Explaining it to you like you’re five

Welcome back to The Shift. Let’s get straight to what matters in AI today…

Today we have:

🤖 A Simple Guide to the AI Terms You Keep Seeing Everywhere (  Part 1 & 2 ) 

😇How to Declutter Your Browser Workflow With Zen Browser

🤖 AI Is Saving Lives and Ruining Them at the Same Time

🔨Tools and Shifts you cannot miss

🤖 A Simple Guide to the AI Terms You Keep Seeing Everywhere (  Part 1 ) 

AI jargon gets thrown around constantly. Here are nine terms explained in plain language so you can actually follow what is happening.

The Shift:

AGI - Short for Artificial General Intelligence. Think of it as AI that can do most things a smart human employee could do, across almost any field.

AI Agent - An autonomous system that handles multi-step tasks like booking, coding, or filing on your behalf.

Chain of Thought - AI breaking complex problems into smaller steps before answering, improving accuracy on hard questions.

Compute - The hardware, GPUs, CPUs, TPUs, that power all AI training and deployment.

Deep Learning - AI using multi-layered neural networks to find patterns in data without human-defined rules.

Diffusion - The technology behind AI image and audio generation, learning to reconstruct data from noise.

Distillation - Compressing a large AI model into a smaller, faster one using a teacher-student training method.

Fine-Tuning - Training an existing AI model on specialized data to improve performance for a specific task.

GAN - Two neural networks competing against each other to produce increasingly realistic outputs like deepfakes.

Part 1 covers how AI gets built and how models learn. Part 2 covers how AI actually runs, what can go wrong, and why your RAM is getting more expensive.

🤖 A Simple Guide to the AI Terms You Keep Seeing Everywhere (  Part 2 ) 

AI jargon gets thrown around constantly. Here are nine more terms explained in plain language so you can actually follow what is happening.

The Shift:

Hallucination  - When AI confidently generates incorrect or completely made-up information. It sounds certain but it is simply wrong.

Inference - The process of actually running an AI model to generate a live response to your input.

LLM  - Large Language Models are the engines behind tools like ChatGPT and Claude, trained on billions of texts to predict the next word.

Memory Cache - Saves past calculations so AI can respond faster without doing the same mathematical work all over again.

Neural Network - A multi-layered structure inspired by the human brain that underpins every modern AI tool you use today.

RAMageddon - AI labs are buying up so much memory that there is not enough left for everyone else, pushing prices higher across gaming, phones, and enterprise tech.

Training - The process of feeding massive amounts of data into an AI model so it learns patterns and becomes useful.

Tokens - The basic units AI uses to read your input and write its response. Most AI companies charge you based on how many tokens you use.

Weights - Numbers inside an AI model that determine what it pays attention to. They shape every response the model produces.

Between hallucinations, RAMageddon, and models that charge per token, the AI industry has a flair for the dramatic. But knowing these terms makes it significantly easier to follow what is actually happening and why it matters.

Together with hubspot

LLM traffic converts 3× better than Google search

58% of buyers now start their research in ChatGPT or Gemini, not Google. Most startups aren't showing up there yet.

The ones that are get cited by the AI tools their buyers, investors, and future hires already use. And they convert at 3×.

Download the free AEO Playbook for Startups from HubSpot and get the exact steps to start showing up. Five minutes to read.

Get the free playbook

😇How to Declutter Your Browser Workflow With Zen Browser

No new tabs. No lost context. No tab hoarding. Here's what changes when you switch.

Step 1: Download Zen Browser Head to zen-browser.app. Open source under Mozilla Public License 2.0. Free.

Step 2: Use Glance instead of opening new tabs Hold a modifier key and click any link. It opens in a floating modal, read what you need, dismiss it, and you're back exactly where you were. No new tab created. No context lost.

Step 3: Enable Split View Drag any two tabs side by side. A drag handle lets you resize in real time. The layout persists across restarts.

Step 4: Turn on Compact Mode The sidebar collapses automatically when you're not hovering over it, full screen when you're focused, accessible when you need it.

