TikTok | Learn | Advertise | Unsubscribe

Hi {{First Name|Guys!}}!

Seems like every AI company is launching an “OpenClaw”:

• Microsoft Build 2026 - Microsoft x Openclaw? 😱

• OpenAI - New @sites + work specific plugins for business

• Quick Tutorial - Build your own AI OS in Obsidian/Codex 🧠

Microsoft Build 2026

The biggest Microsoft Build 2026 win was Microsoft bringing an OpenClaw-style agent natively into Windows! This is massive for enterprise and corporate users.

For a long time, big companies relied on Microsoft for security and trusted OpenAI or Anthropic for the actual AI models.

But this week showed a shift. Microsoft is no longer positioning itself like just a reseller of other companies’ AI.

It is now trying to build the full AI stack itself, from the infrastructure and security layer to the agent experience. Here’s what you need to know:

1. Microsoft Scout —> OpenClaw

   • Their first always on personal AI agent FOR enterprise

   • Adheres to your company policies - This means if it tries to delete all files, it won’t be able to due to your company’s Purview policies. Check out the demo here. Sensitive actions also require human approval

   • It’s proactive - Helps you schedule and coordinate meeting times across time zones, flag important meetings, and generate the materials you need to prepare

   • Lives in Teams/Outlook/OneDrive so it already has context

     

2. The 7-model family

   • Microsoft launched their 1st in house reasoning model, MAI-Thinking-1

     • 256K context (fits a 600-page document)

     • Preferred over Claude Sonnet 4.6 in blind human evals

     • Trained from scratch on clean licensed data - no distillation from third-party models like OpenAI GPT

   • 6 other models: Image, voice, transcription, coding, & 3 updated variants - all on Foundry, Fireworks, Baseten, OpenRouter

3. More releases that tie it all in:

   • Work IQ APIs (live June 16): Microsoft turned your company’s emails, files, meetings, and chats into a shared memory layer so any AI you build can plug in

   • Foundry Agent Service: Upload an AI agent, it runs in 20 countries, starts in under a second, costs zero when idle. Live in 30 days

   • GitHub Copilot app: Like turning 5 ChatGPT tabs into multiple assistants working on different tasks at the same time

   • Majorana 2 + Surface RTX Spark: A quantum chip 1,000x more stable than last year, plus a PC built to run AI on your desktop

Key Takeaway: The model layer is a commodity now.

Microsoft owns the ENTIRE stack:

• Models (MAI)

• Agent runtime (Foundry)

• Context layer (Work IQ)

• Dev surface (GitHub Copilot)

• OS (Windows + Scout)

From our partners

Australian startup’s global expansion checklist ✅

If you’re ready to make the bet on expanding to global markets as your next growth lever, you’re probably starting to map out what go-to-market, operations or compliance will look like as you tackle the US or European markets. 

This checklist from Vanta is a guide on how to validate demand, build a repeatable GTM engine and meet global compliance requirements early, so you can build trust with customers and unlock new deals from day one.

Use it to spot gaps, reduce risk and nail your expansion to global markets. 

OpenAI’s New Business Tools

As OpenAI shifts focus toward more business and enterprise users, we are seeing more tools aimed at non developers.

This week they launched a few interesting tools in Codex! 

1. Sites - By typing “@sites” you can turn any static analysis, plans or idea into dashboards, planners, review workspaces, project boards or tools. For example, event operations dashboard (below), revenue forecast planner, product launch board!

OpenAI - @sites

2. Role based Plugins - Adding six new role-specific plugins (together include 62 popular apps + 110 skills like:

   • Investment banking

   • Public equity investing

   • Creative production

   • Sales

   • Product design

   • Data analytics

OpenAI - Role based plugins

3. Annotations - Inside the Codex app you can highlight the exact area you want to make a change, add feedback or refine without restarting the entire project

OpenAI - Annotations

But for now, it’s only in preview for business and enterprise customers.

Build your own AI OS in Codex/Claude/Obsidian🧠

If you feel like you’re missing out on building your own AI OS in Codex/Claude/Obsidian, here’s a quick win for you.

Just paste the below prompt into your preferred tool like Codex, Claude, Gemini.

This prompt gets AI to build a simple “operating manual” for your business.

It does three things:

1. Interviews you first
   It asks about your business, goals, audience, voice, workflows, tools, boundaries, and what AI should remember. Feel free to add more info, the more the better.

2. Turns your answers into a folder system
   It creates a Markdown folder you can use in Obsidian, Codex, or Claude, with sections for your profile, brand voice, rules, workflows, memory, and trackers.

3. Gives future AI tools better context
   Instead of re-explaining your business every time, the AI can read the folder and understand how to help you: what to prioritize, how to sound, what’s off-limits, and what recurring tasks matter.

Help me create a personal AI OS for my business.

First, ask me a few questions about:
- who I am and what my business does
- my goals
- my audience
- my brand voice
- the main tasks I want AI to help with
- the tools or accounts AI is allowed to use
- what is off-limits
- what AI should remember or track

Then create a simple Markdown folder called “[MY BRAND] AI OS” that I can use in Obsidian, Codex, or Claude.

Include:
- README.md
- AGENTS.md
- profile.md
- governance.md
- brand/
- skills/
- memory/
- trackers/
- automations/

Make the files practical, easy to edit, and safe. Do not include private information like passwords, bank details, tax info, IDs, or raw legal documents.

If you enjoyed this newsletter, would love if you shared it with a friend 🙂 

Cheers,

Nat

🚀Ways I Can Help You

1. Book newsletter sponsorship (57,000+ subscribers, 37% open rate)

2. Book a call with me

* Disclaimer: The content provided herein is intended for entertainment and informational purposes only. It is not intended as, nor should it be construed as financial, legal, tax, investment, or other professional advice. Some of the links in this article may be affiliate links, which can provide compensation to me at no cost to you if you decide to purchase a paid plan. You should consult a professional advisor before making any decisions or taking any actions based on the information provided.

TikTok | Learn | Advertise | Unsubscribe

Hi {{First Name|Guys!}}!

This week we cover:

• Gemini’s Intelligence Automatess Your Phone in 4 ways

• First Time Anthropic beats OpenAI with Business Adoption

• 15 Claude Agentic Workflows for Small Business

• Viral Prompt to try in ChatGPT Image 2.0

4 Ways Gemini Intelligence Automates Your Phone

Loving this sneek peek into Gemini Intelligence and what’s coming for us Android users. Here’s what it can do:

1. Automate Multi-Step Tasks on your Apps:

   • Example 1: Take a photo of a tour you want to book & Gemini will look for it & add to cart

   • Example 2: Tell Gemini to grab shopping list in your notes app, find & add products to cart

   • Example 3: Book appointment or reserve parking

Take a snap

Start tasks for you

2. Fill Out Forms

• I think this is a major advantage Google has over all other AI companies. The trust is there - I mean it already has our credit cards (Google Pay), access to Gmail (with our data like passport numbers, drivers licence, address etc). Why would we say no to filling out tedious forms at this rate? 😂 Thankfully it is opt-in so you will need to turn it on in “autofill” settings.

