Good Morning,

I’m still digesting the Google I/O conference news and some of the key AI trends I’m noticing and while I do so, I wanted to share something a bit different from the usual. It’s a potentially really useful guide since we are in an AI boom of unprecedented opportunity and stock market euphoria that is perhaps a unique window.

Michael Fritzell is an independent financial analyst, investor, and writer based in Singapore. He is best known as the founder and author of Asian Century Stocks, a popular investment newsletter. Michael is a friend of mine and one of the most successful financial writers I know who grew his Newsletter on Substack pretty early on.

This is going to be a guest post by Michael Fritzell who runs one of the most useful investing Newsletters I know of especially for Asian equities. I wanted to share this practical guide since how investors and analysts are using AI in Finance is changing very fast. This is a hands-on guide that should give you a lot of ideas on where to start to use Claude for your investing decision making.

Sign up to Asia Century Stocks

Michael’s Newsletter is no longer on Substack but moved to the Ghost platform. A lot of investors are of course using Claude for their investing activities.

The author briefly describes his Newsletter publication: 1 min 28 seconds. 🎬

• You can find more of his recent interviews with Substack investing, FinTwit and equity folk here.

Many of my readers are also investors and think about technology and businesses at a very deep level.

• Read his recent Deep Dives.

• Follow Michael Fritzell’s Portfolio.

• View his Thematic Reports

If you want to dig deeper to get full access to the work of someone that’s been studying Asian equities for nearly two decades, consider a full sub to Asian Century Stocks.

Get full Access to ACS

One way to think of using Claude for Equity Research is the following:

1. Research

2. Projects

3. Skills

4. Cowork

Claude has evolved a lot when it comes to Agents for Financial Services. Download Claude Cowork.

Some Starting Ideas

• Investors instruct Claude Code to scan the web, pull down the latest 10-K forms, SEC filings, and earnings call transcripts directly into local directories.

• Live portfolio analysis: with Anthropic's Model Context Protocol (MCP), users are plugging Claude directly into live financial databases and brokerages.

Claude for Equity Research

By

Biographical Sketch 🏛️💰💸💳💶

• Originally from Sweden, Fritzell started his career in investment banking in the United Kingdom after completing his university degree.

• After studying Chinese, he moved to Shanghai to work as a buy-side analyst for an emerging markets fund, traveling extensively across China to evaluate investment opportunities and research local companies.

• In 2014, he relocated to Singapore to manage an Asian equity portfolio for a European family office, expanding his coverage to markets like Indonesia, Thailand, Vietnam, South Korea, and the Philippines. He later worked for a hedge fund and a mutual fund in Singapore.

• In April 2021, Fritzell left the traditional fund management industry to write Asian Century Stocks full-time. He’s now based in Singapore.

Michael is a professional analyst, he always includes a disclaimer so the rest of the article is by him:

Disclaimer: This article constitutes the author’s personal views and is for entertainment and educational purposes only. It is not to be construed as financial advice in any shape or form. Please do your own research and seek your own advice from a qualified financial advisor. From time to time, the author might hold positions in the below-mentioned stocks consistent with the views and opinions expressed in this article. This is a disclosure, not a recommendation to buy or sell stocks.

2026 has been the year of Anthropic.

They are the creator of the large language model Claude, which has taken the world by storm:

The company was founded by a group of ex-OpenAI executives, who purportedly set out to create a more ethical generative AI tool. The result was Claude.ai – a chatbot that functions much the same way as ChatGPT and Google Gemini.

However, the popularity of Claude did not take off until late January 2026. At the time, Anthropic released Claude Cowork — an AI colleague that can work with files on your computer using natural language prompts.

And since then, Anthropic has been on an epic run of product releases, including Claude Design, Claude Marketplace, Claude for Excel, Claude for PowerPoint, Claude for Word, Claude for Chrome, and Claude Mythos.

In this post, I’ll explain exactly how I use Claude to analyze stocks. And hopefully, you’ll learn something in the process.

Personalized instructions

First, go to claude.ai and create an account.

There’s a free plan that gives you access to the basic model (Claude Sonnet) with up to 100 messages per day.

There’s also a Pro plan for US$20/month that gives you access to the reasoning model Opus 4.7, with 5x more usage than the free tier, access to Claude for Microsoft Office, and Claude Cowork. I suggest buying the Pro plan.

When you first go to the website, Claude won’t know much about you. So I suggest going to your profile (in my case, “MF”) and then “Settings”. You’ll now be able to provide context on what you’re looking for from Claude.

In my case, I tell Claude that:

“I am Michael Fritzell, an equity research analyst. Whenever I ask about a particular company, I want to understand the fundamentals of the business model – in other words, how it makes money. I also want to understand what makes the company unique in comparison to its competitors. Be brief and to the point.”

If you’ve used Claude for a while, it will have saved a memory of past discussions. You can access that memory by clicking “Memory from your chats”:

If you want to add anything to the memory, click the pen button at the lower-left corner of the screen and add things you’ll want Claude to keep in mind.

You can also import the memory from other generative AI tools. Just click “Settings” and then “Capabilities”, and follow the instructions after clicking “Start Import”.

Claude will now have much greater context on who you are and what you’re looking for.

Managing the main prompt window

While visiting the website, you can reach Claude’s main prompt window by clicking “Ctrl/⌘ + Shift + O”. You’ll then be presented with the following window:

To add files, click the “+” sign, type “/” or drag and drop a file from your computer. If it’s a picture, you can also just paste it by typing “Ctrl/⌘ + V”.

Use Sonnet 4.6 for routine questions and tasks and Opus 4.7 for complex, multi-step analysis (if you’re on the Pro plan).

If you click the sound button in the lower-right corner of the screen, you’ll get to Claude’s voice mode, where you can talk to the AI model using your computer’s microphone. I’ve found Claude’s live mode better than Gemini’s but worse than ChatGPT’s. See what you think.

Save AI prompts in Claude Projects

In the past three years, I’ve been saving my AI prompts in Apple’s Notes app. This has forced me to copy the prompt text each time and then manually paste it into the prompt window.

This is obviously a poor use of time. These days, I’ve instead automated the process by switching to Claude’s Projects feature. You can think of it like a workspace for particular tasks.

The way I use Claude Projects is to save generative AI prompts, such as the following ones:

Company overview

Explain the company’s business model in simple terms. What are its key products and services? Who are its main customers, suppliers, and competitors? What are the contracts and key payment terms?

Bull vs bear

“Act as an institutional-grade equity analyst. Perform a deep-dive, adversarial analysis of the company. Start with the bull case (competitive advantages, moat sustainability, growth levers including secular tailwinds or potential earnings surprises, capital allocation). Then the bear case (2-3 risks that could permanently impair the business, potential margin compression or revenue deceleration, high expectations). Perform a pre-mortem. Consider whether current valuation multiples are too high. Finally, generate a contrarian view of what the market is currently refusing to see.”

Competitive advantages

I want to understand the strength of the company’s products and business. How do the products compare with competitors’ in terms of perceived value, branding, marketing, etc.? What competitive advantages does the company have that will protect it from future competition, so-called economic moats? And what is the company’s bargaining power over other stakeholders?

Supply chain

Explain the supply chain that the company operates in and where the company sits within it. I want the output to be a map from upstream inputs all the way to the end customer, with all stakeholders accounted for. Make sure to capture the names of all the companies interacting with the company.

Segments

Give me a breakdown of revenue, EBITDA, and earnings by segment. Describe how each of these numbers has changed over time and why. Discuss both product segments and geographical segments.

Earnings result

Analyze the company’s latest earnings result. Revenue & profit vs. expectations. Did the company beat or miss consensus? By how much? Key segment drivers. Which business lines drove the result? Any notable acceleration or deceleration? Margin trends. What happened to gross/operating margins and why? Guidance & outlook. What did management guide for next quarter/full year? Any change in tone? Balance sheet flags. Anything notable in cash flow, inventory, receivables, or debt? Market reaction. How did the stock react and what does that signal about what was priced in? Flag anything that looks unusual relative to the company’s recent history.

Earnings calls

Create a summary of the company’s recent earnings calls and tell me what management is focused on. Perform sentiment analysis and describe how sentiment has shifted over time.

Management

For the company, give me a brief assessment of the CEO and key executives: 1) Track record: What have they actually built, turned around, or delivered in prior roles? Quantify where possible. 2) Tenure & insider ownership — How long in the role, and how much skin in the game? 3) Capital allocation history — Do they reinvest wisely, acquire disciplined, or destroy value? ROE/ROIC trend under their watch. 4) Red flags — Related-party transactions, excessive comp, frequent strategy pivots, or promotional behavior. 5) Founder vs. professional manager — Which archetype, and what does that imply for this stage of the business?

Stock price analysis

Help me understand historical catalysts behind the company’s stock price. What news or events moved the stock up or down more than 5% in the last 5 years?

