- Everyday AI
- Posts
- Ep 736: ChatGPT Skills: How to Use The New Feature from OpenAI and The Best Use Cases
Ep 736: ChatGPT Skills: How to Use The New Feature from OpenAI and The Best Use Cases
OpenAI Debuts GPT-5.4 Mini and Nano, Google Expands Personal Intelligence, Anthropic releases Dispatch for iPhone control and more.
Everyday AI
March 18, 2026
š Subscribe Here | š£ Hire Us To Speak | š¤ Partner with Us | š¤ Grow with GenAI
Sup yāall š
Wondering if youāve enjoyed the Start Here Series so far and how long (or if) we should keep it going.
(And if you missed the boat, the Start Here Series is a series of podcasts that you can listen to in order to understand the fundamentals of AI, yet always keep up with trending categories and workflows.)
Thoughts on the Start Here Series?
Thoughts on Start Here Series?š« Vote to see LIVE results š« |
āļø
Jordan
In Partnership With Section
Section: The fastest way to drive, measure, and see returns on AI adoption
Most companies are spending thousands (or even millions) a year on AI tools that employees are barely using, and only 12% are actually getting business value from them.
Section is the platform that fixes that: it coaches employees on real use cases, tracks adoption across your org, and shows you exactly where AI is and isn't creating value.
You go from rolling out tools to proving measurable ROI. Stop guessing if your AI investment is working and check out Section at sectionai.com.
Outsmart The Future
Today in Everyday AI
8 minute read
š Daily Podcast Episode: ChatGPT just launched Skills, a new way to turn workflows into repeatable systems. Hereās how to use it. Give todayās show a watch/read/listen to find out.
šµļøāāļø Fresh Finds: Midjourney Launches V8 Alpha, MiniMax Releases M2.7, Dictionaries suing OpenAI and more. Read on for Fresh Finds.
š Byte Sized Daily AI News: OpenAI Debuts GPT-5.4 Mini and Nano, Google Expands Personal Intelligence, Anthropic releases Dispatch for iPhone control and more. Read on for Byte Sized News.
šŖ Leverage AI: The best teams arenāt prompting better. Theyāre building systems. ChatGPT Skills just made that a lot easier. Keep reading for that!
ā©ļø Donāt miss out: Miss our last newsletter? We covered: OpenAI Trims Side Projects, NVIDIA Launches Vera CPU, Microsoftās big Copilot shakeup, and more. Check it here!
Ep 736: ChatGPT Skills: How to Use The New Feature from OpenAI and The Best Use Cases
Imagine your work life without templates, SOPs or copy and paste. š±
Maybe that's how you've been working in ChatGPT, too.
Yes, ChatGPT's custom GPTs and Projects are useful, but left something to be desired.
The solution? ChatGPT's newest release: Skills.
Join us as we put AI to Work on Wednesdays and show you the ins and outs of Skills and the best use cases and our secrets to get started.
Also on the pod today:
ā¢ ChatGPT Skills unlocked for teams š
ā¢ Modular AI workflows explained š ļø
ā¢ Skills vs Custom GPTs showdown ā”
Itāll be worth your 38 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Hereās our favorite AI finds from across the web:
New AI Tool Spotlight ā Databox Turns business performance data into clear answers your team can understand, explain, and act on ā instantly, Lightfield is a CRM that remembers everything and does the work for you, Unsloth Studio is an open-source, no-code web UI for training, running and exporting open models in one unified local interface.
Midjourney v8 Alpha ā V8 alpha is live: much faster, sharper images, and new --hd and --q 4 modes. Try it and rate images to shape its future.
Google AI Studio Updates - In AI Studio, you can now mix built-in tools (Search, Maps) with your custom functions in one request, and let the model carry tool outputs across steps.
MiniMax launches M2.7 ā M2.7 self-upgrades with agent-driven RL, boosting code, research, and office work. Ready to plug into real workflows.
Copilot Demos ā Copilot shows why backstory changes a story, and how strength of schedule flips bracket picks. Try the demos.