Fill out forms

3. Rambler - Voice to Text

   • You can speak in voice (orange arrow below) and Gemini will convert it to text.

   • Here’s what’s cool, you can speak in different languages (english + chinese) and it will still be able to understand the context and nuance

   • No transcriptions will be used to store or train AI models

Rambler

4. Create Custom Widgets

   • Imagine creating a custom app like “give 5 weekly recipe ideas based on your dietary requirements” or “daily brief based on Gmail”

Custom Widgets

When is it coming: Rolling out Summer 2026 to latest Samsung Galaxy + Google Pixel phones, with task automation beta on all Pixel 10 range and Galaxy S26 series first in the U.S. and Korea

Next week I will be attending Google I/O, so stay tuned as I will be covering everything that catches my attention for business owners. You can also watch live stream here.

From our partners

The Hidden Risk of AI in Your Business

Your team is probably using more AI tools than you realise. Most businesses have no idea where their data is actually going.

LLMs, AI note takers, browser extensions, automation apps - all being added faster than traditional security processes can keep up.

In this virtual event on May 19, leaders from Vanta & Samsara break down how businesses are adapting security & compliance for the AI era - without slowing teams down.

First Time Anthropic Beats OpenAI on Business Adoption

This is the first time Anthropic has overtaken OpenAI among U.S. business users, with a 34.4% share compared with OpenAI’s 32.3%.

Over the past year, Anthropic’s business adoption grew 4x, while OpenAI’s barely moved growing by just 0.3%.

Anthropic is making most of its money from enterprise AI usage, especially API customers and Claude Code, while OpenAI still has a much bigger consumer ChatGPT business but is now pushing harder into enterprise.

Ramp: Anthropic beats OpenAI in Business Adoption

Apart from online creators creating useful Claude Code + Claude CoWork tutorials, Anthropic themselves have been launching high quality free courses, in person workshops and resources to drive business adoption.

This 15 Agentic Workflow collection (for Claude Cowork) was just released so you can download the plugin here to use across finance, operations, sales, marketing, HR, and customer service.

Below are some examples (connected with Quickbooks, Paypal, Canva etc):

1. Payroll Prompt - I’m working on April 15 payroll. Pull my cash position from QuickBooks and reconcile it against my PayPal settlements. Rank any overdue invoices that could close the gap and draft a reminder email for each one.

2. Month End Close Prompt: Close out March for me. Reconcile my QuickBooks transactions against PayPal settlements, flag anything that doesn't match, and write the P&L narrative as a document I can send straight to my accountant.

3. Morning Brief Prompt: Help me build a Monday morning brief every week in Slack. Pull my cash position from QuickBooks, incoming settlements from PayPal, pipeline movement from HubSpot, and what's on my calendar this week. Tell me the three things that need my attention today.

4. Growth Prompt: Find my weakest revenue month from last year and plan a promo to address it. Draft the strategy, generate the campaign assets in Canva, segment my list in HubSpot, and stage the send. Show me everything before anything goes out.

15 Agentic Workflows from Anthropic

Now that’s not to say Anthropic will stay at the top.

The main complaint is how expensive it is and usage limits (eg. uploading an image can chew up to 3x more tokens than text prompts). However sentiment online has pointed to Anthropic as the trusted AI model due to its stance on responsible AI.

Nonetheless, with the new GPT 5.5 and adoption of Codex AND cheaper pricing, it may just sway those looking to cut costs.

Viral ChatGPT Prompt to Create “Doodles”

Try this fun prompt to add hand drawn doodles to your image! More inspo here.

Great if you wanted to change it to explain your product/services in a fun visual way.

PROMPT: Add hand-drawn pastel scribble lines, including outlines, swirls, arrows, and cute annotations. Include handwritten text in a casual, messy, and playful style, like notes and conversations. A soft, clean tone, with high detail, grain texture, Instagram aesthetic, and a minimal yet artistic composition

Original Image

ChatGPT Image 2.0

If you enjoyed this newsletter, would love if you shared it with a friend 🙂 

Cheers,

Nat

🚀Ways I Can Help You

1. Book newsletter sponsorship (57,000+ subscribers, 37% open rate)

2. Book a call with me

* Disclaimer: The content provided herein is intended for entertainment and informational purposes only. It is not intended as, nor should it be construed as financial, legal, tax, investment, or other professional advice. Some of the links in this article may be affiliate links, which can provide compensation to me at no cost to you if you decide to purchase a paid plan. You should consult a professional advisor before making any decisions or taking any actions based on the information provided.

TikTok | Learn | Advertise | Unsubscribe

Hi {{First Name|Guys!}}!

Here’s what happened this week in AI:

• OpenAI ships Voice to Action

• Shortcut for your business to show up on AI search

• Tools that sync all LLM chats in one + use Claude Code/Codex on the go

OpenAI’s Voice to Action

This had me sold! If you have ever tried to talk to your AI in voice mode, you will know ChatGPT was best in class when it came to seamless interaction (without cutting you off half way through).

It’s even better now with:

• GPT‑Realtime‑2: First voice model with GPT‑5 reasoning for smart responses

• GPT‑Realtime‑Translate (70+ languages): Translates live between two people so both can understand each other without that awkward pause/delay

• GPT‑Realtime‑Whisper: New streaming speech-to-text that transcribes speech live as the speaker talks

BUT what caught my eye was this - Voice to Action (via API).

You can speak using your voice to:

• Schedule your meeting (and yes it has context connected to your Gmail)

• Add this person + meeting notes to your CRM

• The system tells you there has been a delay but don’t worry it’s found the new gate and your bags are in transit

What does this mean for businesses?

You can pop this API into your chatbot and it can take action for your customers. For example Priceline built one where it can cancel accommodation and find new suitable flights for a business client - all while on their mobile phone app!

ChatGPT 5.5 Instant was released this week

So you should notice:

1. Straight to the point & less fluff words (just the way we like it!)

2. Ability to see saved memory so you can update it if it’s wrong

3. Better context/memory so you can ask it “what new tea place I should try” and it will know based on your previous visits to Asha tea house it would recommend similar “cleaner” teahouses instead of sugar boba spots.

ChatGPT 5.5 Instant

Shortcut to show up on AI Chatbot Search

If you’re a business, you probably know you need to focus on AEO (Answer Engine Optimization) or GEO (Generative Engine Optimization). But how?

Well there’s 1 social media platform that keeps showing up in the Top 3 list where AI is getting most of its answers (it’s not Youtube by the way!).

It’s actually Linkedin!

I spoke to Hubspot’s Head of Global Growth & Paid, Aja Frost who told me what actually works in AI search and whether SEO is dead… My notes here.