Comps

Generate a comparables table with the company and its key global peers within the industry. Include the Bloomberg ticker, USD market cap, and the following valuation multiples: EV/Sales, EV/EBIT, P/E, and dividend yield. Also include the 5-year average return on equity for each peer.

Forward projection

Estimate the company’s earnings per share in the following three years from 2025 to 2027. Consider industry growth, market share gains, price increases, cost pressures, operating leverage, financing costs and share count dilution in your final earnings per share estimates.

Red flags

“You are a forensic equity analyst. Identify red flags and accounting risks in the financial statements of the company, including revenue recognition, segment reporting, leases, related parties, contingencies, stock-based comp, goodwill/intangibles across the income statement, balance sheet and cash flow statement.”

Management questions

Create 15 precise questions for the CEO about the company’s long-term strategy, competitive advantages, capital allocation and risks that they see on the horizon. Order by information value.

Devil’s advocate

“You are a skeptical short-seller analyzing the company. Your job is to dismantle the bull case. What could structurally break the way this company makes money? Where is the revenue concentrated, and what happens if that concentration shifts? Why might the moat be weaker than bulls think? Who is the most dangerous competitor that bulls are underestimating, and why?What are the worst ways management has allocated capital? Any related-party transactions, aggressive accounting, or misaligned incentives? What assumptions need to hold for the current price to be justified? What happens to the valuation if growth disappoints by 20-30%? What is the single scenario that would permanently impair this business, and how plausible is it?”

Create a separate Project for each of these prompts. Then click “Set project instructions” for each of them. Tell it to use each of the above prompts.

For example, for the “Company overview” prompt above, I input the following and then click “Save instructions”.

What this does is allow us to analyze a stock without having to enter the entire prompt. Instead, I just click the Project name “Company overview”, enter the ticker and then press Enter:

I then get an excellent reply with minimal effort:

If you don’t like the reply from any of the prompts, just go to its Project and adjust it accordingly.

To streamline the process even more, click the Star sign for each of these Projects. They’ll now appear in the sidebar on the left side of your window.

Any time you want to perform analysis, just click each Project, then type in the ticker.

Some investors use Claude Projects the same way you might use NotebookLM: to save files for particular stocks, including annual reports, earnings call transcripts, etc. But I prefer NotebookLM when it comes to analyzing a large number of documents. NotebookLM has a richer feature set. For example, it can generate podcasts, connect to YouTube and Gmail, and easily handle hundreds of sources.

Claude for Word/Excel/PowerPoint

Now let’s install Claude for Word/Excel/PowerPoint.

Go to Microsoft’s Marketplace, click “Get it now” for the app plug-ins you want, then follow the installation and login process.

Next time you open any of these apps, you can ask Claude to analyze or modify the file directly from the associated chat window. For example, you can open the Word file with notes from a previous conference call, and ask for the greatest risks mentioned by management in the call:

Another example from my personal workflow is to open Microsoft Excel, drag and drop Coupang’s 4Q2025 earnings presentation into the Claude sidebar and perform a segment analysis:

“Find 2025 revenues by product segment and geographical segment and summarize them in pie charts”

You’re then presented with the following output:

Other use cases include checking an Excel sheet for formula mistakes, summarizing long documents, and checking for factual or spelling mistakes in my Asian Century Stocks PowerPoint presentations.

Claude Cowork for automating tasks

Next, we’ll download Claude’s desktop app. It’s only available for MacOS right now, but it will probably be available on Windows soon.

Inside the app, there are three modes to choose from

1. Claude Chat

2. Claude Code

3. Claude Cowork

In short, Chat is for thinking through problems. Code is for building software. And Cowork for running tasks on your computer.

Once you’ve installed the Claude desktop app and have it running in the background, you can access it anytime by double-clicking the Options key on your Mac. You’ll then be presented with the following pop-up window:

In my view, this is the fastest way to get an answer to a question, much faster than opening your web browser and searching on Google.

I also have the Caps Lock button connected to Claude’s voice mode. If you click Caps Lock at any time, the voice mode will be activated, and you can simply speak into your microphone to get an answer:

But let’s dig into Claude Cowork specifically. You can find the Cowork button at the top-left corner of the Claude desktop app:

Every time you deal with Cowork, you’ll want to give it access to a folder on your computer. Just click the “Work in a project” button and then click “Choose a different folder”

With that file access, Claude Cowork can now view, create and modify files. You should allow the app to modify files in the folder, but don’t grant it access to your entire hard drive.

Here’s an example of how I might use Cowork. Yesterday, I downloaded the latest eight earnings call transcripts for the Korean e-commerce company Coupang into a folder, and then asked it to analyze the recent trend in its earnings call sentiment. It then gives me the following output, which is saved in a separate Microsoft Word document called “Coupang earnings sentiment analysis.docx”:

If you click “Instructions”, you can tell Cowork exactly how you’ll want the output for any particular task to look. These instructions will be saved in a separate file called “Claude.md” in your chosen folder.

Other than earnings call sentiment analysis, I might use Cowork for finding obscure data, translation tasks, summarizing documents, conducting forensic analysis on financial statements, etc.

But, in my view, where Claude Cowork really shines is where you need to create or modify existing files. In the Coupang folder mentioned earlier, I have an Excel file called Coupang.xlsx as well as its recent annual reports. Knowing this, I can ask Cowork to take Coupang’s income statement and insert it into a new Excel sheet:

“Create a new sheet in Coupang.xlsx where you paste the last 3 years of Coupang’s income statement from its annual reports into a nicely formatted table”

This is what I’m then presented with, inside my Coupang.xlsx Excel file:

Pretty impressive. Just be aware that generative AI tools sometimes make mistakes. So always double-check the output.

You can also connect Claude Cowork with other apps. To do so, click “Customize” in the sidebar on the left-hand side of the app and then click “Connect your apps”. I suggest adding Google Calendar, Google Drive (which includes Google Sheets), Slack and Notion.

You’ll also want to search for “Control Chrome” and install that connector. On top of installing the Claude extension inside Google Chrome. By doing this, you’ll allow Claude Cowork to browse the web for you to download documents, scrape data off websites, etc.

With connections, you can automate processes using each of these apps. For example, I have a 400-line watchlist in Google Sheets that I track weekly. To automate that process, I enter the following prompt:

“Read my watchlist from the Google Sheet “Asian Century Stocks Watchlist”. Find the stocks that have moved more than 10% in the last 7 days. For those stocks, scan the Internet for earnings results, guidance changes, regulatory actions, analyst upgrades/downgrades, and material corporate events. Summarize the findings in a markdown report grouped by market (ASX, HKEX, SET, PSE, etc.), with a one-line summary per stock and a ‘Notable’ flag for anything that warrants deeper analysis. Skip stocks with no material news. Save the report as weekly-scan-[date].md to /Documents/Watchlist/.”

Claude Cowork then saves a file on my computer called “weekly-scan-2026-04-21.md”, containing the following information:

What the text file looks like in a typical note-taking app (in this case, Obsidian)

Since I will want to run this task periodically, I save the routine in what Claude calls a “skill”, so that Claude knows what to do each time:

“Create a skill from this workflow called /weekly-watchlist-scan that captures the process, format, and output location we just refined.”

Finally, I schedule the task to be repeated on each Monday morning at 9:00am by typing “/schedule the weekly watchlist scan every Monday morning at 9:00am SGT”.

As long as the computer is turned on and Claude Cowork is open, it’ll perform the task and save each week’s output as another “weekly-scan” file in my Watchlist folder.

Claude Dispatch

Next, we’ll learn how to control your computer using your phone.

Open the Claude Desktop app, click Cowork and then “Dispatch” in the sidebar to your left:

With Dispatch, you’ll be able to control your computer using the Claude app on your phone, as well as all the apps that are connected to it:

Every time you want to control your computer, just open the Claude app on your phone and click “Dispatch” to open the window where you give instructions to your computer.

When I’m on the run, I might want to retrieve a document from my computer. So I asked Dispatch to retrieve the Coupang earnings sentiment analysis document that I just created:

“Retrieve the Coupang Earnings Sentiment Analysis saved in my Coupang folder”

I’m then presented with the following output within Dispatch:

So it’s very useful, as you can see. With Dispatch, you can work remotely from just your phone using natural language.

Conclusion

Claude isn’t perfect. It’s a little slow at times, and if you’re a heavy user, you’ll quickly exceed your usage quota. Many of us will feel compelled to upgrade to the US$100/month plan.

But I think the Google Chrome plug-in, Microsoft Office plug-ins, Cowork, Dispatch and the other Claude tools will fundamentally change the way we operate. It’s time for us to learn these tools to automate some of our daily work.

In practice, organizing my most-used prompts through Claude Projects might be the one trick that has saved me the most time. Today, when I’m interested in a company, I just click each Project, enter the ticker and voila — I get exactly the output I want.

I’ve also found it helpful to throw a document into Claude for Excel and have it present the data in a nicely formatted table in a spreadsheet.

I still haven’t used Cowork much, as I prefer working with files in the cloud. But I can see why others find it helpful.

