cool that you can also use it to search through X and use it as a read-only X client.
> curl -fsSL http://x.ai/cli/install.sh | bash
One more agent for your team π
> curl -fsSL http://x.ai/cli/install.sh | bash
One more agent for your team π
Anthropic plans Claude memory update with new Memory Files
Anthropic appears set to revamp Claudeβs memory with optional file-based βMemory Files,β replacing one rolling note with structured records by topic.
π #claude @testingcatalog
Anthropic appears set to revamp Claudeβs memory with optional file-based βMemory Files,β replacing one rolling note with structured records by topic.
π #claude @testingcatalog
TestingCatalog AI News
Anthropic plans Claude memory update with new Memory Files
Anthropic is testing a dual-mode memory system for Claude, offering Classic and File Memory options.
π3β€2
π¨ AI News | TestingCatalog
Anthropic plans Claude memory update with new Memory Files Anthropic appears set to revamp Claudeβs memory with optional file-based βMemory Files,β replacing one rolling note with structured records by topic. π #claude @testingcatalog
This media is not supported in your browser
VIEW IN TELEGRAM
ANTHROPIC π₯: Claude will soon receive a new file-based memory upgrade, offering users the option to choose between Memory Files and Classic memory.
> Organized notes Claude writes as you chat and reads when they're relevant. Browse and edit them anytime.
This feature appears to be a new iteration of the previously discovered "Knowledge Bases" and more closely resembles what memory works in always-on agents like OpenClaw and Hermes.
Considering a potential future debut of Claude Conway, Memory Files feature is likely an important preparation step.
> Organized notes Claude writes as you chat and reads when they're relevant. Browse and edit them anytime.
This feature appears to be a new iteration of the previously discovered "Knowledge Bases" and more closely resembles what memory works in always-on agents like OpenClaw and Hermes.
Considering a potential future debut of Claude Conway, Memory Files feature is likely an important preparation step.
SPEACEXAI π₯: The next Grok model is expected to be ready for public release in 2-3 Elon weeks.
> 1.5T V9-Medium base model in comparison to 0.5T v8-Small, used for Grok 4.3
> Cursor data being used for supplementary training
Grok 5? πππ
> 1.5T V9-Medium base model in comparison to 0.5T v8-Small, used for Grok 4.3
> Cursor data being used for supplementary training
Grok 5? πππ
β€5π4 1
GOOGLE π₯: Antigravity now has Gemini 3.5 Flash (Low) option available with a 45% fewer token consumption in comparison to Medium variant.
This addition follows Gemini quota reset as well!
This addition follows Gemini quota reset as well!
16β€22π6 4π΄1
Anthropic to introduce personal AI Fluency scorecard in Claude
Anthropic is testing an AI Fluency scorecard in Claude settings that analyzes user behavior across sessions, scores 11 collaboration habits, and offers targeted guidance, turning its research framework into a personal assessment tool.
π #claude @testingcatalog
Anthropic is testing an AI Fluency scorecard in Claude settings that analyzes user behavior across sessions, scores 11 collaboration habits, and offers targeted guidance, turning its research framework into a personal assessment tool.
π #claude @testingcatalog
TestingCatalog AI News
Anthropic to introduce personal AI Fluency scorecard in Claude
Anthropic is testing an AI Fluency scorecard in Claude, letting users assess and track their AI skills directly within the appβs settings panel.
β€5 2π₯1
MiniMax M3 has been teased π₯
> MiniMax M3 will be based on a new Sparse Attention architecture
> MiniMax M3 is expected to be open source
Soon? π
> MiniMax M3 will be based on a new Sparse Attention architecture
> MiniMax M3 is expected to be open source
Soon? π
β€9π₯5 2 1
GOOGLE π₯: AI Studio Build will soon get support for Themes, where users will be able to choose between 8 pre-defined presets or create their own.
Design MD support would be nice π
h/t @Thomasguka
Design MD support would be nice π
h/t @Thomasguka
β€8π4
Helio launches AI-powered team workspace in Beta
Helioβs public beta turns plain-language goals into AI teams in under 60 seconds. These AI colleagues work in existing channels, handle multi-step tasks, log decisions, learn nightly, and require human approval for high-stakes actions.
π #sponsored @testingcatalog
Helioβs public beta turns plain-language goals into AI teams in under 60 seconds. These AI colleagues work in existing channels, handle multi-step tasks, log decisions, learn nightly, and require human approval for high-stakes actions.
π #sponsored @testingcatalog
TestingCatalog AI News
Helio launches AI-powered team workspace in Beta
Helio is in public beta, letting teams build a working AI team in 60 seconds with zero code or deployment required.
β€4π3π₯1π―1