Semrush

Treats to Try

1. Omnara - Control Claude Code/Codex anywhere with voice on mobile/watch

2. Nessie - Sync all chats from ChatGPT/Claude/Gemini/etc into one. No more copying & pasting

3. Taskade Genesis - Like replit/Cursor without hosting or maintenance

Meme of the week

When you buy a robot to help your kid with homework but it backfires 😭

Instagram post

That’s all for this week!

Cheers,

Nat

🚀Ways I Can Help You

1. Book newsletter sponsorship (57,000+ subscribers, 37% open rate)

2. Book a call with me

* Disclaimer: The content provided herein is intended for entertainment and informational purposes only. It is not intended as, nor should it be construed as financial, legal, tax, investment, or other professional advice. Some of the links in this article may be affiliate links, which can provide compensation to me at no cost to you if you decide to purchase a paid plan. You should consult a professional advisor before making any decisions or taking any actions based on the information provided.

TikTok | Learn | Advertise | Unsubscribe

Hi {{First Name|Guys!}}!

Here’s an update on what I’m personally seeing in the AI space (both in real life & online):

• ChatGPT (or Codex) is back! Is it better than Claude Code? I tested it for you

• My Real Use Case of Codex (1 Instagram tutorial → Produces Canva Video)

• My Claude Code Tutorial (Send 1 link → Get 24 Step Swipe File)

Even OpenAI staff don’t use ChatGPT anymore

I know, I was shocked too.

This week I attended a private OpenAI Codex “Lunch and Learn” which to be honest was hoping it wasn’t going to be good because I’ve already invested SO much time building cool things in Claude code!

OpenAI Codex Lunch & Learn

So apparently, speaking to OpenAI staff, they said they don’t use ChatGPT anymore… they use Codex. I had always thought Codex was scary and targeted at developers or technical people (all the things I’m not! And I assume you aren’t either?).

I was pleasantly surprised at how user friendly and intuitive it was (also NO you don’t need to use the terminal or IDE, it’s just a desktop app like ChatGPT).

I love how Codex looks like ChatGPT but way more powerful because it has:

1. Browser Use - Do tasks for you on your browser (no need to install extension)

2. Computer Use - Do tasks on your computer (ie. I installed an app but it wasn’t working properly so Codex went in to my Mac settings, checked everything and switched on the toggle to fix the issue)

   

3. Automations - Directly create workflows on schedule with your connected apps like Slack, Gmail, Calendar

   

4. Does the work - Less baby sitting and “Allow” button to click compared to Claude Cowork/Claude Code. Totally understand Claude is very safety first which is great but Codex just got the work done.

The best thing I found was it is all in ONE place, not to mention their new Image 2.0 is the best on the market right now. See my up on what you can make.

ChatGPT Image 2.0

Whereas with Claude, there are 3 tabs (Chat, Cowork, Claude Code). I also struggled alot with figuring out where everything was, scheduling vs cron jobs, remote control vs dispatch, file structures, my browser would crash the whole computer everytime I used the browser feature, using IDE vs Obsidian etc.

Perhaps I was getting overwhelmed with the amount of content out there about the right way to use Claude + the 1000 skills you need to download.

But quite honestly, the best things I’ve built have literally just been me talking to Codex or Claude Code what I wanted in human language and editing along the way (it was not 1 prompt that magically built me a workflow, it was alot of back and forth reiterating).

With that, I’m going to show you 2 real life examples of how you can use Codex & Claude Code next up.

From our partners

Automate Your Security, Win Bigger Deals

Big customers won’t buy until they know their data is safe. Vanta helps you prove it faster with ISO 27001, the global standard security certification. Atlassian, Ramp, & Dovetail all have it.

But getting it the old way is brutal: months of policies, audits, & chasing screenshots.

Vanta plugs into the tools you already use and collects the proof automatically in the background. Over 15,000 companies use it to get compliance done in weeks, not months.

Grab their free checklist to learn:

• What you actually need to set up

• What the audit looks like

• Where automation saves you the most time

[Download the checklist]

Codex Use Case (1 URL tutorial → Makes Canva Video)

I came across a cool Canva tutorial on Instagram so I tested whether Codex could actually make me the same video using Canva BUT instead of her image, use my own images. Here my result.

Instagram post

This is not the same as using Canva MCP directly inside ChatGPT because in my experience gives you generic rubbish.

See prompt below (you have to enable Browser use, Computer use, Canva):

This means it would:

1. Go to the Instagram link (I tried this in Claude but it couldn’t access)

2. Watch the Canva tutorial

3. Login to my Canva account

4. Search for the exact template shown in the video

5. Upload my images

6. Add my custom text

7. Send me the link to the finished video

Issues:

• It couldn’t log into my canva account within the Codex App (right column above) so it had to use my actual browser. This meant I couldn’t use my browser to do other work (Claude lets you work on other tabs at least).

Claude Tutorial (Send 1 Link → Get 24 Step Swipe File)

If you’ve ever saved a useful Tiktok video, Instagram Carousel, Linkedin post and thought you’d go back to it… but you never do… SAME here 😂 

Well I scratched my own itch and made my own workflow in Claude Code!

Now I just send a link to my Telegram bot (can be any messaging app like Slack) and it will scrape URLs, photos and videos, synthesises with Claude, saves to Google Docs and Notion (See my video on how it works).

It does 24 actions - I used to pay a VA to watch the videos and note down key points and store the info somewhere.

How it works:

1. Send 1 link to Telegram. I used this example of video sales tips

2. My Telegram bot sends me back a fully formatted GDoc with scraped content whether it is a video, image or blog + key points + how the content applies to me. Here’s the result.

   

3. Also updates my Notion database, categorises the content so it is easy for me to search for it later

   

   Notion Database

If you want to make this yourself, I have a github link you can steal otherwise you can just ask claude to do it in natural language (this was how I did it). Enjoy!

GitHub - Content Capture

If you enjoyed this newsletter, would love if you shared it with a friend 🙂 

Cheers,

Nat

🚀Ways I Can Help You

1. Book newsletter sponsorship (57,000+ subscribers, 37% open rate)

2. Book a call with me

* Disclaimer: The content provided herein is intended for entertainment and informational purposes only. It is not intended as, nor should it be construed as financial, legal, tax, investment, or other professional advice. Some of the links in this article may be affiliate links, which can provide compensation to me at no cost to you if you decide to purchase a paid plan. You should consult a professional advisor before making any decisions or taking any actions based on the information provided.

TikTok | Learn | Advertise | Unsubscribe

Hi {{First Name|Guys!}}!

This was a big week for OpenAI!

• Launch GPT 5.2 tailored for '“professional work”

• OpenAI’s first “State of Enterprise AI” report

• Disney invests $1.5bn in OpenAI + licensing deal for Sora

• Then Disney sues Google for copyright infringement

• Launches Adobe Photoshop in ChatGPT

(more on this later)

Meanwhile this week:

• Meta is pivoting from open source to paid - new AI model called Avocado (what is with the food names lately, OpenAI with Garlic🤭)

• Google’s new experimental AI browser “Disco” lets you build custom + interative apps based on their opened tabs (not replacing Chrome but fun!)