Now that I’ve gone through my usage habits, I’m interested in hearing how you’re finding Claude. What tricks are you using to automate your research process?

Michael

Get Full Access

This was an article from investment newsletter Asian Century Stocks. Sign up for free here.

• About two months ago the celebrated Creator and media analyst interviewed Michael you can catch it here:

Thanks for reading! And I hope you learned something new.

• Visit the Full Library of Asia Century Stocks.

Good Morning,

Are we entering the autonomous AI era? From seemingly useful tools like Claude Cowork to Microsoft’s own spin on a researcher agent called Critique as part of its rath…

Read more

This Newsletter usually goes out at 05:30 AM EST. On rare occasions it’s a bit late (like today).

Good Morning,

I wanted to update you on some of Google’s AI improvements that caught my attention. You can follow the NotebookLM Twitter. At the beginning of March we learned that NotebookLM is introducing Cinematic Video Overviews, a major update to its AI-powered video creation capabilities.

• NotebookLM leverages Gemini 3, Nano Banana Pro, and Veo 3 to offer “fluid animations and rich, detailed visuals to help you learn and engage with the topics you care about. Users with a Google AI Ultra subscription can turn their notes into personalized, fully animated videos. Users can also only generate a max of 20 cinematic video overviews per day.

• Google is saying that NotebookLM has moved beyond a simple “research assistant”, evolving into a full "Research-to-Content" pipeline. Read how to do it.

Read Blog

Google Maps Getting a Huge AI Update

Ask Maps is a new way to get answers to your complex, real-world questions with a simple conversation. These updates to Google Maps look fairly useful. They are calling this “its biggest upgrade in over a decade.” By combining Gemini models with a deep understanding of the world, Maps now unlocks entirely new possibilities for how you navigate and explore. Learn more.

Ask Maps

The first big change is a conversational feature called "Ask Maps", which is designed to let you ask Google Maps more complicated questions that it never could have handled before.

I asked for ideas for beginners on using Gemini with NotebookLM.

The AI-Augmented Engineer

The vibe-worker’s guide to software engineering, AI workflows and coding interfaces.

The AI-Augmented Engineer

Accelerating software engineering careers with AI workflows. I show you how to use AI to write better code, ship faster, and get ahead.

By Jeff Morhous

NotebookLM And Google AI ecosystem related Articles

• Using NotebookLM to learn new things

• 3 Hidden NotebookLM Features Most People Don’t Use

• The ultimate NotebookLM tutorial

• 58 Nano Banana 2 Prompts for Infographics (With Examples)

• 9 NotebookLM Prompts That Supercharge Productivity

• We Blind-Tested ChatGPT vs Claude vs Gemini — Here’s the Winner (2026)

• Every Google AI Tool in 2026: What Each One Does and When to Use It

• How Google Workspace CLI Made My Claude Code Setup 10x More Powerful

• How I Connected NotebookLM to Claude and Changed How I Do Research Forever

Gemini Embedding 2:

Google’s first multimodal embedding model. It can map text, images, video, audio, and PDFs into a single unified space, which is a massive leap for developers building advanced search and RAG (Retrieval-Augmented Generation) systems.

The model is based on Gemini and leverages its best-in-class multimodal understanding capabilities to create high-quality embeddings across:

• Text: supports an expansive context of up to 8192 input tokens

• Images: capable of processing up to 6 images per request, supporting PNG and JPEG formats

• Videos: supports up to 120 seconds of video input in MP4 and MOV formats

• Audio: natively ingests and embeds audio data without needing intermediate text transcriptions

• Documents: directly embed PDFs up to 6 pages long

As you might know, Embeddings are the technology that power experiences in many Google products. While many people have been switching from ChatGPT to Claude, I think Gemini’s offering is worth it now in 2026 if you use Google’s ecosystem.

Considering upgrading your Gemini

Let me just say I love the idea of a smarter Google Maps.

In a Nutshell what does it mean for just the average person though?

If you know someone who uses Google Maps and would want to know, feel free to share it with them.

Share

So convenience wise what’s actually changed or changing?

New Experiences in Google Maps

I’ve tried my best to break down the key new experiences you can expect over the coming months if you are an Android and Google Maps user.

• Ask Maps: A new conversational assistant powered by Gemini. Instead of just searching for keywords, you can ask complex, natural-language questions like, “Where can I charge my phone without a long wait for coffee?” or “Find a restaurant with vegan options and easy parking along my route.”

• Immersive Navigation: This replaces the traditional flat 2D map with a vivid 3D view of your entire route. It uses AI to stitch together Street View and aerial imagery, showing realistic buildings, overpasses, and terrain to help you orient yourself better in unfamiliar areas.

• Enhanced Road Details: During navigation, the map now highlights specific details like lane markings, crosswalks, traffic lights, and stop signs. This is designed to help drivers prepare for tricky turns and lane changes well in advance.

• Transparent Buildings: As you approach complex turns, nearby buildings can become translucent on the screen. This “X-ray” style view allows you to see the road ahead even if it curves behind a large structure.

• Contextual Route Comparisons: When Maps suggests an alternate route, it now explains the trade-offs in plain language—for example, telling you a route is “3 minutes longer but avoids heavy highway construction.”

• Arrival Guidance: To simplify the “last mile” of a trip, the app now highlights the specific entrance of a building and provides recommendations for nearby parking as you get close to your destination.

• Natural Voice Guidance: Voice directions have been updated to sound more human and use landmarks. Instead of “In 500 feet, turn right,” you might hear, “Go past this exit and take the next one for Illinois 43 South.”

Google Gemini is Taking Marketshare

According to Similarweb, Gemini (in pink) continues to make strides in terms of worldwide traffic and active users.

Gemini has been taking ChatGPT Marketshare in some domains

We like to talk about the Claude vs. ChatGPT rivalry lately, but Google is the one that’s made the most progress in terms of building a full-stack AI ecosystem.

More ChatGPT Users are also using Gemini

As Gemini, Claude, Grok and others reach near parity with ChatGPT, it becomes more about preference, tone and the “personality” of the chatbot (or brand of the company) for many everyday users.

• Simliarweb has a lot of recent data they have shared on this here (in mid March, 2026).

NotebookLM Overtook Perplexity is Website Visits in Early 2026

Using NotebookLM with Gemini

Read more

View more guides like this one.

Good Morning

It’s Chinese New Year period in these parts so my cadence may be lower than usual in mid to l…

Read more

Good Morning,

You can do so much with AI. The building and DIY aspect also keeps getting more nuanced, and more powerful. Open-source AI is providing a new array of capabilities even at the local and individual level.

I’m a huge fan of and I’ve wanted to share more about his work for so long. Today, we finally have the chance. Ben is an independent AI researcher (LLM, NLP) with two really useful blogs and I have a huge respect for his work: (don’t let the funny names fool you, these are serious resources).

The Kaitchup – AI on a Budget 🍅

Hands on AI tutorials and news on how adapting language language models in a DIY setting to your tasks and hardware using the most recent techniques and models.

The Kaitchup – AI on a Budget

Weekly tutorials and news on adapting large language models (LLMs) to your tasks and hardware using the most recent techniques and models. The Kaitchup proposes a collection of 170+ AI notebooks regularly updated.

By Benjamin Marie

The Kaitchup publishes invaluable weekly tutorials with info that’s hard to find elsewhere.

• By being a paid subscriber to The Kaitchup, you also get access to all the AI notebooks (160+), hands-on tutorials, and more in-depth analyses of recently published scientific papers.

Learn Deeper

Read The Salt 🧂

Reviews and in-depth analysis of bleeding edge AI research and how-tos. The Salt is a newsletter for readers who are curious about the Science behind AI. If you want to stay informed of recent progress in AI without reading much, The Salt is for you! I do my best to offer articles that might be interesting for a wide variety of readers.

The Salt - Curated AI

The Salt offers weekly reviews and in-depth analyses of the latest AI papers. If you want to stay informed of recent progress in AI without reading much, The Salt is for you!

By Benjamin Marie

Benjamin’s technical and practical knowledge is invaluable depending on how deep down the rabbit-hole you want to go in DIY with models. It’s not overly technical but it is on technical topics, useful for a wide range of readers interested experimenting locally DIY with models or in small teams.

Selected Works

I asked him for a basic beginners tutorial on how to run LLMs locally (something I sometimes get questions about). He’s able to add so much practical know-how and insights into the latest models where for me he is an authority. If a new model comes out his opinion represents hands-on experience and being up to date on the latest scientific papers.

1. Qwen3-VL: DeepStack Fusion, Interleaved-MRoPE, and a Native 256K Interleaved Context Window.

2. Did the Model See the Benchmark During Training? Detecting LLM Contamination

3. Making LLMs Think Longer: Context, State, and Post-Training Tricks

is an independent researcher focused on hands-on AI and the tools around modern language models. He helps people and companies cut costs by adapting models to their specific tasks and hardware. I hope you learn something from it. While my work doesn’t touch on machine learning professionals that much, more and more individuals and small teals are playing with these open-source models locally.