Mistral Forge Released ā Train models on your companyās own knowledge so agents act like insider experts. Curious how Forge does it?
Mamba 3 Released ā Mamba-3 cuts inference memory in half while keeping model quality, so GPUs spend less time idle. Click to Learn More
OpenAI Compensation Study ā OpenAIās ChatGPT is giving people fast, accurate pay benchmarks for hard-to-search jobs. That clarity can shift career moves and boost negotiating power.
AI Copyright ā Two dictionary makers sue OpenAI for allegedly scraping 100,000 articles to train ChatGPT. Could this sink AIās free ride on publishers?
AI Photo App Banned ā An AI photo-editing app was banned after an ad claimed it could erase anything, including images that sexualize women. Click to Learn More
1. OpenAI launches faster, cheaper GPTā5.4 mini and nano š§
OpenAI today rolled out GPTā5.4 mini for ChatGPT, Codex, and its API and introduced GPTā5.4 nano for API use, bringing faster, more efficient small models that close the gap with larger GPTā5.4 on coding, reasoning, multimodal understanding, and tool use.
The mini is available now to Free and Go users via the Thinking menu and serves as a rateālimit fallback for other tiers, while the nano targets highāvolume, costāsensitive tasks like classification and data extraction. These releases signal a push to make advanced capabilities more affordable and operationally practical for production workloads without sacrificing much performance.
2. Google expands Personal Intelligence across Search, Gemini app and Chrome š£ļø
Personal Intelligence is rolling out now in the U.S., bringing connected, personalized responses across AI Mode in Search, the Gemini app, and Gemini in Chrome.
The feature links your personal Google apps to deliver tailored results for shopping, travel planning, troubleshooting, and more, while letting you control which apps are connected and when. Google says the system does not train directly on your Gmail or Photos and that training uses limited interaction data to improve functionality.
3. Anthropic rolls out Dispatch for remote Cowork control on iPhone š²
Anthropic today released Dispatch, a research-preview feature that lets Max subscribers remotely control a Mac-based Cowork session from a mobile device, with Pro access promised in a few days. The setup uses a QR pairing between the Claude Mac app and your iPhone, and while Dispatch can search files and work with Connectors, testing shows it is slow and only about half the time delivers expected results.
I one review, several actions like summarizing notes, finding screenshots, and adding items to Notion worked reliably, but more advanced tasks such as opening apps, fetching active Safari URLs, sending iMessages, and accessing some services failed.
4. Senators demand ByteDance pull Seedance 2.0 now š¤Æ
ByteDance temporarily halted the global rollout of Seedance 2.0 over the weekend, and U.S. Senators Marsha Blackburn and Peter Welch have now demanded the app be "immediately shut down," saying it threatens intellectual property rights and creators' livelihoods.
The senators argue Seedance was trained on copyrighted material and likenesses without permission, framing ByteDance's pledge to improve safeguards as insufficient and a stalling tactic. The move follows pushback from filmmakers and a cease-and-desist from the Motion Picture Association, and coincides with a proposed bill to give artists access to AI training records and stronger IP protections.
5. NVIDIA resumes China-specific AI chip production amid geopolitical tensions š„ļø
NVIDIA confirmed it is restarting manufacture of a China-tailored variant of its powerful AI chips, a timely move that addresses urgent demand from Chinese customers while navigating tightening export controls.
The company said the chips will have restricted performance compared with global models to comply with U.S. regulations, signaling a practical workaround that preserves market access without flouting rules.
6. Anthropic releases massive 81,000āperson interview on what people actually want from AI š
A new Anthropic report out today summarizes interviews with 80,508 Claude users across 159 countries about their hopes and fears for AI, finding most want tools that free time, boost work and learning, and offer emotional support while also worrying about job loss, unreliability, and loss of agency.
The research shows 67% hold a net positive view, with stronger optimism in lower and middle income regions and clear tensions where the same capabilities that deliver benefits also create harms. Anthropic used Claude classifiers to code themes and found productivity, personal transformation, and life management top desires, while hallucinations, economic disruption, and dependence lead concerns.