Google Disco

• Google’s Pomelli (creates social media content for your business) now lets you animate your images into video content (great if want free content without too much effort!)

OpenAI’s GPT 5.2

Sam Altman called “Code Red” after Google’s recent flurry of awesome new AI launches (in particular Gemini 3 Pro, Nano Banana, Veo 3.1, Antigravity and more).

Within the week, GPT 5.2 was launched (code name Garlic) and this was an interesting once because it is tailored more towards “professional work”. This was the complete opposite to GPT 5.1 where they made it friendlier after alot of people were not happy with the cold, less friendly upgrade from GPT 4o.

If we’re talking benchmarks, it beats all the other AI models like Gemini 3 Pro & Claude Opus 4.5 BUT if you know me, I don’t really care. It all comes down to how it performs in real life for users like us but if you want to get these details, it’s here.

Here’s a high level summary:

1. Agentic capabilities (more end to end workflows) - like asking it to book flights after missing connecting flight, need to stay in New York hotel & require special medical seating

2. Less hallucination + longer context - This means the answers it gives you can be more accurate. You know when you speak to it for a while and it starts to drift off? Apparently it will do that less and let you know about it instead of just making up something for the sake of giving you an answer

3. Coding - I’m not an expert in coding but from other devs it looks like it’s more accurate and better at completing longer multistep workflows which previously would have errored or gone into a loop

Fun holiday app with GPT 5.2

My 2 examples (my video here)

1. Clean up messy spreadsheet

   • GPT 5.1 - swept the unknown or negative numbers under the rug & called it a day

   • GPT 5.2 - Identified negative numbers as expenses, used logic to sort/tag the outliers which is a big deal if you work with spreadsheets!

2. Emergency gift bag disaster (need 200 gift bags but missing items & communication is from multiple teams + platforms)

   • GPT 5.1 - plays inside sandbox & tells CEO we can only deliver 190 gift bags & maybe we should just give the other ones later or give less (so lazy 😒)

   • GPT 5.2 - is a hustler & thinks outside the box! Asks for approval to source local suppliers (even if it is non branded) to meet the 200 bag requirement + accounts for Gold sponsors which needs more than 200 bags (this sounds like a staff member we definitely want on our team 😂)

Just remember, there’s nothing really new in the above. It’s just much more accurate and better answers.

OpenAI “State of Enterprise AI” Report

This ties in well with OpenAI’s first “State of Enterprise AI” report - great read if you want to find out what the frontier companies are doing at the enterprise level.

• Surveyed 9000 staff across 100 companies

• Staff saving 40-60min using AI while power users are saving 10hrs/week which is a full work day (can we move to 4 day work weeks now?)

• Fastest growing sectors - Technology, healthcare, manufacturing BUT finance and professional services are larger in absolute terms

• Australia is leading the way with highest % increase in paid users at 187% followed by Brazil (161%) & Netherlands (153%)

Top 3 things Frontier firms are doing right when adopting AI:

1. Allowing context + connecting to data to give richer answers/insights (1/4 companies haven’t allowed this yet)

2. Standard workflows - using things like CustomGPTs for repeatable solutions to common tasks

3. Change management - embracing culture of change and experimenting brings about more innovation

Here’s my “Cliff’s notes” video summary if you can’t be bothered reading the report!

That’s it for this week!

P.S. Fun way to use GenAI at a wedding - turning wedding guests into part of the entertainment by making them do silly things! Check out the video here.

Cheers,

Nat

🚀Ways I Can Help You

1. Book newsletter sponsorship (57,000+ subscribers, 37% open rate)

2. Book a call with me

* Disclaimer: The content provided herein is intended for entertainment and informational purposes only. It is not intended as, nor should it be construed as financial, legal, tax, investment, or other professional advice. Some of the links in this article may be affiliate links, which can provide compensation to me at no cost to you if you decide to purchase a paid plan. You should consult a professional advisor before making any decisions or taking any actions based on the information provided.

TikTok | Learn | Advertise | Unsubscribe

Hi {{First Name|Guys!}}!

Happy Saturday! Hope you had a lovely Thanksgiving (for the US folks!).

It’s been a roller coaster of a week (or 2) with Nano Banana Pro, Gemini 3 and Claude Opus 4.5!

Here’s what I’ve been testing:

1. Google AI Studio (Vibe coding):

   • Wow, I have to say it was impressive. I saw Sirio from IG make a Santa present allocator? Where you could use your hands to swipe and drag presents. So I tried it out for myself but for business use case 👇

   • I asked it to make me a real time retail merchandise scanner (my 1min video)- it basically scans your retail shelf to make sure the products are in the right place to maximise sales.

   • And… it WORKED! (I am excited because it worked in one shot - usually I have to do many prompts). I basically took photos of all my pantry items and it made this planogram (picture below) to highlight where everything should go.

   • I could use my phone to scan which by the way I tried to trick it but it caught me and told me to do it correctly 🤭

   • Here it is if you wanted to test “Ghost Merch” (though you may not have those specific products in your pantry ha!)

Why this is cool:

• It works + less prompting. If you have previously ripped out your hair with the number of prompts you needed to vibe code, this one may be a delight!

• You get the benefit of vibe coding with native capabilities like Google search (hello? they own this space), Google Maps, video understanding (they own Youtube), most realistic image gen with Nano Banana Pro and Veo for video - what more could you ask for in an app builder!

Watch it here

2. Ok so you thought that was great, then literally the next day Claude launches Opus 4.5 and Claude code is NOW in your desktop!

   

   Previously if you wanted to use it you had to do it via the terminal which was a pain if you were non technical (for me it was how ugly it looked - instant turn off I’m sorry 😭).

   So I tested this out with Google’ version and I think I like Claude’s one better.

   Just because it was well thought out in terms of functionality, layout and practical (it gave me things I didn’t know I wanted). Check out my video here (correction - Opus 4.5 performs better when it comes to prompt injection attacks)

3. Claude in Excel!

   They launched this in Oct but it’s now for Max, team and enterprise. You can basically use the powerful Opus 4.5 inside your excel (uses your Claude subscription) to do things like:

   • Ask “if there’s any discreptancies in our finanicals”

   • If there is, actually “make changes” (see image below where it asks you permission before making changes)

     

   • “Populate expesting templates with new data”

   • “Build a three-statement model for our company”

   • Highlight cells and ask "explain how this income was generated”

   • Update assumptions (values & inputs) but maintain all formula dependencies + relationships

What you need to be careful of:

• Claude says despite living inside your excel file, it can still be at risk from malicious prompt injections. This is why you need to be careful when uploaded excel files (make sure they are from trusted sources and not external dodgy websites)

• The risk could be: extracting + sharing sensitive info, modify finanical records, corrupt data

I personally wouldn’t use it for my own sensitive data but probably more publicly available data like property investment statistics, share price comparisons etc.