So I’m very proud to be able to bring you a guide like this:

Run Language Models on Your Computer with LM-Studio

A practical guide to running local models and picking the right one for speed or accuracy.

Running large language models (LLMs) locally used to mean wrestling with the GPU’s software layer (like CUDA), scattered model formats, and a lot of trial-and-error. Today, it’s surprisingly approachable. With tools like Ollama or LM Studio, you can download a model, load it in a few clicks, and start chatting on your own machine, without sending prompts to a cloud service.

This article walks through the practical path from “installing the app” to “running my first local model,” and then zooms out to the part that really matters: what determines whether a model runs smoothly (or not) on your hardware. Along the way, we’ll cover installing LM Studio, the memory (simple) math behind model sizes, how to pick trustworthy GGUF builds and compression levels, how to sanity-check model output, and why “thinking” models can be dramatically better on hard prompts while also being noticeably slower.

The goal is not to turn you into an engineer. It’s to give you enough intuition to choose models confidently, understand what an application like LM Studio is telling you, and avoid the most common “why is this slow / why is this wrong” surprises.

Read more

Good Morning,

I hope this guide finds you well. To check out our growing archive of guides go here. Don’t forget at the end of this article to check out our Claude Code readin…

Read more

If 2025 was the first year of AI agents, how to build them? What should you know before you start? I asked for an in-depth guide on Agent protocols. To check out our other recent guides go here.

A fellow Canadian🍁🇨🇦, she’s now also a Founder: she writes the Data Letter.

video:

THE DATA LETTER

The Data Letter

The Data Letter prevents data disasters that end careers. Built from a decade fixing Fortune 500 systems: PII exposures, cost explosions, pipeline failures, broken models. Weekly toolkits for data practitioners and leaders. Get the free Audit Kit.

By Hodman Murad

Read her introduction to the brand new company she’s build, Asaura AI (Beta launch April, 2026).

Background and Startup Story

Hodman Murad is the founder of Asaura AI, where she pioneers AI designed for executive function. As the author of The Data Letter, she identifies common failures in AI and data systems, such as hallucinated growth metrics and the costly loss of customers due to optimization.

Beyond her technical expertise, she writes Asaura, a publication dedicated to exploring ways AI can be improved to enhance human productivity and mental health. She’s going to be building in public so you can follow along. Her work focuses on connecting high-performance systems with their users.

This article is essentially a mini-book, 8,000 words and 61k characters. As a data scientist and consultant with so much B2B experience, I think she’s a great person to outline some of how AI agents work and for you to learn a bit more about the new protocols. This article might be a bit more technical than most of our pieces, but is meant for actual learning and a deep introduction. That being said, we’ve made it as easy to read as possible.

Her new startup called Asaura AI, is specifically designed for adults with executive dysfunction, people who know what to do and have time, but can't make themselves start. Since the pandemic and with the affordability crisis, and with more people identifying being on a spectrum of various cognitive divergences (ASD, or other), the concept of her product is relevant to a lot of folk. (Building in public notice): She’ll be sharing everything: the research, user interviews, design decisions, technical challenges, and the actual build process.

Asaura AI

A newsletter at the intersection of AI, productivity, and mental health.

By Hodman Murad

Asaura AI’s newsletter is primarily about AI at the intersection of productivity and mental health.

More from Data Letter Newsletter 📊

• How Netflix Does Data Reliability

• Who Owns Data Quality, Anyway?

• How to Detect Model Drift When You Can’t Measure Performance

• Bad Data’s Hidden Toll

Do writers become Builders?

• As more Newsletter writers and creators are becoming startup and product founders, knowing more about how AI agents work might be helpful.

• 2025 was all about the protocols and required scaffolding getting in place and maturing that enables everything about what comes next from Generative AI perhaps finally delivering some societal and broad based macro ROI.

• Hodman’s crystal clear writing style was the right fit to introduce this crucial and complicated topic for AI’s emergence in the decade ahead. Hopefully 2026 will not just help democratize knowledge (as displayed in reasoning models in 2025), but democratize opportunity. A new breed of entrepreneur is being born with AI’s coding capabilities now available in natural language.

In a Nutshell

This is the tl;dr of AI Agent protocols as of January, 2026.

Integration & Data

• Model Context Protocol (MCP): Standardizes how agents connect to local or remote data sources (e.g., Google Drive, Slack).

• OpenAI Function Calling: Standardizes the JSON format for agents to trigger external software tools.

• agents.json: A discovery file that tells agents which capabilities an API or website offers.

Collaboration & Handoff

• Agent-to-Agent (A2A): Allows agents from different vendors (Google, Microsoft, etc.) to talk and share tasks.

• Agent Communication Protocol (ACP): Manages the “memory” and state of long-running agent tasks.

• OASF (Open Agentic Schema): Defines an agent’s “resume” (skills and costs) so other agents can find and hire them.

API & Commerce

• Agent Protocol: A universal API (e.g., POST /task) so developers can swap agent frameworks without changing code.

• AP2 (Agent Payments): Enables agents to pay each other autonomously for services using digital wallets.

Why read this?

Understanding AI agent protocols is critical because they are the "connective tissue" of the 2026 AI economy. Just as TCP/IP turned isolated computers into the Internet, these protocols turn isolated bots into a collaborative ecosystem. If you are hoping to work with AI agents, orchestration and build stuff, these are basic requirements.

Share this post if you think you know a colleague, coworker, founder, friend or entrepreneur / AI enthusiasts who might benefit from it.

Share

The Agent Protocol Handbook

December, 2025, last updated January, 2026.

Full title: The Agent Protocol Handbook: Navigating the New Era of Agentic AI Standards.

Protocol Convergence

The evolution of AI from conversational chatbots to action-oriented agents demands a new infrastructure layer, which the industry has now standardized around a layered stack of open protocols. This is about building the economic infrastructure for a future of autonomous digital entities.

MCP and AP2 form the protocol layer between AI models and the external world.

We’re witnessing the early stages of a fundamental architectural shift in artificial intelligence. The conversation has moved beyond model capabilities to infrastructure design, specifically, how AI systems reliably and securely interact with the external world.

While most attention focuses on model capabilities, a lack of execution infrastructure currently blocks agentic AI’s progress. How do autonomous systems safely access tools, manipulate data, and conduct transactions at scale?

This represents both technical standardization and strategic positioning. The December formation of the Agentic AI Foundation (AAIF) has formalized this standardization, shaping ecosystem development much like TCP/IP defined internet communication or HTTP enabled the web.

In this analysis, we examine MCP (an AAIF founding project) and AP2 (the industry-standard payment layer) as foundational protocols within the emerging agentic ecosystem, exploring their architectural roles and strategic implications for the autonomous agent economy.

🏗️ Chapter 1: Building Agentic AI’s Missing Infrastructure

The transition from chatbots to autonomous agents represents the most significant architectural challenge since the advent of cloud computing. Despite their proficiency in conversation, current AI systems are constrained when it comes to executing tasks in the real world.

Execution as Bottleneck

The bottleneck for agentic AI is its execution infrastructure. Current AI systems can generate detailed instructions for complex tasks, such as booking international travel, outlining investment strategies, or explaining supply chain optimization; however, they cannot execute these tasks autonomously.

Consider the gap:

• AI models can describe processes in detail

• They can analyze options and make recommendations

• They can generate step-by-step plans

But they cannot:

• Access the tools needed to execute those plans

• Conduct transactions across multiple systems

• Maintain security and auditability while acting autonomously

The limitation is the lack of a standardized, secure, and scalable infrastructure for action.

Protocol Layer Opportunity

Historically, technological revolutions follow a predictable pattern: breakthrough invention → protocol standardization → ecosystem explosion. We saw this with:

• Internet TCP/IP-enabled global networking

• Web: HTTP/HTML created the browsing experience

• Cloud: REST APIs standardized service integration

• Mobile: App stores created distribution ecosystems

Agentic AI reached its standardization milestone in December with the launch of the AAIF. In this new landscape, value capture has shifted from controlling the protocol itself to building specialized agent skills, such as Block’s goose framework, and the proprietary orchestrators that run on top of these shared standards.

Two Complementary Standards

The protocol landscape has crystallized into a layered stack:

Model Context Protocol (MCP): The AAIF Standard for Context

• Open standard for tool and data access

• Client-server architecture with local/remote flexibility

• Developer-friendly with a growing open-source ecosystem

• Strategic bet: commoditize integrations, compete on model quality

Agent Payments Protocol (AP2): The Emerging Standard for Secure Transactions

• Focused on secure payments and transactions

• Built on the existing Google ecosystem and partnerships

• Enterprise-grade security and compliance focus

• Strategic bet: leverage existing enterprise relationships and payment infrastructure

These specifications, whether stewarded by the AAIF (like MCP) or backed by broad industry coalitions (such as AP2), serve as the neutral technical baseline for the entire industry.

