- Everyday AI
- Posts
- Ep 683: 5 Practical AI Workflows That Actually Matter
Ep 683: 5 Practical AI Workflows That Actually Matter
5 AI workflows that matter, OpenAI’s new Audio Tech, ChatGPT Takes aim at Apples App Store, xAI Acquires third Data Hub and more
Everyday AI
January 02, 2026
👉 Subscribe Here | 🗣 Hire Us To Speak | 🤝 Partner with Us | 🤖 Grow with GenAI
Outsmart The Future
Today in Everyday AI
8 minute read
🎙 Daily Podcast Episode: AI shouldn’t feel overwhelming—it should feel useful. These five workflows show how to make it practical, fast, and valuable. Give it a watch/read/listen.
🕵️‍♂️ Fresh Finds: Pickle releasing Smart Glasses, Samsung releases AI Projector, Chinas 'Silicon Valley' is building a fortune telling AI Robot and more Read on for Fresh Finds.
🗞 Byte Sized Daily AI News: OpenAI’s new Audio Tech, ChatGPT Takes aim at Apples App Store, xAI Acquires third Data Hub and more Read on for Byte Sized News.
đź’Ş Leverage AI: Aside from 5 AI Workflows that actually matter, we go over how AI should change your future of work mindset. Keep reading for that!
↩️ Don’t miss out: Did you miss our last newsletter? We Covered: Updates on Apple’s AI efforts, Softbank’s $41B investment in OpenAI, xAI launces Grok Business and more Check it here!
Ep 683: 5 Practical AI Workflows That Actually Matter
Might internal memos be a thing of the past?
When you can just build something as fast as writing a memo about it, why wouldn't you just build the demo?
In this episode of Everyday AI, we sit down with Google Cloud’s Richard Seroter to break down five simple ways to use AI with Google. No technical background needed.
We talk faster research, better learning, building ideas without overthinking, and why “demos over memos” might change how teams work.
If you want practical, no-BS ways to actually use AI in your day‑to‑day, this one’s worth a listen.
Also on the pod today:
• Gemini Deep Research: smarter searches 🔎
• NotebookLM: personalized learning assistant 📒
• Turning inbox chaos into insights 📬
It’ll be worth your 33 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Here’s our favorite AI finds from across the web:
New AI Tool Spotlight – Resell-AI Get gives you instant AI-powered valuations in seconds, 70 Lives is Real-time tax tracking and smart pricing for beauty pros, CareerFinder asks you to answer a few intuitive questions and they will turn your strengths and goals into a clear, step-by-step career roadmap.
AI Glasses — Pickle 1 smart glasses challenge Meta’s lead—shipping starts Q2 2026.
AI Portable Projector — Samsung’s AI projector aims to replace your TV—see what’s new.
AI Daily Tasks — AI agents are quietly taking over everyday tasks—are you ready?
China AI Tech — China’s tech hub bets big on robots—and AI that predicts your future.
1. OpenAI Gears Up for Audio-First AI Devices 🔊
OpenAI is doubling down on audio tech, uniting teams to overhaul its voice models and prepping a next-gen personal device aimed for early 2026, according to The Information.
The company wants its new audio AI to sound more natural, handle interruptions like a real conversation, and even talk over you—something current tech can’t do. Industry giants and startups alike are racing to make audio the go-to interface, signaling a shift away from screens as the main way we interact with technology.
2. OpenAI Takes Aim at Apple’s App Store Monopoly 💰
OpenAI is pushing into the app space, letting users access services like Instacart and Spotify straight through ChatGPT, in a move that could eventually shake up Apple’s dominance.
The rollout, still in its infancy, shows flashes of promise but is riddled with glitches and limited functionality, according to recent Wall Street Journal tests. CEO Sam Altman boldly claims his real competitor isn’t Google, but Apple, as OpenAI eyes a future where chatbots replace traditional apps and even the iPhone itself.
3. xAI Supercharges AI Ambitions with Third Data Hub Near Memphis 🔌
Elon Musk’s AI startup xAI is turbocharging its computing power, announcing the acquisition of a third building—playfully called “MACROHARDRR”—to nearly double its AI training capacity.
The new facility, set outside Memphis, will help xAI scale its Colossus supercomputer cluster toward rivaling the biggest players like OpenAI and Anthropic. As xAI eyes 1 million GPUs and a dedicated natural gas plant, the move underscores just how fiercely Musk’s company is racing to keep pace in the global AI arms race.
4. NVIDIA Doubles Down on AI Open Source Push 🤖
NVIDIA just announced its acquisition of SchedMD, the developer behind the crucial open-source Slurm workload manager, signaling a bold move to cement its leadership in high-performance AI computing.
The company has pledged to keep Slurm open and vendor-neutral, ensuring broad compatibility in the fast-evolving AI hardware landscape. This follows recent launches of NVIDIA’s Alpamayo-R1 vision language model and expanded open-source tools for autonomous and physical AI research.