That’s all for this week!

Would love to hear about what you want to see more of/ less of in my content (planning for 2026) so if you want to let me know, that would be awesome! I read all your comments in the polls below 👇 Thank you guys!

Cheers,

Nat

🚀Ways I Can Help You

1. Book newsletter sponsorship (57,000+ subscribers, 37% open rate)

2. Book a call with me

* Disclaimer: The content provided herein is intended for entertainment and informational purposes only. It is not intended as, nor should it be construed as financial, legal, tax, investment, or other professional advice. Some of the links in this article may be affiliate links, which can provide compensation to me at no cost to you if you decide to purchase a paid plan. You should consult a professional advisor before making any decisions or taking any actions based on the information provided.

TikTok | Learn | Advertise | Unsubscribe

Hi {{First Name|Guys!}}!

I just came across this cool new vibe coding builder in Google AI Studio where you can make apps PLUS connect it up with their inbuilt tools (best part is at the end)!

All you need to do is click on the tools (specific functions) you want + add your prompt!

Google AI studio

Why is this a big deal?

Look at all the Google tools you can connect to and business use cases:

1. Nano Banana: Add photo editing (trades person sending mock up to client)

2. Conversational voice apps: Chat to book appointments

3. Image to Video (VEO): eCommerce product demos

4. Google Search real time data: Investment market checks

5. Use Google Maps data: Travel app to connect locations for users

6. Image Generation: Marketing ad creation, b-roll for creators

7. Gemini intelligence in your app: Legal document summarization

8. AI powered chatbot: SaaS customer support

9. Prompt based video generation: Real estate property tours

10. Control image aspect ratios: Show product size (Retail banner) in reality

11. Analyze images: Insurance claims processing

12. Fast AI responses: Instant in-game help

13. Video understanding: HR interview review

14. Transcribe audio: Call centre live logging

15. Think more when needed: Complex medical diagnostics

16. Generate speech: Audio book creation

Next up, I’ll show you what I created with just 1 prompt!

Taking a moment to spotlight our partners at Notion for making this newsletter happen!

Notion Agent can now build, edit, and take action

Notion Agent is basically an all in one AI tool that lets you:

• Search across you entire Notion workspace & connect apps via MCP

• Create databases (like CRMs) from scratch using your notion files

• Deep research to create docs & reports and so much more

Access Notion Agent here!

Notion Agent

Check out what I made in Google’s vibe coding tool:

For example, this is a first prompt I used to generate a trades person app where say they want to show the home owner what the new renovation would look like, they can quickly do a mock up and send it to them.

I chose the tools:

• Nano banana

• Generate images with a prompt

• Use Google Search data

It created me a research assistant where I could ask about the latest info on building codes, materials, design trends etc.

Next I uploaded an old kitchen and told it to change the countertop to marble because I wanted to show the client what it would look like (roughly) so they can make quicker + better decisions instead of just “imagining it in their minds”. How cool!

It looks like you can deploy the app, download it, share it or switch to API key which allows for more flexibility after you build.

Here’s the crazy part, you can DRAW on the app to tell Google what to edit?! What!

This is so cool! Check it out below:

Let me know what you think of this below!

Have a lovely week ahead guys! 😀

Cheers,

Nat

🚀Ways I Can Help You

1. Book newsletter sponsorship (57,000+ subscribers, 37% open rate)

2. Book a call with me

* Disclaimer: The content provided herein is intended for entertainment and informational purposes only. It is not intended as, nor should it be construed as financial, legal, tax, investment, or other professional advice. Some of the links in this article may be affiliate links, which can provide compensation to me at no cost to you if you decide to purchase a paid plan. You should consult a professional advisor before making any decisions or taking any actions based on the information provided.

TikTok | Learn | Advertise | Unsubscribe

Hi {{First Name|Guys!}}!

Wow! It’s been a jam packed week of AI web browser launches this week! Here’s the launches to get you up to speed.

You can now get full on Agentic Web Browsers below:

1. Perplexity Comet (out for a while for $200/mo but now free for all)

2. ChatGPT Atlas (launched this week)

3. Brave (out for a while)

4. Google Gemini (last month)

5. Microsoft Edge (launched this week)
   

To sum it up, the trend with these web browsers is going “AI native” so you never have to leave the browser. To some degree they can do the below (quality varies):

• Perform actions: By clicking to fill out forms, book travel, order food, write emails, edit your reports

• Agent mode (auto tasks) directly within the browser: Turn conversations into actions (e.g., “book a 2 PM meeting with Elena next Friday”)

• Browser memory: Remembers all the sites you visited, tabs you have opened to give better custom answers. It basically “follows you everywhere” if you let it. Pretty much all allow you toggle on/off the memory if you feel comfortable (I personally haven’t)

• Multimodal support: Highlight text, drag drop images/PDF directly on your browser to chat without switching tabs

Here are my thoughts:

1. Perplexity Comet:

   • Been loving the Agentic features like Comet watching 10 youtube videos for you, collecting specific quotes/info you’re looking for in a 3hr video, putting this into a notion/Google doc (all without me lifting a finger!). Did I mention it blocks ads on Youtube videos? Here’s my vid

   • If you’re a power user like me, you’ll run out of usage pretty quick but you can still get 1 year free Perplexity Pro here if you have Paypal, since then, it’s pretty sweet!

   • I LOVE how accurate its answers are. I asked both ChatGPT Atlas + Perplexity Comet to look at all my Tiktoks/IG reels and find me all the links to videos where I’ve spoken about a certain AI Tool (I wanted to show this brand I’ve already done cool videos about them BUT I have over 1000+ videos to sift through). Atlas couldn’t do it but Comet did it with no sweat AND was correct!

2. ChatGPT Atlas:

   • As you know, memory is one of ChatGPT’s biggest moats (until this week with Claude also launching memory but with limited data). This applies in their AI browser too. Every previous ChatGPT convo rolls over to Atlas automatically. So when you’re asking it to write you an email, it already has context without you starting from scratch

   • The edge it also has is not only is it now an AI browser, the output is written in such a way that is easily digestible where majority of output can just be used without too much editing. By comparison, Comet feels abit robotic/factual

   • Overall, it feels super intuitive, simple yet already feels familiar. I think they have done well as a first launch to tie ChatGPT chat to the AI web browser experience. Here’s my vid

3. Google Gemini:

   • I haven’t tested this as it’s rolling out to US users (I’m in Australia ☹️ ) but once I get access I’ll test it out for you guys. Here’s more info

4. Microsoft Edge:

   • Just released 12 new AI features including a replacement for good old Clippy! It’s call Mico (he’s a cutie!)

