Meta Just Put a Giant AI on Your Desk

Hola Decoder😎

If someone forwarded this to you and you want to Decode the power of AI and be limitless, then subscribe now and Join Decode alongside 30k+ code-breakers untangling AI.

🧠 Meta releases an 816B AI model that runs locally on apple macs

Meta has released Muse Spark Open, an 816-billion-parameter open-weight multimodal AI model designed for coding and autonomous tasks, with Apple helping bring local execution to Mac Studio hardware.

The Shift:

816 Billion Parameters Run Without the Cloud - Meta says two M5 Ultra Mac Studios with 1TB of combined memory can run Spark Open using Apple's updated MLX toolkit, enabling local AI processing at an unusually large scale.

AI Handles Coding and Video Editing - In Meta's demonstrations, Spark Open fixed issues in an iOS app and edited video using DaVinci Resolve, showing how the model can work across software development and creative applications without cloud processing.

Meta Trails Chinese Models in Rankings - Spark Open scored 39 on Artificial Analysis' Intelligence Index, behind China's GLM 5.3 (45), Kimi K3 (44) and Qwen3.8 (40). It performs competitively on multimodal and knowledge tasks but trails those models in coding.

Western Open AI Models Are Expanding - Spark Open is the third major Western open-model announcement this month, following Reflection AI's Beam and Mistral's 1-trillion-parameter Le Chonk, whose model weights are scheduled for October 27.

Meta's release demonstrates how increasingly large AI models can operate on local hardware, potentially reducing dependence on cloud infrastructure for demanding tasks. However, Chinese open-weight models continue to lead several coding benchmarks, keeping competition focused on both capability and deployment efficiency.

Together with Omnisend

Your BFCM Plan Is Still Waiting on Better Prompts

How much time goes into pulling last year’s reports and explaining them to AI? If every useful answer takes another round of clarification, your BFCM prep barely moves forward.

Omnisend’s BFCM AI playbook gives you detailed prompts for the work you need to get done, so you can spend less time figuring out what to ask.

Use last year’s results to decide which campaigns deserve another run and check deliverability before increasing send volume.

Then turn those findings into a campaign calendar your team can work from.

Better yet, Omnisend connects with Claude or ChatGPT, so you can analyze performance and create editable email templates without leaving the chat.

New to Omnisend? Its email and SMS platform serves 150,000+ ecommerce brands, reporting an average $79 for every $1 spent. Migration is free and takes up to five business days.

Start now, with time to put the answers to work.

Try the BFCM prompts

🤑How to run your finances with ChatGPT

Connect ChatGPT to your real bank data and get answers based on what you actually spend, not generic budgeting advice.

Step 1: Connect your bank accounts - Open ChatGPT > Plugins > search for your bank or a financial data connector. Plaid-connected plugins let ChatGPT read your real transactions. If your bank is not listed, export a CSV from your bank’s app and upload it directly.

Step 2: Find hidden subscriptions - “Go through my last 90 days of transactions. Find every recurring charge, group them by category, and flag anything I am paying for more than once or anything I probably forgot about.”

Step 3: Build a budget from real numbers - “Based on my actual spending from the last 3 months, build me a zero-based budget where every dollar has a job. Show what I am currently spending in each category and what I should be spending to hit [savings goal] per month.”

Step 4: Spot where the money is going - “What are my three biggest spending categories? Where am I overspending compared to a reasonable benchmark for my income? What is the single biggest change I could make this month to save more?”

Step 5: Plan ahead - “Based on my income and fixed expenses, tell me how long it would take me to save [amount] if I cut [category] by 20%. Show me two or three different scenarios.”

Real data in. Real answers out. Not guesses.

📊 Harvard study finds gender gaps in AI use, training and media coverage

Harvard researchers analyzed surveys of more than 5,400 American adults, browsing data, and 78,000 news articles, identifying differences in how men and women use AI, receive workplace support, and appear in AI-related coverage.

The Shift:

Women Report Lower AI Adoption - 29% of women said they had never used AI, compared with 22% of men. However, browsing data showed similar rates of trying AI, with differences emerging in time spent using it.

Women Receive Less AI Training at Work - 64% of employed women reported receiving no AI training, compared with 49% of men. Researchers associated workplace training and encouragement with greater AI usage and confidence across both groups.

AI Knowledge and Confidence Differ - Men answered more AI knowledge questions correctly, while women selected "not sure" more frequently. However, men were also more likely to claim familiarity with fictional AI figures.

