☕🤖 Perplexity Unveils Computer, Its Most Ambitious Product Yet
PLUS: Learn how to start with Claude Code...
Hey fam 👋 Luis & Rui here with your Thursday AI coffee break.
Perplexity just dropped a full AI computer that runs entire projects end-to-end, Meta bet $60 billion on AMD chips to challenge Nvidia, and the Pentagon threatened to blacklist Anthropic over AI safeguards.
Grab your coffee!
(3-minute coffee break ☕)
AI News around the world 🌍
🖥️ Perplexity launched Computer, a general-purpose AI system that researches, codes, and deploys entire projects using 19 specialized models. (continue reading)
💰 Meta signed a $60 billion deal with AMD for AI chips over five years, taking direct aim at Nvidia’s GPU dominance. (continue reading)
⚠️ Defense Secretary Hegseth gave Anthropic until Friday to drop AI safeguards or face Pentagon blacklist and the Defense Production Act. (continue reading)
⚡ Notion introduced custom AI agents that autonomously handle Q&A, task routing, and status reports across Slack, Linear, and Figma. (continue reading)
🤖 Cursor launched cloud agents that build software in isolated VMs, now producing over 30% of the company’s merged pull requests. (continue reading)
Trending AI Tools 🛠
🔍 Lorka AI: Boost your productivity with AI-powered automation designed to simplify your workflow and help you work smarter. Discover how Lorka AI helps you save time, optimize tasks, and get better results faster. 👉 Try Lorka AI now!
⚡ Miniloop: Describe go-to-market tasks in plain English and get production-ready automations across HubSpot, Apollo, Gmail, and more.
🧠 Mindrop: AI-powered second brain for WhatsApp, Telegram, and web that automatically summarizes, tags, and organizes every link you save.
🔬 LLM Council: Open-source tool by Andrej Karpathy that sends queries to multiple LLMs for anonymous peer-review, then synthesizes a final answer.
Tutorials 📝
Career & Capital 💼
AI Job Board
🏢 EisnerAmper: AI Solution Architect - $120,000 to $200,000 (view job)
📊 Cotality: Principal, Machine Learning Scientist - $135,300 to $185,000 (view job)
🛡️ Northrop Grumman: Principal Signal/Image Processing Engineer - $125,300 to $234,600 (view job)
Investments in AI
💎 MatX raised $500M Series B to build custom AI chips competing with Nvidia, backed by Jane Street and Leopold Aschenbrenner’s fund. (link)
📊 Basis secured $100M at $1.15B valuation for AI agents that autonomously handle accounting tasks, led by Accel with GV and Lloyd Blankfein. (link)
🇫🇷 Mistral AI made its first acquisition, buying cloud startup Koyeb to become a full-stack AI platform with on-premises deployment. (link)
Prompt of the Week 🧠
Meeting Notes to Action Items → Turn rambling calls into accountability systems that actually get followed up on.
Prompt:
You are a senior operations manager who is obsessive about execution and follow-through. I'm going to paste in notes (or a transcript) from a meeting. Your job is to extract every commitment, decision, and action item, even the ones that were implied but not explicitly stated.
For each action item, provide:
1. Task: What specifically needs to be done (rewrite vague items into clear, completable tasks)
2. Owner: Who is responsible (flag if ownership was unclear)
3. Deadline: When it's due (flag if no deadline was set and suggest one)
4. Context: The 1-sentence reason this matters or what it blocks
5. Priority: High / Medium / Low based on business impact and dependencies
Then generate two outputs:
OUTPUT 1 - Action Item Table
A clean markdown table sorted by priority, then deadline.
OUTPUT 2 - Follow-Up Message
Write a concise follow-up message I can send to all attendees that:
- Summarizes the 2-3 key decisions made
- Lists each person's action items with deadlines
- Flags any unresolved items that need async decisions
- Ends with the next check-in date
Tone: direct, professional, zero fluff. If anything in the notes is ambiguous, call it out explicitly rather than guessing.
Here are the meeting notes:
[PASTE YOUR MEETING NOTES OR TRANSCRIPT HERE]Where to use it: Paste into ChatGPT, Claude, or Gemini immediately after any meeting. Works with rough notes, polished minutes, or raw transcripts from Otter.ai, Fireflies, or similar tools. Especially powerful for cross-functional meetings where ownership tends to get fuzzy.
📮 Need help with AI? Here’s how we can help:
✅ Become a PRO member → Get exclusive AI tutorials, ready-to-use prompts, custom GPTs, and +50 automation workflows. Upgrade here
🧁 Book an AI Audit → We’ll review your business and show you where AI can actually make an impact, cut unnecessary costs, and build workflows that work. Your AI growth partner. Book your AI Audit
📣 Sponsor Us → Put your brand in front of 96,000+ founders and professionals. Work with us
🙌 That’s a wrap for today’s AI Break
Thanks for hanging out → don’t forget to hit subscribe, drop your thoughts below and share this update with your fellow AI enthusiasts.
Luis & Rui
PS: Catch us on X @sousa_brothers for daily updates on the latest AI developments.
PPS: If you enjoyed today’s AI Break, sharing is caring (and our algorithms are 😉 programmed for gratitude)
Perplexity Computer using 19 specialized models feels like the logical next step after Claude Code and Cursor. I've been running parallel agents in my own workflows for research tasks, but having it packaged as one interface saves the mental overhead of orchestrating everything manually.
I know I'm not going deep with Perplexity, but it could be a great alternative for people who's not comfortable with Claude Code or the similar.
Thanks for sharing the meeting notes prompt, very useful!
I just can’t see how Perplexity wins here in the long term (which is a couple of years in AI time). The labs are all going to build the same product, and once computer use and planning become effectively solved problems there’s no real differentiator Perplexity can offer. At that point the labs will be able to win on cost since Perplexity has to make their own profit on top of the token cost.