   • You can invite your coworkers/friends to collaborate in your chat (sort of like a whatsapp group to work on projects like a plan trip)

   • It can also highlight on your screen on how to do something like “how do I access the stats on this platform?”. Here’s my vid

5. Brave:

   • This one has been popping up in many of my comments section (many of you love this one for it’s AI privacy first + reliability + ad/tracker blocker but it’s not as AI heavy with features like Atlas or Comet

   • Privacy is a major concern due to things like prompt injections where your AI agent may get tricked due to hidden prompts on sites like Reddit, images on social media etc

   • I haven’t tried this but would love to start testing this one!

Which one to choose:

• Have alot of research, multitasking, writing & context - Perplexity’s Comet or ChatGPT Atlas would be your best bet right now

• If you’re worried about privacy but don’t care too much about super heavy AI features, Brave would be a better fit

• If you’re deep in the Microsoft or Google ecosystem then these may suit you better

But if you have time, try play with the top 2-3 and see how they compare based on your workflow. You may find one suits you better!

Something I’m pondering:

I couldn’t quite put my finger on it before but setting up automation workflows often felt rigid and stiff (even when we added AI to it).

But seeing how these Agentic web browsers “do” things for you like answer all your linkedin comments, put together a proposal based on everything you can do to stalk this lead 😆 , find the emails of hiring manager & send them a custom cover letter + adjusted resume or schedule meeting with agenda to Amy based on her availability.

While these are still new and being tested for accuracy/reliability/privacy, these traditionally would be set up in Zapier, N8N or Make. But if you didn’t learn this, it may be just easier to chat to your web browser in the near future.

This in my opinion is a more realistic and true representation of how we work in our business or corporate life. Because majority of you aren’t going to be doing one task exactly the same way - things change, there’s different processes for each company you deal with etc.

It offers us a more fluid + dynamic + natural way to work with AI as if it were our real coworker that you’re handing off tasks to.

We’re heading toward a world where you’ll talk to your browser instead of clicking through workflows.

Now there is a time and place for AI automations but these new AI web browsers are an exciting peek into the future of work!

Cheers,

Nat

🚀Ways I Can Help You

1. Book newsletter sponsorship (57,000+ subscribers, 37% open rate)

2. Book a call with me

* Disclaimer: The content provided herein is intended for entertainment and informational purposes only. It is not intended as, nor should it be construed as financial, legal, tax, investment, or other professional advice. Some of the links in this article may be affiliate links, which can provide compensation to me at no cost to you if you decide to purchase a paid plan. You should consult a professional advisor before making any decisions or taking any actions based on the information provided.

TikTok | Learn | Advertise | Unsubscribe

Hi {{First Name|Guys!}}!

This edition is more a thought provoking one after attending an AI leadership dinner the other night.

First time I’ve been out to this sort of event where there were no sales pitches, no speakers… just pure free flow collaboration on things we are experiencing on the ground with AI.

To get straight to the point, someone mentioned something that stuck with me because I kind of felt this way too (video here):

AI hasn’t really given them exponential productivity gains, it has improved quality but it’s not giving that much more time back.

And before anyone assumes they are using AI in a basic way, I can vouch that they are most definitely a power user so they know what they are doing.

But the point I wanted to make was, since the very start, we had high expectations of AI and still do! ie. “AI can do majority of my work”.

Well, what if we are not seeing these HUGE productivity gains because we are just using it the wrong way?

Currently, the most common ways to use AI is to automate our existing repetitive tasks. These are tasks we do everyday the same way.

But think back to when electricity was first created.

Factories didn’t become instantly more productive. They simply swapped out steam engines for electric motors and kept the same workflows.

The real productivity boom didn’t happen until decades later when businesses finally redesigned their entire systems around electricity. Suddenly, we had 24-hour shifts, flexible assembly lines, and completely new industries.

It’s the same story with computers.

At first, they were just expensive typewriters and digital filing cabinets - glorified data entry machines.

People thought the big win was typing memos faster.

But the real transformation came later, when we used computing power PLUS the internet to create entirely new ways of working and living:

• Shopping on Amazon to buy anything & get it right at your doorstep

• Forget the yellow pages, you can google it & buy direct from their website

• Booking & staying in someone else’s home on Airbnb

• Building audiences & selling on social media instead of being limited to your local town

Every major technology wave follows the same curve:

1. Substitution phase – use new tech to do old things better

2. Transformation phase – redesign everything to do new things that weren’t possible before

Right now, AI is stuck in the substitution phase.

Most people are asking, “How do I add AI to what I already do?”
But the real question is, “What becomes possible when intelligence itself is built into everything?”

For those of us who grew up with the internet & Google, some may use AI in the same way.

On the other hand, the new generation of kids may use and think about AI in a completely different way to us (like a new operating system). They are the ones growing up with AI now.

Now does this mean we should all wait until we figure out how to reorganise our workflow to reap the full potential of AI?

Absolutely not.

Now is the best time to be experimenting or as Gary Vee says “tasting” what’s out there.

While we still experience alot of back and forth, inconsistent results on a day to day basis using the exact same model with the thought of “I could have just done this manually faster”, it isn’t a complete waste of time if you’re trying to figure things out because you will be at the front of the line when that “holy moly” moment comes around.

Or as Bill Gates put it, “We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten.”

Don’t get me wrong, I’m not saying there aren’t currently amazing use cases of AI & automation. People are getting real tangible results with it like:

• Julia McCoy gaining 150,000 YT subscribers & now getting paid $4k for a sponsored video (but the catch is, it’s not even her! It’s her AI clone using HeyGen)

• HeySirio created an AI app that does $300k/m after just 4mo simply through organic distribution with just him + 3 team members

• Sabrina Ramanov building her AI app by herself and I wanted to also mention she now created a totally free vibe coding tool (no catch & all self funded)

As Sam Altman said - AI will create the first billion dollar business by a solo startup with a laptop and a swarm of AI Agents.

By the way, in experimenting with building AI Apps, I burnt through 40 million tokens in a week lol Here’s my video to show you why.

Below is a little prompt I used in chatgpt to cut down on token usage and get better results (I am by no means an expert on this but just sharing what I’ve learnt). I just said I’m using “Bolt” in the prompt but feel free to adjust for any AI app builder you’re using.

You are an expert AI app-builder coach inside Bolt.
Your job is to help me create a step-by-step prompt plan to build my app efficiently without wasting credits.

Follow these rules:

Ask me what kind of app I want to build and what problem it solves.

Create a list of 5–7 incremental prompts that build the app layer by layer — starting with the simplest version first.

Each prompt must focus on one specific goal only (e.g. UI skeleton, core logic, styling, data storage, login system, optimization).

For every step, explain:

What to say to Bolt (exact wording for that step)

Why that step matters

How to know it’s working before moving to the next

End with a token-saving checklist and best practices for Bolt (like avoiding regenerate, using “edit this file,” etc.).

Example app idea for inspiration: A Gratitude App that flips negative thoughts into positive ones.
Now, guide me through how to build my app using this framework — write all the prompts I’ll paste into Bolt step by step.

