💻 Apple unveils MacBook Pro M5 Max

Good Morning! Apple just dropped the MacBook Pro M5 Pro and M5 Max, and it’s claiming AI speeds that once required massive Nvidia rigs now fit inside a laptop. Plus, you will learn how to fully configure Claude as your ultimate AI workspace.

Plus, in today’s AI newsletter:

• Apple Unveils MacBook Pro M5 Max

• Gemini 3.1 Flash-Lite Gets a Major Upgrade

• OpenAI Launches GPT-5.3 Instant

• How to Set Up Claude the Right Way

• 4 new AI tools worth trying

AI NEWS

💻 Apple Unveils MacBook Pro M5 Max

Apple refreshed the MacBook Pro lineup with its new M5 Pro and M5 Max chips, boosting AI performance, unified memory, and storage, while raising prices slightly.

• Up to 4× faster LLM prompt processing and major gains in AI image and video generation

• Up to 128GB unified memory with bandwidth reaching 614GB/s

• Faster SSD speeds (up to 14.5GB/s) and improved battery life up to 24 hours

• Starts at $2,199 (14” M5 Pro) and goes up to $3,899 (16” M5 Max)

Apple’s biggest weakness in AI used to be slow inference speeds. With M5 Max, it’s positioning the MacBook Pro as a portable AI workstation, capable of running large models locally without cloud dependence. If the performance claims hold up, Apple’s long-term silicon strategy may finally be paying off.

AI MODELS

⚡ Gemini 3.1 Flash-Lite Gets a Major Upgrade

Google DeepMind released a preview of Gemini 3.1 Flash-Lite, boosting intelligence by 12 points while keeping blazing-fast speeds and a 1M-token context window.

• Scores 34 on the Intelligence Index (+12 vs 2.5 Flash-Lite) while generating 360+ tokens/sec

• Beats top-tier models on multimodal benchmarks (78% on MMMU-Pro)

• Delivers first token 2.5× faster and 45% higher output speed vs Gemini 2.5 Flash

• Pricing more than tripled: $0.25/M input tokens and $1.50/M output tokens

Gemini 3.1 Flash-Lite now punches above its weight class in reasoning and multimodal tasks, making it one of the strongest lightweight models available. But with output costs jumping 3×, developers will need to balance performance gains against significantly higher inference bills.

AI MODELS

💬 OpenAI Launches GPT-5.3 Instant

GPT-5.3 Instant is the new default ChatGPT model, focused on making everyday interactions feel more natural and useful. OpenAI says it improves answer accuracy, reduces unnecessary refusals, and delivers stronger web search performance.

• Hallucinations reduced by up to 26.8% in web searches and 19.7% in internal knowledge tasks

• More natural writing style, with fewer robotic or overly cautious responses

• Mixed safety results: better than GPT-5.1 Instant, but slightly weaker than GPT-5.2 Instant in some areas

• Rolling out to all ChatGPT users now; available via API as “gpt-5.3-chat-latest”

Instead of chasing flashy new features, OpenAI is refining the core ChatGPT experience, improving reliability, tone, and search quality. Small iteration upgrades like this quietly shape how millions experience AI every day.

HOW TO AI

🗂️ How to Set Up Claude the Right Way (And Leave ChatGPT Behind)

In this tutorial, you will learn how to fully configure Claude as your ultimate AI workspace, moving beyond simple chatbots to leverage local file reading, in-spreadsheet AI, and specialized plugins for seamless knowledge work.

🧰 Who is This For

• AI researchers helping small business owners integrate models to improve efficiency and profits

• Founders and recent grads hunting for their next big startup idea who need a powerful brainstorming partner

• Finance enthusiasts and traders wanting to automate complex Excel modeling and P&L spreadsheets

• Anyone who wants to stop writing endless prompts and start using an AI that actually reads their files and asks clarifying questions

STEP 1: Install Claude Cowork and Build Your Context Files

Instead of treating Claude like a generic text box, you need to set up Claude Cowork, a desktop app that interacts with your actual files. Go to the official Anthropic website, download the app, and ensure you are on a Pro plan.

Once installed, create a dedicated folder on your computer filled with Markdown (.md) files containing your business context, writing style, or specific project briefs. Point Cowork to this folder so it absorbs your background knowledge automatically. Instead of writing long prompts every time, your first prompt just needs to be: "Read the uploaded files completely, and ask me clarifying questions before executing the task."

STEP 2: Select Opus 4.6 and Enable Extended Thinking

Most people fail to get elite results because they are using the default, lighter models. Open your Claude chat, click the model selector at the bottom, and switch to Opus 4.6.

Crucially, you must toggle on the Extended Thinking feature. This acts as an internal monologue, forcing the AI to reason through complex tasks, like analyzing market trends, coding, or mapping out software integrations, before it generates a response. This simple toggle drastically reduces hallucinations and improves logic.

STEP 3: Bring AI Directly Into Your Spreadsheets

If you work heavily with financial models, trading data, or client metrics, stop uploading CSVs to a chat window where the formatting gets destroyed. Open Microsoft Excel, go to the Add-ins menu, and install the official "Claude by Anthropic" extension.

Once you sign in with your paid account, Claude lives directly inside your workbook. It can read exactly what is in specific cells. You can ask it to explain complex formulas, locate #REF errors, or even build a dynamic 3-year projection spreadsheet from scratch, all while keeping your cell structures and formulas perfectly intact.

STEP 4: Install Plugins and Generate Artifacts

To turn Claude from a general assistant into a specialized expert, open Cowork, go to the Plugins directory, and install the packs that fit your daily work (like Data Analysis, Sales, or Marketing). Once active, you can use simple slash commands (like /build-dashboard) to instantly execute complex workflows without needing to write paragraph-long instructions.

Additionally, you can ask Claude to generate interactive Artifacts. Simply ask it to "Create an interactive HTML calculator" or "Build a visual comparison chart," and Claude will generate a fully functioning, interactive UI directly inside your chat window that you can use and download.

Google details Coruna, an exploit kit used to hijack iPhones via malicious websites; iVerify suggests it may have been originally built for the US government.

In an all-hands meeting, Sam Altman said OpenAI does not “get to make operational decisions” regarding how its technology is used by the Department of Defense.

Google adds new Android features, including custom calling cards, real-time location sharing in Google Messages, and luggage tracking with partner airlines.

OpenAI is developing an alternative to GitHub; the project is nascent, and the decision came after OpenAI engineers experienced an increase in outages.

🍌 Nano Banana 2: Google’s new top AI image model

💼 Claude Cowork: Anthropic’s team AI platform with plug-ins

🤏 Qwen3.5 Small: Alibaba’s tiny AI models that perform like systems 13× bigger

🎥 Replit Animation: Turn text prompts into animated videos

THAT’S IT FOR TODAY

Thanks for making it to the end! I put my heart into every email I send, I hope you are enjoying it. Let me know your thoughts so I can make the next one even better!

See you tomorrow :)

- Dr. Alvaro Cintas