Why Infrastructure Matters Now

Three converging trends make protocol standardization urgent:

1. Enterprise Demand: Companies need agentic AI that works with existing systems without security compromises

2. Regulatory Scrutiny: Autonomous systems require audit trails and accountability mechanisms

3. Economic Scale: The agent economy needs standardized transaction protocols to achieve network effects

Strategic Stakes in Protocol Wars

For AI companies, the protocol layer represents both defense and offense:

• Defensive: Prevent being locked into competitors’ ecosystems

• Offensive: Create ecosystems that generate network effects and switching costs

For enterprises, the choice is about future-proofing AI investments while maintaining security and control.

For developers, it’s about building on infrastructure that platform changes won’t obsolete.

What’s at Stake

The adoption of these standards will determine:

• Who captures value in the agent economy

• Which companies become ecosystem gatekeepers

• How much interoperability exists between AI systems

• What security and privacy standards become normative

• Whether we get open ecosystems or walled gardens

This analysis examines MCP and AP2 as the foundational layers (context and commerce) that enable a cross-platform agent economy.

Let’s examine MCP’s architecture and analyze the technical and strategic implications of Anthropic’s open protocol approach.

🔌Chapter 2: MCP: Protocol Layer for Agentic AI

MCP sits between AI models and external resources, creating a standardized protocol layer.

Unlock everything.

Read more

Good Morning,

This article is a bonus guide for premium readers.

Why am I so Bullish on Claude Code?

It’s my opinion that Claude Code was the single most impactful AI tool developed in 2025 that will be significantly upgraded and refined in 2026. If I post a lot of articles around Claude Code in the weeks ahead, this is why.

• In 2025 and moving into 2026, Claude Code signifies a shift from "assistive" AI (think: simple autocomplete) to "agentic" AI. While agentic AI may take some months and years to improve, there’s a lot of enthusiasm over the progress made already in 2025 with protocols like MCP and A2A and more to come.

• While Anthropic focuses on Enterprise AI customers, for founders and entrepreneurs, and even Creators, this can translate fairly well.

Beyond Context Engineering: Agentic Orchestration

Claude Code fluency will equal the professional fundamental skills of agentic AI orchestration. In short, if you spent time on AI tools like NotebookLM or ChatGPT in 2024 and 2025, Claude Code could be a worthwhile focus in 2026. I have a number of guides in the queue related to this topic.

I think the bottom line is, context engineering in 2026 will begin to require more Claude Code fluency.

Anthropic's Emergence in AI coding

Claude Code is no longer just a chatbot that writes snippets; it is treated as a “junior developer” that can:

• Navigate Entire Codebases: It maintains awareness of project structures and can find information across hundreds of files.

• Execute Multi-step Tasks: It can plan a feature, modify multiple files, run tests, debug the failures, and create a git commit—all from a single natural language prompt.

• Subagents & Parallelism: Version 2.0 introduced the ability to spawn “subagents.” These specialized threads work in parallel (e.g., one agent writes an API endpoint while another simultaneously writes its unit tests), significantly reducing development time.

In short, prompting becomes far less relevant with things like Claude Skills. Anthropic is expected to continue to lead with the best LLMs for coding and this scales well into new use cases and capabilities.

Of course the capabilities of Claude Code in 2026 are powered by newer models like Claude Sonnet 4.5 and Claude Opus 4.5 that were launched fairly recently in late 2025. So why does this matter? In the context of Claude Code, now Anthropic's flagship agentic CLI tool, Skills are specialized, executable modules that give the AI "procedural expertise" beyond its general training.

How useful this can work for your work in content creation or building products is up to you.

Mohammed is the creator of the Newsletter Digital Creator. He’s from the Republic of Maldives 🇲🇻 🏝️.

The Digital Creator

The Digital Creator

Accelerating Creators with AI Agents. I show how to use AI to start, create, grow, and scale faster.

By Sharyph

He builds roadmaps and guides for builders, content creators and system thinkers. The Digital Creator (about) is a scrappy Newsletter full of tips, how-tos, tricks and some good advice to automate content related tasks. Sharyph hopes to be impactful in your journey of how you write smarter, faster, and bolder with the power of AI.

He’s identified five core skills that handle 80% of product creation work that he goes over in this guide, today’s article. Similar to authors like or that I have featured multiple times (via Brazil and Indonesia respectively), more builders all over the world are experimenting deeply with using AI to improve and automate creator products, content and Newsletters using AI tools.

If you’re ready to systematize your product creation process, Sharyph has also built something for you:

The Creator Launch Kit is his free course. I personally like the idea of digital creators helping other creators. Not everyone can turn these AI tools into actionable products that save stressed out creators time. I try my best to feature international creators from around the world who are emerging writers.

Sharyph’s content seems finely tuned for Creators who hope to build their own products with AI agents. You can see this by some of his approaches to adding value:

Join DC Premium

• Substack SEO + AEO Playbook: Your Substack Posts Are Invisible (Here’s How to Fix It)

• How I Use AI Agents in My Business

• Build Your Digital Product in 7 Day

The following article comes with context engineering templates to use in Claude Code. He’s giving you valuable prompting guide context to speed up your content production.

The 5 Claude Skills Every Digital Product Creator Needs

Unlock this guide

Read more

Get Premium All-Access

Good Morning,

As one of the older AI Newsletters on Substack, I pride myself on getting a variety of voices and perspectives on here with the help of guest contributors. From learning about how to us…

Read more

This is our third article in our learning series on Genspark. See the first here, and the second here. I think my readers know how bullish I am on the skills of - who I asked to take another look at this AI startup (I keep bugging him about). There’s a good reason for this.

Editor’s Introduction to Genspark

In my mind Genspark used to be a Chinese Perplexity of sorts, with many ex members of Baidu. I incredibly liked the product when it focused on search. But it has morphed commercially into a rather impressive Agentic AI first startup. Honestly I can’t believe that they only recently just did their Series B funding.

You might recall seeing Genspark if you ever came across ’s Lean AI Leaderboard 🏆

View Leaderboard

Genspark at of the time of writing ranks 16th globally in the most lucrative lean AI startups in the world. (This is not a sponsored post). As someone that studies AI startups and AI tools, Genspark is a case study that continues to fascinate me. It’s a fair example of how scrappy Chinese AI engineers can be (wherever they choose to be based). The pivots in product and pricing are textbook lean growth.

Try Genspark

Eric Jing, Kay Zhu, and Wen Sang founded Genspark, an AI company specializing in agentic AI for knowledge workers. Eric Jing serves as CEO, Kay Zhu as CTO, and Wen Sang as COO. The company was established in 2023 by Jing and Zhu, former Baidu executives, and later expanded to include Sang as a co-founder. Genspark has since achieved unicorn status with a valuation exceeding $1 billion following a $275 million Series B funding round in November 2025.

Genspark: Unloved in a World full of Loveables (GPT Engineer).

• Genspark remains the sort of AI tool that you rarely hear about compared to the value it can provide knowledge workers. If I only paid for one AI tool, this one might actually be it.

A Prototype of the Flourishing AI lean Startup

However the crazy part is how they keep pivoting adding features. Genspark AI is an all-in-one AI workspace that leverages advanced models (literally the latest days after they are released) to automate complex tasks with a single prompt. It acts as a super agent, generating content such as slides, sheets, documents, images, videos, and even full apps or spreadsheets.

Bring your own AI (BYOAI) - a Hidden Gem 💎

Originally launched as an AI search engine for synthesizing information, it has evolved into a no-code platform for personal agents. Genspark excels in productivity by handling research, phone calls, presentations, and creative workflows without robotic outputs.

• Basically if you like Perplexity and use Chatbots for research, you’ll likely like Genspark. I cannot praise this thing enough.

On November 20th, 2025 the small AI startup announced Genspark AI Workspace. As you might have guessed Genspark AI Workspace is an “all-in-one” AI platform designed to automate knowledge work by consolidating research, communication (email/meetings), and content creation (docs, slides, sheets, visuals) into a single environment, putting tasks on “autopilot” with high-level prompts instead of manual steps. Basically, it can do a lot of useful stuff at a fairly high level.

What is Super Agent?

Genspark is an AI super agent that automates complex, multi-step tasks by combining multiple AI models and tools to execute high-level prompts, going beyond simple assistance to deliver finished results like presentations, videos, code, and even making phone calls for you. So for example tasks it excels at are for research with citations, content creation (slides, videos, podcasts, reports, websites), data analysis (sheets), and real-world actions (phone calls for bookings/inquiries), functioning as an autonomous agent that plans, researches, and executes tasks with minimal user input.

How does it work?

Super Agent and the approach that Genspark has evolved into is being fairly utilitarian (which I can respect).

Key Capabilities:

• Multi-Agent Workflows: Uses several specialized AIs (chat, code, image, video) for one request, cross-verifying and compiling outputs for better results.

• Autonomous Task Execution: Breaks down complex tasks (e.g., “plan a trip to Japan”) into steps, executes them, and provides the final output.

• Deep Research & Reporting: Conducts comprehensive research, fact-checks, and generates detailed reports or Wikipedia-style pages with citations and chatbots.