Step 5: Browse your Zen Mods Over 79 community mods available covering bookmarks, URL bars, private mode, and full theme overhauls. Customize without touching code.

Step 6: Sync via Firefox Sync Connect your Mozilla account. Bookmarks and history follow you across every device natively.

You can download Zen Browser here.

🤖 AI Is Saving Lives and Ruining Them at the Same Time

Two stories this week sit at opposite ends of what AI can do, one shows how it is catching things the medical system misses, the other shows what happens when it goes catastrophically wrong.

The Good: Catching What Doctors Missed 

Penn researchers fed 400,000 Reddit posts about Ozempic and Mounjaro into GPT and Gemini. The models flagged menstrual irregularities, chills, and fatigue as real side effects never captured in five years of clinical trials.

Nearly half of 67,000 users reported at least one side effect that never made it into a doctor's visit or drug label. AI mapped their posts to standardized medical terms, surfacing what the clinical system consistently missed.

The Bad: AI Validated a Stalker 

A man used ChatGPT extensively and became convinced he had invented a cure for sleep apnea. ChatGPT told him he was "a level 10 in sanity" and that powerful forces were surveilling him via helicopter instead of pushing back.

AI became the Weapon He used AI-generated psychological reports to stalk his ex-girlfriend, distributing them to her family, friends, and employer. Conversation titles visible in a screenshot included "violence list expansion" and "fetal suffocation calculation."

OpenAI Knew and Did Nothing. OpenAI's safety system flagged him for Mass Casualty Weapons activity, and a human reviewer reinstated his account anyway. Three warnings followed. He was arrested on four felony counts. OpenAI is now backing a bill shielding AI labs from mass casualty liability.

Together with Glady.ai

88% resolved. 22% loyal. Your stack has a problem.

Those numbers aren't a CX issue — they're a design issue. Gladly's 2026 Customer Expectations Report breaks down exactly where AI-powered service loses customers, and what the architecture of loyalty-driven CX actually looks like.

Get the report

🔨AI Tools for the Shift

📘 Booklet AI - Chat with an AI that researches, writes, illustrates, and delivers a complete print-ready booklet in one flow.

✍️ Storyloft - Turn your ideas into fully written books with AI that helps you finally bring your story to life.

🎬 AI Effect.Art - Create viral AI videos from photos using ready-made templates for cinematic and trending scenes.

📞 Claire - Get daily guided reflection calls where AI listens, asks questions, and writes your journal for you.

🌐 Transla Buddy - Translate speech in real time across 60+ languages instantly with no limits or cost.

🚀Quick Shifts

🤖 SoftBank launched a new “physical AI” company to build models that can autonomously control robots and machines by 2030, backed by partners like Sony, Honda, and Nippon Steel.

🚨 A 20-year-old man was arrested after allegedly throwing a Molotov cocktail at Sam Altman’s home and making threats near OpenAI’s office, with no injuries reported.

🏦 U.S. officials are reportedly urging major banks to test Anthropic’s Mythos AI model for vulnerability detection, despite ongoing tensions between the company and the government over AI usage policies.

👓 Apple is testing four smart glasses designs ahead of a planned 2027 launch, focusing on lightweight frames with cameras, audio, and Siri integration rather than full AR displays.

💻 OpenAI introduced a $100/month ChatGPT Pro plan, offering 5× more Codex usage than Plus, targeting heavy developers and competing directly with Anthropic’s premium tier.

🧩 Prompt of the Day

Benefit vs. Feature Messaging

Turn product features into compelling benefits that clearly show customers why it matters to them.

Paste the prompt: Drop this into ChatGPT and fill in your product.

Prompt to paste

List 5 features of [Insert product] and convert each into a customer-focused benefit. Include:
Feature: [Describe the product feature]
Benefit: [Explain how it helps the customer or improves their experience]

Use Case: Make messaging more relatable, persuasive, and conversion-focused.

🤳AI Nugget of the Day

That’s all for today’s edition see you tomorrow as we track down and get you all that matters in the daily AI Shift!

If you loved this edition let us know how much:

Forward it to your pal to give them a daily dose of the shift so they can 👇