5. DeepSeek’s New Paper Rethinks AI Training 📰
China’s DeepSeek has just published a new AI training method that analysts are already calling a major breakthrough, signaling fresh momentum in the global race to build bigger and better models. According to Business Insider, the approach aims to let large language models scale up while staying stable and efficient, tackling one of the biggest technical headaches in AI development.
Experts say the paper shows DeepSeek’s growing confidence and internal firepower, and could quickly push rivals to explore similar techniques. The timing is notable, as it fuels speculation that the research will underpin DeepSeek’s next-generation models, even as the company faces chip shortages and fierce competition from Western labs.
🦾 How You Can Leverage:
Richard Seroter pretty much said that writing requirements documents is turning into a waste of time.
Richard is the Senior Director and Chief Evangelist at Google Cloud and he dropped a straight up reality check today on Everyday AI that most enterprises aren’t ready to hear: The era of "pixel pushing" documents to explain an idea is over.
Why?
Wild timeline we’re living in.
Richard broke own 5 time-saving AI workflows, which we’ll bullet point at the end of this lil newsletter, but that’s not the real level up here.
Instead, we are begging you to listen to today’s show, as Richard gives us a legit masterclass in changing the way we think about work.
1. The "Demos Over Memos" Mandate 🚀
"Demos over memos."
That’s the new motto inside Google’s engineering product areas. (We friggin love this, BTW.)
Richard explained that teams often waste months debating specifications, formatting tables in docs, and arguing over theoretical implementation.
Stop it shorties.
With tools like Gemini CLI and Code Assist, non-technical staff can now "vibe code" functional prototypes. You don't need to know Python syntax. You need to know your intent.
The most important programming language today is English. (Or, the language you speak.)
If you can articulate what you want, you can build a working demonstration. This shifts the entire organizational bottleneck.
You aren't waiting for a developer to validate if an idea is possible. You build the ugly version, prove the concept, and then hand it to engineering to make it production-ready.
Try This:
Identify a manual process that frustrates you this week. Maybe it's reconciling a specific spreadsheet or tracking team updates.
Instead of writing a ticket to IT or complaining in a meeting, open a coding assistant (like Gemini Code Assist or even standard LLMs) and describe exactly what you want a simple web app to do.
Use natural language. "I want a simple page that takes this CSV and highlights rows where column B is greater than column C." Build the ugly version. Prove it works.
2. Deep Research vs. Confirmation Bias 🔥
That’s a trap ya’ll.
The real power of Gemini Deep Research isn’t speed—though Richard noted it analyzed 150+ sites in six minutes for him—it’s the ability to act as an adversarial partner.
Say buh bye to sycophantic AI and hello to a deep researcher that challenges your preconceived notions.
Richard said he used Deep Research to challenge his own project assumptions. Instead of spending three days clicking blue links to find data that supported his thesis, he asked the AI to find every reason why he was wrong.
Six minutes.
That’s the difference between a three-day research project that leads to a failure, and a coffee break that saves your quarter. The competitive advantage isn't getting answers faster. It's asking better questions.
Try This:
Next time you’re about to launch a project or pitch a strategy, stop.
Open Gemini Deep Research (or your preferred reasoning model) and feed it your core thesis. Then, explicitly ask it to act as a skeptic and find data-backed reasons why this approach will fail.
You aren't looking for validation here—you're looking for the blind spots that usually cost you three months of work. If the AI can dismantle your strategy in 30 seconds, you just saved yourself a fiscal quarter of headaches.
3. Context Engineering > Prompt Engineering ⚡
They’re focusing on the wrong thing.
The future isn't about being a "prompt engineer." It's about being a "context engineer." Tools like NotebookLM and the new agentic capabilities in Gemini don't work because of magic words; they work because of the data you feed them.
Richard highlighted a massive shift: Moving from stateful one-off chats to long-running contexts.
When you feed an agent your calendar, your email, your 50-page SOPs, and your style guides, you aren't asking it to be smart. You're asking it to be you.
This is how you solve the "blank page" problem. You don't ask AI to write a strategy from scratch. You feed it the last three years of performance reviews, the new product documentation, and the competitor analysis, then ask it to find the gap.
Try This:
Take that massive PDF contract, technical manual, or compliance document you’ve been avoiding.
Dump it into NotebookLM right now. Don't just ask it for a summary. Treat it like a new hire. Ask: "Based on these rules, where is our current workflow non-compliant?" or "What is the weirdest clause in this agreement?"
Breakdown: Richard’s 5 Simple AI Strategies to Supercharge Your Workflow 👇
Gemini Deep Research for Analysis: Use Gemini to quickly synthesize info from across the web (and your files) for smarter, faster research.
NotebookLM for Exploration: Ground AI in your own docs/data to learn, organize, and digest information your way.
Gemini CLI & Code Assist to Build: Use command line and code tools to prototype, automate, and build—no advanced dev skills required.
Jules for Background Work: Offload tasks to autonomous AI agents that work in the background, then review their output when ready.
AI Rolling Out Everywhere: Google is adding AI features across its products (Search, Sheets, Drive, etc.), making everyday tasks smarter and easier.
Reply