Men Dominate AI News Coverage - Across 78,000 news articles, men appeared in AI headlines 14.6 times more often than women. Researchers partly attributed this to male-led companies, although the imbalance also appeared among quoted experts.

The findings suggest AI adoption gaps involve more than access to technology. Workplace training, confidence and public representation also matter. Expanding training opportunities and encouraging participation could help address differences as AI becomes more common in professional settings.

Together with Playbook Pro

The Playbook for Profit That Funds Itself

Great operators don’t just count cash, they make it multiply. They use timing and leverage to turn everyday revenue into rapid expansion.

Financial Alchemy is your cheatsheet packed with real operator insights and ready-to-use templates. Inside, you’ll learn how to:

💰 Master cash-conversion cycles to grow while suppliers fund your inventory, just as Wayfair scaled with a –34-day cash conversion cycle.

📊 Use the Cash Flow Simulator to forecast scenarios and uncover up to $100K in hidden operating cash within a single quarter.

🧮 Pinpoint profit levers with the Unit Economics Analyzer to cut CAC payback windows from 6 months to as little as 3 months.

🤝 Negotiate better terms using the Supplier Toolkit, a framework inspired by Skinny Pasta’s strategy that extended supplier terms to 120 days and fueled 100%+ Y-o-Y growth.

Start compounding capital instead of chasing outside funding.

Get your ebook now and turn your cash flow into the most powerful growth engine!

🏆 Tools you Cannot Miss:

📱 ClipMatch – Uses AI to turn footage from a phone's camera roll into social media content.

🎞️ AVE – Provides local-first video editing on Mac using transcripts and AI-assisted editing tools.

🎥 Vizard Agent – Uses an AI video workflow to help create and edit different types of video content.

🎨 Scumble – Offers open-source AI inpainting for editing selected regions of images instead of regenerating an entire visual.

🎵 Music Dawn – Records musical ideas and helps musicians develop recordings into more finished-sounding tracks.

🚀 Quick Hits

🌐 OpenAI upgraded its ChatGPT Chrome extension to access and compare open browser tabs. ChatGPT Work can also fill forms, send messages, download files and organize tabs, while the sidebar now supports full conversations.

🛡️ Elon Musk said SpaceX will probably allow independent evaluators to assess its AI systems, potentially following Anthropic’s approach. The tentative commitment comes amid growing calls for external oversight of advanced AI models.

⚖️ The FTC has reportedly subpoenaed OpenAI and Anthropic over AI security incidents, requesting documents on safety practices, privacy policies and marketing materials. The agency is also reportedly planning to question senior executives about potential consumer protection violations.

🛡️ Anthropic is granting six METR researchers employee-level access to its AI systems for independent safety evaluations. METR will publish findings on risks and incidents, although Anthropic can require redactions to protect sensitive information.

🏦 Bank of America CEO Brian Moynihan said the bank hired 1,000 contractors to fix software vulnerabilities exposed by Anthropic’s Mythos and other AI models, highlighting how advanced AI is changing cybersecurity requirements.

👨‍💻 What’s trending on socials

🚗 Tesla Roadster reportedly performs midair flip during unveiling - A viral video shared by Collin Rugg claims the Tesla Roadster launched off a ramp, flipped midair and landed safely during its unveiling. The post describes the car as a Tesla-SpaceX collaboration, though the stunt remains unverified.

📦 Elon Musk considers launching an Amazon competitor - Elon Musk suggested starting a competitor to Amazon after reportedly blocking Grok Bot from placing orders on the platform. Musk posted, “Maybe we should start a competitor to Amazon,” but announced no concrete plans.

🎮 Pokémon HeartGold fully decompiled in one month - A developer claims Pokémon HeartGold’s entire 4.1 MB of compiled code has been decompiled into C/C++, taking approximately one month and an estimated $56,000 in AI API costs.

💘 Former Bumble and Tinder employees build anti-AI dating app - Former Bumble and Tinder employees reportedly created Hoppy, an anti-AI dating app inspired by MySpace, featuring customizable profiles with videos and memes. The app has reportedly attracted over 50,000 waitlist signups.

Thanks for Decoding with us🥳

Your feedback is the key to our code! Help us elevate your Decode experience by hitting reply and sharing your input on our content and style.

Keep deciphering the AI enigma, and we'll be back with more coded mysteries unraveled just for you!