• Content Generation: Creates professional slides, spreadsheets, short videos, animated explainers, podcasts, websites, and lead magnets from simple prompts.

• Real-World Actions: Features a “Call For Me” function to make phone calls (e.g., to check restaurant availability or book tables) and summarize the interaction.

• Productivity Tools: Generates resumes, organizes data into tables, assists with legal/finance matters, and integrates with platforms like Google, Slack, and Notion.

The Human Interface and backend?

1. Input Prompt: You give Genspark a high-level goal (e.g., “Create a presentation on AI trends for investors”).

2. Agent Orchestration: Claude (or similar) coordinates multiple specialized AI models (like GPT-4) to tackle parts of the task.

3. Execution: The agent researches online, analyzes data, generates content (text, images, code), and performs actions.

4. Verification: Outputs are checked for quality and accuracy.

5. Deliverable: A finished product (like a slide deck, video, or summary) is provided, ready for use, often with options to export or customize further.

Key Aspects of Genspark in 2025:

• Autonomous Workflow: It uses multiple AI agents to fulfill requests, acting as an all-in-one workspace to reduce app switching.

• Multi-Modal Creation: Generates various content types, including documents, presentations, spreadsheets, images, and videos from simple prompts.

• Advanced Technology: Orchestrates various frontier models and open-source AI, supported by extensive tools and datasets.

• Agentic Capabilities: Can perform complex actions like making calls and integrating with business apps.

You can expect this list to change as well. Genspark is actively finding not just their product market fit but a path to really stand out from other AI startups in my opinion.

has this Newsletter Why Try AI that is super useful not just for keeping up to date on AI News but exploring AI tools. I would consider it beginner friendly but also specialist friendly. Daniel is a down-to-earth Ukrainian guy living in Denmark with a keen eye for experimenting with AI tools. His writing style is super approachable, casual, useful, informative and funny.

“Daniel loves experimenting with creative GenAI tools like image generation, AI video, AI music platforms, and so on.”

Why Try AI

Hype-free, hands-on AI for non-techies.

By Daniel Nest

Genspark Related by Daniel & Greatest Hits

• Here Are My Go-To AI Tools (kind of viral)

• Don’t Sleep On Genspark’s Super Agent

• Minimum Viable Prompt: Your Cure for AI Overwhelm

• “Help Me Help You.” How to Get AI Chatbots To Prompt Themselves.

• The Skeptical Writer’s Guide to AI

Support hands-on for AI Technies

14 New Genspark Features You Must Try

True to his method to find the most actionable AI tools, Daniel lists some of what makes Genspark really useful for knowledge workers in a fairly comprehensive listicle format. If you are considering trying Genspark in depth, you really should read this. This is a pretty deep dive but also an easy to follow overview of Genspark’s main capabilities (it’s deeper than you think).

The article also catches you up on Genspark’s frantic product innovation - and I mean considerable progress in mere months. It covers essentially everything Genspark has launched since May 2025 (the link to the first article is at the beginning of this one). See Genspark’s pricing feature list here. Genspark offers three service tiers with different usage limits with credits now.

Anyways, literally one of the best AI tools people still haven’t heard of. Put down NotebookLM and try this! While the majority of its employees are Chinese, Genspark (Mainfunc Inc.) maintains an office in Singapore as part of its global operations alongside its Palo Alto headquarters. Singapore where some of its top investors are based.

If I could combine ’s clarity and ’s curiosity about AI tools - well here you go: (bonus)

240+ AI Tools That Defined 2025

Genspark Is a Unicorn Now. Let’s Get up to Speed.

Read more

Good Morning,

Looking forward to 2026, I wanted to find a topic that might also be relevant to so many of the Creators, writers, solopreneurs and builders on this platform. Many of whom support AI Supremacy and which I do my best to give back to.

Share

This is a guide for Founders, the bravest kind, the ones who go at it alone.

I’m really proud of finding guest contributors from all around the world 🌍 who are uniquely qualified to talk about the topic they undertake on AI Supremacy. is from Lebanon 🇱🇧 and a 3-time Founder who was at Google for nine years. I know him from his viral LinkedIn posts, check out his Notes here. He’s recently launched Ruya Advisory, working with founders on building, growing and fundraising in their early days.

“Let’s Launch, Grow, Fundraise For Your Startup Together.”

Check out his Newsletter: Helping early-stage founders validate, launch, and raise with in-depth guides, frameworks and tactical strategies to future proof your startup with agentic AI and relatively recent methodologies.

Majd's Newsletter

Helping early-stage founders validate, launch, and raise. Follow for in-depth guides, frameworks, and tactical strategies that actually move the needle.

By Majd ALAILY

Get his Free Report: Try the Problem-Solution Pulse: get an 11-page personalized breakdown of your startup’s foundation. Majd is obsessed with design-led growth and systems thinking for startups. His tactical insights and battle-tested advice are drawn from real founder experiences and what’s actually working right now in a dynamic environment.

Pledge your Support

I asked Majd for his expertise on how Generative AI is changing how solopreneurs are operating and building their own companies, with insights that might be relevant also to many writers building media startups here on Substack.

In this edition: tl;dr ✂️

This article keeps the natural conversational tone of the author in full.

• Why 2026 is THE era for AI solopreneurs, and why the shift is bigger than “no-code”

• 5 forces enabling solo founders to build what once required full teams

• The best business models & categories for 1-person AI companies this year

• A practical solo-founder playbook to build, monetize, and stay sane while doing it

  

On the journey: ✍🏻

• How my biggest growth mistake got me 1M users

• Build a pitch that actually impresses investors (Based on 100+ decks)

• Insider Lesson: How a $55 validation built a $100M+ company

• The 4 Circles Behind Every Great Founder

If you run a Newsletter like a business, you might be a media founder in a manner of speaking.

Why 2026 Is the best time (ever) to become an AI solo-founder

In 2021, I left my job at Google to start my own business.

Coming from a “mega-corporation”, I wanted to stay lean.

But I couldn’t. It just wasn’t possible.

Too many moving parts. Too much friction. Too little infrastructure.

So I had to scale, staff up, build systems the “old-school” way.

Fast forward to 2026, EVERYTHING is changing.

Editor’s Note: If you have the itch to build a small business as a solo founder this article is a mini crash course on some important considerations, strategies, mindsets, growth-hacks and Generative AI implications for doing it in 2026.

✅ No-code tools are everywhere

✅ Agents can replace critical functions

Read more

Good Morning,

With Google’s Gemini 3 release they have improved their AI coding capabilities and agentic possibilities for developers and for Enterprise on the Cloud. Google Antigravity is an ‘agent-first’ coding tool and it’s clear Google is going all-in on Agentic AI.

But what are we talking about here exactly, what does it do and what is the full scope?

The Gemini 3 launch has gotten its fair share of critical acclaim not just for its benchmarks and performance but for its multimodal architecture and vision for what the future of AI is going to be like.

I asked for a guide to this elusive Antigravity agentic IDE product.

The AI Augmented Engineer

The AI-Augmented Engineer

Learn how to use AI tools to write better code, ship faster, and stay ahead of the curve

By Jeff Morhous

• What does a next-generation IDE look like?

• How to use it and what can it do?

Dive into Antigravity 🧲

Gemini 3 is SOTA on SWE Bench Verified with a standard agent harness (all models using the same independently created harness).

Building the Foundations for the Agentic Future

What’s obvious about Gemini 3 is that it has been carefully optimized also for vibe coding and as an agentic coding model. Google says that Antigravity, which supports multiple agents and gives them direct access to the editor, terminal, and browser, is designed for an “agent-first future.”

It’s clear Google is going after Cursor, Github Copilot and Claude Code with Antigravity. It does bring a few new features along with it. Antigravity will be available for free in public preview for Mac, Windows and Linux as of November 18th, 2025, with what Google calls “generous rate limits” for Gemini 3 Pro usage.

Join over 2 million developers building on Google’s AI Studio

AI Studio

How to think like an AI augmented engineer?

• How to use NotebookLM to actually learn things

• What the heck is vibe coding?

• How I vibe coded an iPhone app I now use everyday

• Building a micro-SaaS in an afternoon with AI

• Deep Research: A guide to speedrun learning anything

In a nutshell: tl;dr

Here Google is pushing how the future IDE might operate with a few subtle but powerful tweaks:

• Agentic Platform: Google Antigravity is an experimental, free AI-powered development tool that evolves IDEs into an “agent-first” experience, using AI agents for autonomous coding tasks powered by Gemini 3.

• Asynchronous Workflow: Unlike traditional IDEs focused on synchronous editing, it enables AI agents to run tasks in the background, like debugging or testing, without blocking user input.

• Multi-Model Integration: Supports Gemini 3, Claude Sonnet, and GPT-OSS models natively, offering more flexible AI assistance than single-model IDEs like Cursor.

• Browser Control: Features deep browser integration for agents to interact with web apps directly during development, a capability absent in most IDEs.