Your team doesnāt need better prompts.
It needs full-blown AI SOPs that actually stick the script.
(And sorry... if your people are STILL re-explaining the same task in every chat in the year 2026, a sharper competitor has prolly already turning that chaos into process.)
OpenAI has caught up with the pack with its recently released Skill function available to teams on a paid plan, which we unwrapped on todayās show.
Letās be real. The winners in the business world arenāt reinventing the wheel for each deliverable.
Theyāre using templates and SOPs to spread their domain expertise and control outcomes.
And thatās exactly the agentic adherence that we have now in ChatGPT with skills.
So on todayās āAI at work on Wednesdaysā show, we walked you through the ins and outs of skills, some of our favorite use-cases, and the secret way to instantly grab the most out of Skill instead of just chasing shiny solutions.
Ready to SOP your AI?
Letās get it.
1. Stop hoarding prompt hacks š
Most teams are still using ChatGPT like a smart coworker with short-term memory and zero process discipline. Helpful? Yep. Reliable at scale? Not even a little.
Thatās the real unlock here. Skills let you package repeatable workflows into something reusable. Not just a clever prompt. The actual bones of how your team gets work done. SOPs, templates, references, instructions, assets. The whole burrito.
That matters because prompt hacks do not scale across a company. They live in one personās brain, one saved chat, or one cursed Notion page nobody updates. Skills take that fragile little tower of duct tape and turn it into a reusable workflow your team can actually trigger on purpose.
So the business value is not ācool, new feature.ā Nah. Itās consistency. Itās output control. Itās giving your team a way to stop re-teaching AI the same job every single day.
Try This
Pick one deliverable your team repeats all the time.
Not the sexy one. The one that quietly eats hours every week. Grab the best example, the docs behind it, and the edits people always make after the first draft.
That is your first Skill. Start there.
2. Build from the mess š„
Hereās the cheat code from the episode. Donāt start by browsing for popular Skills and copying what everyone else is doing. That is how you end up with a fancy hammer and no clue what needs building.
Instead, start with your own work.
Look at the long chats where your team had to go back and forth a bunch of times, but eventually landed something GREAT. Those messy threads are not a sign the AI failed. They are a map of what your business actually values. Tone corrections. Missing context. Structural tweaks. All the little ānah, not like thatā moments.
That is pure gold.
Why? Because those corrections reveal the hidden SOP you were already using. You just had not formalized it yet. Skills give you a way to trap that logic in a repeatable workflow instead of letting it evaporate into chat history and vibes.
The team that figures this out first is gonna look way smarter than it actually is. Which, honestly, is a pretty good business model.
Try This
Audit five chats where the final output was worth keeping.
Then pull out every meaningful correction and rewrite it as a standing instruction. What should always happen? What should never happen? What context needs to be pulled in every time?
Congrats. You just found the raw material for a Skill.
3. Turn tribal knowledge into leverage ā”
The biggest opportunity with Skills is not personal productivity. Itās team infrastructure. This is where things go from ānice little AI trickā to āoh, this could actually change how we operate.ā
Think about the workflows your company repeats every week. Executive briefings. Budget versus actuals explainers. Marketing performance summaries. Contract review summaries. Operations reporting. None of that stuff is glamorous. All of it matters.
And thatās why Skills are such a big deal.
They let you take domain expertise that usually lives with your best operators and spread it across the team in a more controlled way. So instead of relying on memory, heroics, and copy-paste chaos, you get reusable process. That means fewer dropped steps, fewer weird outputs, and fewer moments where someone has to manually rescue the work at the eleventh hour.
This is how ChatGPT stops being a chatbot and starts acting more like an operating layer.
Try This
Choose three workflows where inconsistency is expensive.
Start with the ones that create the most rework, confusion, or executive side-eye when they come out sloppy. Build Skills for those first.
Boring wins here, fam. The companies that operationalize the unglamorous stuff first are usually the ones that end up pulling away.
Reply