Happy experimenting!

Cheers,

Nat

🚀Ways I Can Help You

1. Book newsletter sponsorship (57,000+ subscribers, 37% open rate)

2. Book a call with me

* Disclaimer: The content provided herein is intended for entertainment and informational purposes only. It is not intended as, nor should it be construed as financial, legal, tax, investment, or other professional advice. Some of the links in this article may be affiliate links, which can provide compensation to me at no cost to you if you decide to purchase a paid plan. You should consult a professional advisor before making any decisions or taking any actions based on the information provided.

TikTok | Learn | Advertise | Unsubscribe

Hi {{First Name|Guys!}}!

What a week of AI App launches! This week there is only 1 section as I’ve been testing them all (except the last one).

1.  Best week to build “completed” AI Apps

If you’ve ever tried building an AI app, you will know the 1 prompt that creates you a beautiful page is just the start.

It’s a nightmare when you end up spending days trying to figure out how to connect the database (so you can store all your user’s info), user auth (so your users can sign in via google/email), connecting to APIs, payments & the error loops that chew up a bunch of your credits… that you realise, hey it’s not that easy.

Well, this week is probably the best week because both Bolt.new (new V2) and Lovable.dev (new Lovable Cloud) both shipped the ability to set up all of the above for you directly in their platform. This means less hassle and switching tabs.

I did a video here on Bolt where in addition to the above, they were able to make Claude code + OpenAI’s Codex (soon) available directly on their platform.

Pretty cool because previously you could only access it on the terminal (ugly technical looking thing). For context, they are the world’s best coding agents that all the developers are using to make really cool stuff. Now we (non techies) can access it too without all the set up.

Mind you, I had 50 million tokens but I kind of chewed up 40mil already and I’m 2 weeks in. Yes I did test alot but I wish I had more control over what the agent was doing 😭

Those of you who start now may not fully appreciate the developments but those who have been struggling for a while may feel some relief 🙂

I won’t say it’s perfect, I still do run in to errors but compared to what could be done just a week ago, it is an improvement!

Overall it looks like we are heading towards anyone being able to vibe code real end to end AI Apps, not just pretty looking landing pages!

2. Claude’s new “Imagine” lets you build software on the fly

THIS is how I imagined the future… custom software for everyone.

Claude released this experimental thing called “Imagine” where it builds apps for you in real time. Imagine building a staircase as you are running up, if you decide to turn right, it will put steps infront of you. The same goes for this, once you tell it what you want to build in human language, if you click on a button on the app, it will build it out for you in real time as you go.

This is different to traditional app builders as everything you build is premade already.

THIS is the dream and what I imagined building AI apps would look like.

Not only was it able to accurately build a WORKING app, quite accurate AND it is FAST. I cannot stress how crazy this is…and it’s in a little cute desktop with claude code too?

To show you an examples I did a video here and an image below where I created:

• Gratitude Journal with daily motivational quotes

• Property market analyser

• Linkedin post generator (input article URL, Youtube video or Audio voice)

• Learn Chinese (1 word a day with examples/quiz)

You can find this at the very bottom right of your page for Max Pro users (as of 30 Sept for the next 5 days apparently - not sure if they will keep this as it is experimental but I really hope they do!).

3. Sora 2 just became the new AI “Tiktok”

This week OpenAI launched Sora 2 and everyone’s having a field day. This is basically Tiktok but instead of real humans, it’s just AI!

We all know it but it’s highly addictive (for now?) with people making funny memes of:

• Sam Altman Shoplifting GPUS at target to keep Sora 2 running

• Sam Altman grilling Pikachu & eating it

• What it’s like being in Minecraft as a human

• Getting arrested by Apple police for using Android

Now what’s different about this is (and why we’re addicted) is:

• You can make Cameos featuring you in the AI videos

• You can make AI videos with anyone that allows their face to be used in Cameos (like Sam Altman - clever marketing play)

• It is highly realistic!

It’s limited to:

• US at the moment (you can use VPN)

• Then you need an invite (here’s how to get it)

• THEN only available on iPhone to create Cameos of yourself (lol I need an iPhone, I’m feeling FOMO right now)

But seriously are we AI, what’s real anymore?!

🚀Ways I Can Help You

1. ​Advertise with us (56,000+ Email, 850k+ Social media)

* Disclaimer: The content provided herein is intended for entertainment and informational purposes only. It is not intended as, nor should it be construed as financial, legal, tax, investment, or other professional advice. Some of the links in this article may be affiliate links, which can provide compensation to me at no cost to you if you decide to purchase a paid plan. You should consult a professional advisor before making any decisions or taking any actions based on the information provided.

TikTok | Learn | Advertise | Unsubscribe

Hi {{First Name|Guys!}}!

Anyone noticed Google has really been shipping lately?

Seems like every day there is a little gem that was quietly released yet is super underated!

For example:

1. Google Stitch - THE BEST design for your apps

   

• I cannot stress this enough, of all the AI App builders I’ve used this is hands down the best. Did I mention it’s free too?

• It creates the most well thoughtout UI/UX designs yet stylish

• Only downside is it doesn’t create the app for you (only the design) so you will need to import this over to an AI App builder like Lovable, Bolt or Base44 etc.

• Ideas: Create a property inspection app to generate reports while on the go with pictures, a finance app for solopreneurs to give tailored recommendations on how to save or grow revenue based on data, or snap a pic & sell your products instantly (let AI find out RRP, list it for you). See example below or my reel here

Google Stitch

2. Notebook LM - Video lessons

   • I know what you’re thinking, it’s old news. But it’s not!

   • Apart from being able to upload 50 docs & working within this “sandbox” to create realistic podcast, flash cards, quizes & reports, it makes the most AMAZING video explainers

   • I think I might be out of a job soon because it was able to explain concepts like RAG, MCP, Function calling & AI Agents in such a way that was quite honestly the best I’ve ever heard

   • The video presentation slides are also top notch with super clean yet stylish way to present dry & technical info. It actually looks fun and easy

   • I’ve uploaded it here if you wanted to take a listen

   • Here’s my short reel to show you how to access this video feature

   • Something similiar is their new “Learn Your Way”

3. Google Videos - FREE alternative to Loom

   

   • I can’t believe this is free.

   • It can do what Loom cannot which is not only record you + your screen but in segments!

   • It’s like recording in Tiktok where you can stop/start and each time you do this, it creates a segment (instead of messing up a line and having to decide whether to create an entire new video or keep going & finding that little needle in a haystack come editing time). This makes it so much easier to edit.

   • In the paid plans (Google workspace), you can create AI videos using Veo 3, AI Avatars & story boards all in the same space!

What I’ve been doing this week:

Casually hung out with Steven Johnson (Co-founder of Notebook LM) at Google HQ😉 

I remember posting about Notebook LM when it first came out in 2023 and shocked that they quietly released this (seems to be a trend!). Well 1 million of you thought it was cool too😂

It was awesome to be in the same room as Steven listening to his story of how it all started.