• Agent Dashboard: Includes a dedicated manager for overseeing multiple AI agents, contrasting with the editor-centric interfaces of tools like VS Code.

• Natural Language Commands: Provides tab autocompletion and voice-like code generation via natural language, streamlining workflows beyond autocomplete in standard IDEs.

“At its core, Antigravity is still a fully-featured AI IDE, but the Agent Manager provides a new mission control surface for background agents to tackle backlog tasks, perform codebase research, and reduce context switching.” Get started.

Explore Use Cases

But let’s rewind a bit in a more guide-like reading of this new tool by Google. What about for non-technical users? This is our established how-to-guide intro to Google Antigravity. It will live in our guide section. Also check out Jeff’s own guide section here. If you are impressed with his fluency and teaching style consider supporting his work:

Get Augmented

Your guide to building apps with Google Antigravity

By , Mid November, 2025.

Read more

I recently wrote the viral NotebookLM post read it here. This is supplementary material and a continuation of that. Follow me on LinkedIn and read my Newsletter: The AI Mak…

Read more

Today we continue our series on using Anthropic Claude and Claude Code, the fourth article in our series (go here to see the others), focusing on the latest agent skills. This is a guide.

Frank Andrade is The PyCoach of the Newsletter Artificial Corner.

If you know someone with an interest in getting better at Anthropic’s features and Claude, this agent skills guide may be useful for them.

Share

Context of Agent Skills

In mid October, 2025 Anthropic announced its next step toward general purpose utility of its Enterprise AI and world leading AI coding models: Skills for Claude which are extremely pragmatic and useful.

Agent Skills by Anthropic

🔗 Official sources:

• See Github

• See Agent Skills in Claude Docs

• Available Skills

• Introducing Agent Skills

• Real world applications with Agent Skills

Artificial Corner

Artificial intelligence in plain English, join thousands of AI enthusiasts to make the most out of AI.

Artificial Corner

Artificial Intelligence in plain English. Join thousands of AI enthusiasts to make the most out of AI

By The PyCoach

Frank is an AI & Python instructor who has helped thousands on YouTube and Substack master AI through beginner-friendly guides and in-depth tutorials. He’s starting a new journey on Instagram and is offering his ChatGPT course free to anyone who follows him and DMs him on Instagram.

Specialized Capabilities you can customize

📢 Agent Skills are modular capabilities that extend Claude’s functionality. They are designed to transform a general-purpose agent (Claude) into a specialized one for specific tasks.

Good Reads

• How to Use AI with Excel (Reliably): From Generating Functions to Complete Reports

• Deep Research: One of ChatGPT’s Best Features (with of Wonder Tools)

• Your saved prompts are useless now (Claude Skills guide) by

• ChatGPT Atlas Will Change Web Browsing Forever. Here’s Where It Beats Chrome

• The Genius of Anthropic’s Agent Skills

• How I Use Every Claude Code Feature by

• Claude Skills Are Taking the AI Community by Storm by

Upgrade for Full Access

This Guide Covers ⚗️ tl;dr

Images, screenshots and walk-through how-to guide & tutorial:

• Introduction to Claude Skills: Claude’s Skills feature allows users to create reusable instructions for specialized tasks, such as adhering to brand guidelines, saving time on repetitive workflows; these skills are compatible with ChatGPT and Gemini for broader use.

• Usage Example: By enabling Anthropic’s brand guidelines skill, Claude generates presentations that align with specific company styles, like Anthropic’s, compared to generic outputs without skills.

• Enabling and Applying Skills: Access Skills via Settings > Capabilities, toggle on pre-built ones, and invoke them either manually via “Try in chat” or automatically through prompts; always enable the skill-creator for best results.

• Creating Custom Skills: Prompt Claude to build a skill (e.g., for Google branding), receive a downloadable ZIP file with Markdown instructions and resources, then upload it to Claude for use in new chats.

• Cross-Platform Compatibility and Limitations: Export ZIP files to ChatGPT or Gemini for similar automation; however, Skills require a paid subscription, may yield inconsistent results, and can activate unintentionally if not managed.

Let’s get in to the deep dive:

How to Use Claude Skills to Automate Your Workflow

An amazing feature that even ChatGPT and Gemini users can benefit from

By , October, 2025.

If you’ve ever struggled to get tasks done quickly (whether for yourself or your team) using custom GPTs or projects, you should try Claude Skills.

Claude’s new Skills feature lets you generate reusable instructions to help it perform specialized tasks like following your company’s standards more reliably

Skills are good for:

• Teaching Claude how to deal with specific tasks

• Saving hours on repetitive tasks

• Creating custom workflows

The best part is that skills generated with Claude can also be used on ChatGPT and Gemini (more on that later).

Links to best pieces:

• How to Go from ChatGPT Beginner to Pro in 2025

• 25 Things I’ve Learned About ChatGPT and Other AI Tools

• GPT-5 prompting guide

• Sora 2 prompting guide

• How to Create Your First AI Agent (for Non-Coders)

Let’s see an example.

Read more

Good morning,

When we think about the future of the human interface an ambient computing comes to mind popularized by such movies as Her, in 2013. This implies a universal personal assistant in the form of agentic AI chatbots that are completely personalized to us. I believe the rise of Voice AI experienced by “advanced voice mode” of current leaders best illustrates how close we are to this point.

In June, 2025 Madrona did an interesting exploration of the Emerging AI Agent Infrastructure of Voice AI. The way we connect with the world via voice-AI is about to shift. I asked of Why Try AI, to look more into the consumer AI aspect of Voice.

Voice AI is having its moment in 2025 on the AI startup front as well. In the past few months, we’ve seen a surge of Indian startups leveraging voice AI in powerful ways, according to Lightspeed India.

“Latency: down from >1s to <300ms (human level). Anything above 1.5s is highly noticeable and considered socially awkward
Expressiveness: new STT and TTS models which can capture emotion, pitch, tone in ways that make the output indistinguishable from humans VAD & interruption handling: real-time turn-taking, background noise filtering, and multi-speaker context, all of which are important components to deliver high quality conversational experiences” - Rohil Bagga, Lightspeed India

A little over a decade ago, the movie Her introduced us to Samantha, an AI-powered operating system whose human companion falls in love with her through the sound of her voice. Twelve (almost thirteen) years later in the real world AI companionship is a real thing for millions of young people and Gen AI users.

The societal debate surrounding AI companions isn’t just about their effects on humans but how corporations entertain and retain users. With Google Gemini gaining on ChatGPT in 2025, OpenAI’s ChatGPT will soon allow ‘erotica’ for adults in major policy shift. The move could represent a major shift in OpenAI’s policy, which currently bans NSFW content even as services such as Character AI are still incredibly immersive. Current LLMs designed to be sycophantic (agreeable and manipulative) will only get and become more persuasive. As such, be cautious with how you interact with AI. (AI Psychosis is a real thing).

Advanced Voice Mode Gets Human Level in 2026

Among the seven leading AI products, they all have Voice-AI implications many with very functional voice mode (and voice search) options:

As conversational AI is embedded in more devices and institutions, we’ll be taking for granted all of this soon enough. This has profound meaning of how we will be relating with whatever the future of AI, chatbots and BigAI “Super apps” become and software that’s increasingly personalized to us.

With latency now at human levels, models that can mirror tone and emotion, and interruption handling that makes conversations natural - voice AI is moving fast from promise to reality.

A new stack of Agentic Voice AI agents is coming:

​Daniel’s Why Try AI newsletter focuses on practical guides to AI for non-techies. He demystifies generative AI for the average person via humorous, beginner-friendly takes without hype. In addition to large language models, Daniel loves experimenting with creative GenAI tools like image generation, AI video, AI music platforms, and so on. For general interest AI enthusiasts, it should be one of your go-to sources in my opinion:

Why Try AI

• Hype-free, hands-on AI for non-techies.

Why Try AI

Hype-free, hands-on AI for non-techies.

By Daniel Nest

Let’s Explore AI Tools

Her is a selection of ’s recent posts that caught my eye that you might find of value:

• 14 Niche AI Tools You Should Try

• I Tested Three Different AI “Study” Modes

• Here Are My Go-To AI Tools

• I Fed My Voice to 10 Free AI Voice Cloners. Only One Nailed It.

Trending AI Tools of 2025

Commercial voice applications are about to hit a tipping point

I believe we’ve very close to the sort of general purpose utility Voice AI that was once thought of as the holy grail of ambient computing and science fiction even 20 years ago.

Daniel’s complete guide to Voice AI follows with links to many of the most important tools, startups and products around Voice AI. The list is easy to read with dozens of the key players.

It’s more like a launching pad to further exploration than a comprehensive guide, but important if Voice AI is something you want to go deeper into. You will find the best:

• Note takers & transcription services

• Conversational AI agents

• The 8 leading chatbot voice modes

• Speech generation & voice cloning

• Voice infrastructure (APIs / SDKs), and more.

The voice AI interface is posed to make more progress in the next three years 2026-2029, than in the previous thirty years. But which are the key real world use cases in 2025 and who are the leaders today?

