In partnership with
🖱
Meta just gave its AI agent the keys to your Mac.
Muse can now dig through your local files, launch apps, and actually do things on your desktop not just chat about them.
This is AI moving from helpful assistant to active participant on your computer.
But wait, there's more:
A startup built an AI that ditches text generation entirely for pure decision-making
SpaceXAI's new transcription model claims double the accuracy at the same dirt-cheap price
I'm Alex. Welcome to L8R by Innov8.
Let's dive deep 🐰
In today's post:
Meta brings Muse to your Mac finally, an AI that touches your files
TypeSafe's Jev beats GPT-5.6 without writing a single sentence
SpaceXAI's new transcription model: twice the accuracy, same ten cents
Meta's New AI Agent Wants to Run Your Mac
Meta just dropped a desktop version of its AI agent that actually does your work for you.
Muse for Mac can dig into your local files, read your emails, and organize your folders.
It is a massive step up from normal chatbots that just spit out text.
The details:
Meta released Muse for Mac on September 17, letting the AI interact directly with Apple apps like Mail, Messages, Calendar, and Notes.
You can tell the agent to find lost files, build end-of-day summaries, and clean up your messy folders while it runs in the background.
The tool runs inside a totally isolated virtual machine to keep your data safe and asks for your permission before doing risky things like deleting files or sending texts.
The mobile version of Muse hit 600,000 downloads in its first five days after launching on September 8, showing massive consumer demand.
The desktop app is currently only available in the United States and uses a freemium model with premium tiers costing between $20 and $100 per month.
Why it matters:
We are finally moving past simple chat windows.
AI is becoming an active worker on your computer instead of a passive tool.
This puts Meta in a direct fight with OpenAI and Google to control your desktop experience.
Whoever builds the best local agent wins the next era of computing.
💡 L8R's Take:
Giving Meta access to my local Mac files sounds like a privacy nightmare, even with their new secure virtual machine.
But the convenience of an AI organizing my messy folders might just win me over.
Apple needs to wake up, because Meta is eating their lunch right on their own hardware.
From Our Partner
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.
Jev AI Crushes GPT-5.6-Luna in Speed and Cost
Forcing giant AI models to spit out perfectly formatted JSON is a massive waste of time and money.
A new startup called TypeSafe AI just fixed this with a model called Jev.
It is a blazing fast AI built only for making quick decisions, not writing poetry.
The details:
TypeSafe AI launched Jev on September 15, 2026, backed by a $40 million seed round.
Jev does not generate text, code, or summaries at all.
It looks at your data and instantly returns structured choices, scores, or yes/no answers.
It answers in about 100 milliseconds, which is way faster than the 700-plus milliseconds of traditional models.
It costs just $0.042 per million input tokens, and output tokens are completely free.
Why it matters:
We are finally splitting AI tasks into two smart buckets.
Jev acts like your brain's fast 'System 1' for quick choices like sorting data or catching spam.
You can save expensive 'System 2' models like GPT-5.6-Luna for deep thinking.
This cuts out the messy retry loops developers build just to get structured answers.
💡 L8R's Take:
I absolutely love this approach.
Using a frontier model to output a simple yes or no is like using a sledgehammer to crack a nut.
Jev will become the default router for every major AI app by the end of the year.
From Our Partners
Chemical-Free Farming Is Now Open To Investors
For decades, farmers have relied on herbicides and heavy equipment to control weeds.
Greenfield Robotics is working to change that.
BOTONY robots navigate crop rows and mechanically remove weeds, giving farmers an alternative to blanket herbicide applications and intensive tillage.
The potential benefits are significant:
Clean food supply
Reduce reliance on agricultural chemicals
Reduce unnecessary soil disturbance
Help farmers protect crops from weed competition
Support more sustainable farming practices
And the market is enormous.
Greenfield estimates more than 250 million acres of broadacre cropland in the U.S., with more than 100 million acres in its initial serviceable market.
With the Reg A+ now live, investors can participate in Greenfield Robotics' growth story.
This Reg A+ offering is made available through StartEngine Primary, LLC, member FINRA/SIPC. Please read the Offering Circular and related disclosures before investing. This investment is speculative, illiquid, and involves a high degree of risk, including the possible loss of your entire investment.
SpaceXAI Drops Grok Voice 2.0 for Ten Cents an Hour
SpaceXAI just dropped a massive upgrade to its speech-to-text model, and it costs almost nothing.
Grok Voice Transcribe 2.0 promises double the accuracy of the old version without raising the price a single cent.
This is a huge win for developers who need cheap, flawless audio transcriptions.
The details:
SpaceXAI launched Grok Voice Transcribe 2.0 on September 18, 2026, keeping the price at just $0.10 per hour for batch processing.
The new model claims to double the accuracy of version 1.0 and easily handles noisy phone lines and thick local accents.
Premium features like speaker tracking, word-level timestamps, and mid-sentence language switching are included at no extra cost.
Atlassian is already using the new model to power video transcriptions for its Loom platform.
Why it matters:
Audio data is usually messy and hard to process.
Most cheap transcription tools fail when they hear background noise or heavy accents.
SpaceXAI is fixing this by using real-world training data from Tesla cars and customer support calls.
They are turning high-end voice AI into a cheap, everyday tool.
💡 L8R's Take:
Ten cents an hour is practically free.
SpaceXAI is going to crush the competition if this model actually works as well as they claim.
If you build apps with voice features, you need to switch to this API right now.
🚀 Quick L8R Summary
Muse for Mac: Meta's AI agent now lives on your desktop and can mess with your local files.
TypeSafe's Jev: A new model that skips text generation entirely just gives you classifications and scores, fast.
Grok Voice 2.0: SpaceXAI claims their new transcription model is twice as accurate for the same $0.10/hour price.
📩 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…👋