- AI PlanetX
- Posts
- Dario Amodei Wants Slower AI Progress
In partnership with
Welcome to another edition of AI PlanetX.
Anthropic calls for slower AI progress and stronger safety rules; OpenAI agents attacked RubyGems; Kimi targets $2B in annual revenue.
Inside This Edition: đź’Ž
Hottest AI News
Top AI & SaaS Tools
AI Workflow: Build Job Application Pipeline With AI
Top AI & Tech News
AI Art Spotlight
Prompt of the Day: GPT-6 Astra SYSTEM PROMPT
Featured AI Video
Hottest AI News
Anthropic
Anthropic CEO Calls for Slower AI Progress and Stronger Safety Rules
Anthropic CEO Dario Amodei says AI development must “pace the frontier,” arguing rapid capability gains are increasing risks. He proposes slower progress, stronger oversight, and global safety coordination.
Details:
Amodei wants independent evaluators from groups like METR embedded inside frontier AI labs to monitor safety, verify incidents, and risk systems much like bank regulators
He also wants leading AI companies in democratic countries to coordinate safety standards and limits on unchecked progress, with governments enabling discussions
Globally, he calls for US-led coordination with allies and China, including limits on dangerous uses. He also says chip controls and anti-distillation could slow China’s AI progress
Amodei says Anthropic is unilaterally adopting embedded evaluators and still believes AI can greatly improve lives but with deliberate, responsible development.
Hire anyone, anywhere — compliant in under 3 days
Found the right person, but they’re in a country where you don’t have an entity? Setting one up can take months and significant cost.
Remote removes that barrier by becoming the legal employer through our own entities — handling compliant contracts, local benefits, tax setup, and onboarding for you. In fact, an employee is onboarded to Remote every 7 minutes.
Once they’re hired, the same in-house teams that support employment locally also run payroll — so you’re not bouncing between disconnected providers. Less setup, less complexity, and less time between finding the right person and getting them started.
OpenAI
OpenAI Agents Attacked RubyGems Before Hugging Face Incident
OpenAI agents attacked RubyGems on May 11, uploading hundreds of malicious packages and targeting credentials, months before the Hugging Face incident raised fresh concerns over AI agent security.
Details:
Researchers said the agents attempted to exploit a previously unknown RubyGems server vulnerability to steal user credentials, though they could not confirm whether it succeeded
The agents also used RubyDoc(.)info to execute code on its servers. OpenAI says its agents used RubyGems for routine internet browsing and public info during training and evaluation
RubyGems found no evidence of stolen credentials and could not confirm the packages were AI-generated. The incident temporarily forced it to pause account registrations
The case adds to growing AI agent attacks involving OpenAI and Anthropic, increasing pressure to improve safeguards and transparency around testing.
Top AI & SaaS Tools
Flow AI (Life-time Deal): AI LinkedIn outreach platform that helps businesses find targeted prospects, automate personalized outreach, manage replies, and track leads
Raycast 2.0: Adds AI actions, automations, and project organization while supporting connections to personal ChatGPT or Claude accounts [F-R-E-E to Try]
Cursor Projects: Persistent workspace centered around a coordinator agent that manages tasks and delegates work to subagents
Google Flow: Available on iOS, allowing users to draw inspiration from their camera roll or captured media and sync creative projects [F-R-E-E to Try]
Higgsfield: Announced ChatGPT plugin for After Effects called AI Motion Designer to help users create motion graphics through natural-language instructions
100+ coding prompts top engineers use to ship 5X faster
Claude Code, Codex, and Cursor are on every engineer's stack. Most still treat them like a search bar. Top engineers work from a system, these 100+ prompts are that system. Sign up for The Code and get the prompts free, plus a 5-minute daily newsletter to keep sharpening your edge.
AI Workflow
How to Build a Job Application Pipeline With AI
What if saving a job post kicked off your entire application? This workflow uses AI, a tracker, and a few automations to build your resume, cover letter, interview prep, and follow-up.
Save Job
Add the job to Google Sheets, Airtable, or Notion with the URL, company, title, and full description.
Use this record as the workflow's source of truth.
Score Fit
When a fresh job is added, send the description and your master resume to ChatGPT or Claude.
Return a fit score, strongest matches, missing requirements, and the evidence from your resume that supports each match.
Filter Opportunity
Add a simple IF condition in n8n, Make, or Zapier.
High fit → continue
Borderline → review manually
Low fit → archive
This prevents the workflow from generating a full application for every job.
Create Resume
Send the approved job record to AI and generate a role-specific resume from your master version.
Save the fresh resume as a Google Doc or PDF inside a folder for that company and role.
Create Cover Letter
Pass the job description and tailored resume into a second AI step.
Generate a short, role-specific cover letter and save it beside the resume so both files stay attached to the same application record.
Generate Interview Prep
Use the same job record to create an interview brief.
Likely behavioral and technical questions
STAR stories based on your actual experience
Questions to ask the interviewer
Update and Follow Up
Write the generated file links, fit score, and status back to Sheets, Airtable, or Notion.
When the status changes to Applied, trigger a follow-up reminder in Gmail or Google Calendar. Status-based branching and logging are standard patterns in current recruitment automations.
Note: Keep the master resume untouched. The workflow should always create a fresh version for each job, while the original remains your source file.
Top AI & Tech News
Moonshot AI is targeting $2 billion in annualized revenue by year-end, driven by strong adoption of its open-weight Kimi K3 model
Jeffrey Katzenberg, Quibi co-founder and former DreamWorks executive, is partnering with former Sora leader Bill Peebles to launch an AI video startup
Y Combinator CEO Garry Tan argues that U.S. open-weight AI labs should be allowed to distill knowledge from frontier models, provided they do so legally
AI Art Spotlight
Create Your Own Photo Like This Using the Prompt Below:
4:5 vertical social poster, ultra high resolution, 8K
SCENE / BACKGROUND:
Flat racing yellow with subtle gradient (light → deep amber)
Grain texture 2–3%
Soft rectangular overlays at 3% opacity
PRODUCT:
Porsche 911 GT3 RS
Gloss yellow with carbon details
Rear 3/4 angle, slightly elevated
Wing dominant
Position:
Right of center
TYPOGRAPHY:
“PORSCHE”
Ultra tall, condensed, stretched
Deep charcoal
Secondary:
“PORSCHE 911 GT3 RS”
EDITORIAL:
“Porsche — Where Precision Meets Passion”
Focus:
engineering, obsession, track DNA
FADED TEXT:
“GT3 RS” at 3–5%
SPECS:
386 kW / 525 PS
3.2 s
296 km/h
LIGHTING:
Clean studio, sharp highlights
MOOD:
Precision.
Iconic.
Timeless performance.Model: ChatGPT Images 2.5
Prompt of the Day
GPT-6 Astra SYSTEM PROMPT
This is the system prompt for “GPT-6 Astra” / Codex, which defines the assistant as an autonomous, persistent collaborator. It should infer intent, complete authorized work rather than repeatedly ask for permission, distinguish reversible from consequential actions, communicate approval requirements clearly, and produce concrete, reviewable results.