In partnership with
🖱
Perplexity just made AI agents free to run.
Their new "Portable Computer" platform lets you execute complex AI tasks on your own hardware no cloud costs, no data leaks, no per-token billing.
But wait, there's more:
Meta's cooking up a new model called "Watermelon" that supposedly matches GPT 5.5
Apple dropped M6 and M5 Ultra chips built specifically for on-device AI
I'm Alex. Welcome to L8R by Innov8.
Let's dive deep 🐰
In today's post:
Perplexity's Portable Computer lets you run AI agents on your own machine
Meta's "Watermelon" model drops in October
Apple's M6 and M5 Ultra chips bring serious AI horsepower to your desk
Perplexity's Portable Computer
Perplexity just changed the game by letting you run their smart AI agents right on your own desk.
They launched a new tool called Portable Computer.
Now you can do heavy AI work without paying cloud token fees or leaking your private data.
The details:
You need serious hardware like the new Nvidia DGX Spark or a PC with an RTX GPU and at least 24GB of VRAM.
It runs powerful 27-billion parameter models like Qwen 3.8 or Perplexity's own PPLX 27B totally offline.
If the local AI gets stuck on a hard step, it asks your permission before sending just that one task to the cloud.
Your private files stay on your machine, and you pay zero token costs for all the local work.
Why it matters:
Cloud AI costs add up fast when agents run in loops. By moving the brain to your local machine, Perplexity makes AI agents a real tool for everyday work. Companies will love this because their private data never leaves the building. It turns your fancy hardware into a serious AI worker.
From Our Partner
Cut Lead Review From Hours To Minutes
Sign up for a free trial of Attio, the agentic CRM.
Ask Attio to build a daily workflow that surfaces the deals that need your attention today, like anything with a stage change, a recent reply, or a new signal in the last 24 hours.
Review your pipeline in Claude, synced live from Attio via MCP.
That's it.
Meta's 'Watermelon' AI Wants to Crush GPT-5.5
Meta is getting ready to drop a massive new AI model called Watermelon this October.
Mark Zuckerberg wants to prove his company can match OpenAI's best tech.
This new model will also power expensive AI agents that surf the web for you.
The details:
Meta's superintelligence chief Alexandr Wang claims Watermelon matches OpenAI's GPT-5.5 on internal tests.
The company used ten times more computing power to train this model compared to their last release in April 2026.
Meta is also launching an AI agent platform called Hatch that can use apps like DoorDash and Reddit for you.
Hatch will test a crazy new subscription model, charging users up to $199.99 per month for premium access.
You will also see third-party AI agents pop up inside WhatsApp starting this week.
Why it matters:
Meta is tired of giving away AI for free. They spent billions on chips and now they want your money. If Watermelon really matches GPT-5.5, Meta becomes a top player in the AI race. They are moving past simple chatbots into agents that actually do your chores.
💡 L8R's Take:
Charging 200 bucks a month for an AI agent is a massive gamble. Nobody will pay that much just to order DoorDash unless the AI is flawless. Meta needs to prove Watermelon is actually smart before asking for our wallets.
From Our Partners
See how AI delivers fast, secure, personalized experiences
Watch Fin and Plaid now on demand to see how AI can resolve issues directly in customer conversations. This includes everything from reducing bank-linking friction, enabling proactive support, and helping customers complete real financial tasks.
Apple Drops M6 and M5 Ultra Chips for Local AI
Apple just dropped a massive hardware update that turns your desk into an AI supercomputer.
The new M6 and M5 Ultra chips are built entirely for heavy AI workloads and always-on agents.
You do not need to pay for cloud AI when your local machine is this powerful.
The details:
The new M6 chip is Apple's first 2-nanometer processor, making it crazy fast and power-efficient.
You can get the M6 in a refreshed Mac mini starting at $899.
The real beast is the M5 Ultra chip inside the new Mac Studio, starting at $5,499.
The M5 Ultra supports up to 512GB of unified memory, which is exactly what you need to run giant AI models locally.
Apple added Neural Accelerators to every single GPU core to speed up AI tasks.
Why it matters:
Cloud AI gets expensive fast. Plus, you have to worry about data privacy. With 512GB of memory on the M5 Ultra, you can run massive, frontier-class AI models right on your desk. This pushes the whole industry toward agentic computing, where your local machine does the heavy lifting for you all day long.
🚀 Quick L8R Summary
Perplexity Portable Computer: Run AI agents on your own machine no cloud costs, no data leaks.
Meta's Watermelon Model: Meta's October release might match GPT-4o, signaling a push beyond ads into AI monetization.
Apple M6 & M5 Ultra: Apple's new 2nm chips turn Macs into local AI powerhouses for faster, private workflows.
📩 Innathe L8R engane undarunnu 👇?
We read every reply - just reply to this email and let us know how we can improve !
Appo adutha L8R il kanaam bie…👋