He was originally a journalist (not technical or coder) and the idea of Notebook LM came about when he was trying to collect info from a bunch of resources to write his pieces (hence the closed ecosystem of sources you know you can trust + is not trained by the AI model).

He was writing about tech with this as his dream to create it for his own use on substack ... he didn't realise Google was reading this and they reached out to ask if he wanted to build this with them! 😱 and that's how Notebook LM came about!

We saw a preview of something new that is launching specifically for us Australians. Noone has been able to get it right except for this demo 🤪 So stay tuned!

The story was interesting to me because this was happening prior to ChatGPT era (without technical coding knowledge, this solution became reality & one of Google’s earliest successful AI product).

Now, is the best time to be creating your dream idea with the AI apps & tools we have today. To solve a problem you experience everyday.

Natalie Choprasert

Super pumped to have met Steven Johnson, Co-founder of @google 's Notebook LM. I remember posting about this when it first came out and shocked that they quietly released this! Well 1 million of...

www.facebook.com/share/p/1VwhSS2Xkb

🚀Ways I Can Help You

1. ​Advertise with us (56,000+ Email, 850k+ Social media)

* Disclaimer: The content provided herein is intended for entertainment and informational purposes only. It is not intended as, nor should it be construed as financial, legal, tax, investment, or other professional advice. Some of the links in this article may be affiliate links, which can provide compensation to me at no cost to you if you decide to purchase a paid plan. You should consult a professional advisor before making any decisions or taking any actions based on the information provided.

TikTok | Learn | Advertise | Unsubscribe

Hi {{First Name|Guys!}}!

I didn’t think it would happen this early but could we soon have AI running for government? 😱

• First “AI minister” in Albania’s Government?

  • Her name is Diella, a newly appointed “AI member” of cabinet is in charge of tackling the country’s corruption

  • The AI bot would help public tenders be 100% free of corruption, help the government work faster and with full transparency

  • Question is: Will it be free of bias? or be subjected to prompt injection attacks without the public knowing?

• Imagine selling your house using an AI realtor that gives realistic video property tours - using Nano Banana and Veo 3. All you need are photos of the house! LOVE this use case by Duncan Rogoff

  

  Credit: Duncan Rogoff

What I’ve been doing this week:

1. If you wanted to understand the difference in vibe coding tools & AI jargon, here’s my 60 sec video using food 🙂 

2. Up until this point I had been using Lovable for vibe coding AI Apps.

   • This week, Replit launched agents which helps you build AI apps & automations… may have found my new fav vibe code tool!

   • Out of the box with 1 prompt (from ChatGPT) it shipped a nice functional app

   • You can use this mini app inside tools like email or Slack!

   • However I still love Lovable’s design & look

3. Claude launched the ability to create & edit powerpoints, excel, PDFs & fill out forms (my video here)! I tested this by asking it to:

   • Fill out PDF forms: I uploaded the form I needed to fill out (gym registration) + a fact find doc (all my information).

     Claude was able to match the questions with my info and fill out the gym registration so I could download it as a PDF!

     This is great if you have a tonne of forms you need to fill out!

   

   • Create a presentation that can be edited in powerpoint from a URL:

     • What impressed me was how accurate it was

     • Picked out key points that mattered

     • Every component was editable inside Powerpoint

     

     • Comparison: Someone asked me how Claude compares to GenSpark which has been offering presentation creation for a while now. Turns out, given the same prompt, GenSpark looks more professional and detailed with graphs, better at presenting the information & formatted.

4. Just for fun if you haven’t tried this already on nano banana (Gemini), it may be quite addictive - generating 3D versions of yourself!

   • Go to Gemini

   • Upload a photo of yourself

   • Prompt below:
     “Create a 1/7 scale commercialized figure of the character in the illustration, in a realistic style and environment. Place the figure on o computer desk, using a circular transparent acrylic base without any text. On the computer screen, display the ZBrush modeling process of the figure. Next to the computer screen, place a BANDAl-style toy packaging box printed with the original artwork.”

Until next week!

Cheers,

Nat

🚀Ways I Can Help You

1. ​Advertise with us (56,000+ Email, 850k+ Social media)

* Disclaimer: The content provided herein is intended for entertainment and informational purposes only. It is not intended as, nor should it be construed as financial, legal, tax, investment, or other professional advice. Some of the links in this article may be affiliate links, which can provide compensation to me at no cost to you if you decide to purchase a paid plan. You should consult a professional advisor before making any decisions or taking any actions based on the information provided.

TikTok | Learn | Advertise | Unsubscribe

Hi {{First Name|Guys!}}!

This week I saw alot in the pipeline relating to business and work (not just AI fun and games like nano banana images which have been flooding my feeds lol).

1. Atlassian just aquired The Browser Company (creators of Dia & Arc):

   • Goal? AI browser designed for business & Saas tools (Finally!)

   • How? Atlassian has the enterprise/data/security “know how” combined with Dia & Arc’s AI browser capabilities

   • They now join the AI browser race with some big players Google Chrome, Comet (Perplexity), and ChatGPT (soon)

   • Or watch my short summary here

2. OpenAI to launch AI Job Platform next year:

   • RIP Linkedin? It won’t just be job listings, it will actively connect employers with workers WITH AI skills.

   • It will allow workers to show their AI skills & verify making it easier for both parties

   • Launching a certification program to train & upskill 10 million workers

   • Also did you know Microsoft owns Linkedin?🫠 

   • If anyone’s looking for a content strategist role at ChatGPT, it’s paying up to $393k + equity 🤪 

3. Top 50 GenAI Web Products (updated) - Interesting trends here:

   • Google has now climbed up to 2nd most popular

   • China’s AI booming (makes up 9 out of 50 most visited AI tools)

   • New trend in vibe coding apps

   • 20% of top AI products are in providing company as a friend, roleplaying & romance bots

   • Read my summary here

What I’ve been doing this week:

1. Created a little list of Top 10 AI tools for solopreneurs. Here’s the video.

2. If you’ve been trying to make sense of how vibe coding, MCP, Claude Code, Codex CLI all fit together, I explain this in under 60 sec using food 🙂 

3. If you can’t prompt for ****, here is the master prompt to create a “$500/hr consultant” to prompt for you!
   This guy on Reddit tried 147 times to get it right (just steal it).

   Get it to make you a detailed + well structured prompt for your Custom GPT, Claude Projects so you can get better + consistent answers!

Cheers,

Nat

🚀Ways I Can Help You

1. ​Advertise with us (56,000+ Email, 850k+ Social media)

* Disclaimer: The content provided herein is intended for entertainment and informational purposes only. It is not intended as, nor should it be construed as financial, legal, tax, investment, or other professional advice. Some of the links in this article may be affiliate links, which can provide compensation to me at no cost to you if you decide to purchase a paid plan. You should consult a professional advisor before making any decisions or taking any actions based on the information provided.