Check out our guides and upgrade to premium for full-access:

Subscribe now

Read more

I’ve been studying this past week Google’s incredible 2025 of AI product releases and if you feel behind, you are not alone. In the scattered bureaucracy that is Google, there is light.

There’s a lot to unpack. When they doubled down on LLMs and related products since ChatGPT, they weren’t joking. Nearly three years later, we starting to see the full force of the results. We are starting in mid to late 2025 to see the fruits of their labor. I want to take you on a quick tour where you can dive in deeper if you see something interesting. My hope is this will enable you to familiarize yourself better with what Google’s AI ecosystem is becoming and what they are up to in AI products.

Why Read this? 📚

So if you like discovering AI tools, you might enjoy this article. But I go into some of Google’s research ambitions too near the end.

This article/directory of Google AI products can help you explore everything from Google Deep Think to tools out of Google Labs you’ve never heard of. Or perhaps you have, but never actually tried. It’s all about utility as Google in 2025 has had the year of the decade in terms of AI innovation (though I hate to admit it). This post is a portal to that world.

BigAI’s Ecosystem 🌌

If ChatGPT is convenient and Anthropic’s Claude is pragmatic, Google Gemini and GDM’s various AI Products fills the gap between, and it’s quickly become a fairly wide array of products. Not all of them are that useful, but they are becoming more capable, cost-effective and gaining ground on rivals.

In 2025, the way we conceive of search, research, content generation and discovering the physical world with AI has all suddenly transformed. This momentum in the late 2020s is only going to speed up.

Rise of Gemini App 🌍

According to Google, while AI Overviews has 2 Billion users and Google Gemini now has in the area of 500 million monthly active users as of September, 2025. That’s catching up to ChatGPT fast.

Access Gemini on the Web

With all these new AI capabilities this means it’s perhaps the best time to own a Pixel phone or Google device. Google Gemini is now the 2nd most popular AI product and AI app in the world just behind ChatGPT. It once seemed like ChatGPT had an insurmountable lead as a Generative AI chatbot, but in 2025 that’s become just another myth.

AI Mode ✨

AI Mode the continuation of AI Overviews has been shipped out by default to more countries.

AI Mode is new Search mode expands what AI Overviews can do with more advanced reasoning, thinking and multimodal capabilities so you can get help with even your toughest questions. You can ask anything on your mind and get a helpful AI-powered response with the ability to go further with follow-up questions and helpful web links.

AI Mode is a powerful, conversational AI-powered search experience in Google Search that allows users to ask complex, multi-part questions and get detailed, AI-generated responses with citations to web content.

Try AI Mode

NotebookLM ⚗️

NotebookLM is a AI-powered research assistant from Google that acts as a personalized virtual research assistant, allowing users to upload documents and other sources to get summaries, answer questions, and brainstorm ideas.

Some find it useful or even entertaining and educational.

• Google recently proposed: 6 ways to use NotebookLM to master any subject

Guides:

• How to use NotebookLM for personalized knowledge synthesis

• How to Use Google's Deep Research, Learn About and NotebookLM Together

Agentic Design Patterns (Free Book) 📖

A senior AI engineer at Google just dropped a 400-page book for review: Agentic Design Patterns. If you want to learn more about AI agents this might be a good place to start.

It’s got everything about AI agents + code:

• Prompting

• MCP

• Tools

• RAG

• Memory

• Humans in the loop

• Multi-agent

• Guardrails

• You name it

You can read it here (Google Drive + links to chapters):

Read the Book

Google Labs 🔬

Google Labs is a platform where users can try and provide feedback on experimental features and products, particularly those involving artificial intelligence, to help shape their development.

• A space for experimentation:

  Google Labs hosts experimental AI products and early-stage features from across the company.

• A platform for feedback:

  It allows users to get hands-on with these new ideas and provide direct feedback to the teams developing them.

• A glimpse into the future:

  Users can explore and try out cutting-edge AI technologies and help Google develop responsible and innovative tools.

Visit Google Labs

Flow 🎬

A new AI filmmaking tool that lets you seamlessly create cinematic clips, scenes and stories with consistency using Google’s most advanced AI models, including Veo 3.

Create with Flow

Genie 3 🧞‍♀️

Genie 3 (released in August, 2025) is the latest "world model" AI from Google DeepMind that can generate interactive, physically consistent 3D virtual environments from simple text prompts. Users can navigate and interact with these dynamic, game-like worlds in real-time.

• Real-time interactive worlds: Unlike models that produce only static images or pre-rendered videos, Genie 3 creates worlds that a user can explore and move through continuously. It generates these 3D environments at 720p resolution at 24 frames per second.

• Promptable world events: Users can dynamically change the generated world by issuing new text commands while exploring it. For example, you can alter the weather, add new characters, or trigger other events on the fly.

• Persistent visual memory: Genie 3 can remember changes and object locations in the environment for up to a minute, which helps maintain visual and physical consistency. If you paint a wall and then turn back to it later, the paint remains.

• Learned physics: Rather than relying on a hard-coded physics engine like traditional video games, Genie 3 learns the rules of the world, such as how objects fall and collide, by observing massive datasets of videos.

Learn More

Career Dreamer 💼

A new experimental tool that uses AI to make it easy to explore career possibilities personalized to your skills & interests. (Available only in the U.S. right now)

Try it Now

Shop with AI Mode

This is new shopping experience is rolling out in AI Mode, where shoppers can view product visuals and other AI-powered guidance that leverages product data. This is popular for both product discovery and travel plans.

This is supposed to get us in the habit of using AI Mode.

Google AI Pro 😎

Google AI Pro is a subscription plan offered by Google as part of its Google One service, providing enhanced access to Google's advanced AI tools and models. If you enjoy this article, you might want to get it if you foresee yourself using Google’s AI products more. Start your free trial.

Get Google AI Pro

Nano Banana 🍌 (Gemini 2.5 Flash Image)

Get Full Access

Also check out more AI guides and AI tools.

Read more

If you know someone who might appreciate insights on Nano Banana, please don’t hesitate to share this guide with them.

Share

I’m of the Newsletter AI Blew My Mind:

Going Paid Launch 🚀

My Newsletter is finally going paid (as of yesterday) so support if you appreciate the work that I do. Editor: 🌎 Support global writers, the author is from Bucharest, Romania. 🇷🇴 🏰

Support ABMM

Learn how to turn AI into your unfair advantage…🔥👀

AI blew my mind

Learn how to turn AI into your unfair advantage with tested tools, practical prompts, and workflows you can apply today.

By Daria Cupareanu

🍌 Why is ‘Nano Banana’ by Google special?

If image alterations, design and visual nostalgia interests you, this guide may be helpful:

"Nano Banana" is the codename for Gemini 2.5 Flash Image, an advanced image generation and editing AI model developed by Google. It’s a major upgrade in image generation, editing and multi-modal design possibilities. It features seamless:

What’s new and different?

1. Prompt-based image editing

2. Character and style consistency

3. Image blending

4. Real-time and fast editing

5. Multimodal capabilities

Google used this launch of Nano Banana as a major campaign in early September, 2025 to net new Gemini users to boost their consumer AI adoption.

What I’m working on 🦄

• See my Prompts and Workflows

• My exploration of AI Tools

• I run an active Chat.

☕ Read my best works

I dive into the wild, fast-moving world of AI tools, prompts, and workflows so you don’t have to:

• Recursive Prompting: The Process That Makes AI Outputs (and You) Smarter

• How I Taught Myself to Work with AI in Domains I Had Zero Expertise In

• How I Use ChatGPT to Catch My Biases (Daniel Kahneman Would Approve)

• I Tested ChatGPT Agent on Real Work Tasks: How You Can Use It Too

• How to Test-Drive Any Career or Life Decision with AI

Browse my Archives

Learn more about my Newsletter launch.

Nano Banana: Everything You Need to Know About Google’s Viral AI Editor

This guide covers around 4,000 words or 22 pages in a Google doc, 25+ images and dozens of business use cases.

From Photoshop-killing features to creative use cases, glitches (and fixes), and the business ideas it unlocks

When Google launched Nano Banana on August 26, the world wasn't prepared for what came next.

Social media exploded with the cultural phenomenon. Banana emojis flooded feeds, meme challenges went viral, and debates over whether Nano Banana was the "Photoshop killer" dominated conversations from Silicon Valley to design studios worldwide.

In less than two weeks, over 10 million new users had signed up for the Gemini app, drawn by word-of-mouth buzz and viral social media posts showcasing the tool's capabilities. Users immediately started creating, and the platform processed more than 200 million images, a volume that rivals what established editing tools handle in entire years.

What we’ll explore:

• Why Nano Banana stands out: what makes it different from other AI image tools

• How to get started: access, pricing, and Google’s prompting guidelines

• Hands-on experiments: 15 creative and professional use cases (comics, outfit swaps, headshots, thumbnails, book covers, UGC, ads, landing pages, and more)

• Limitations and fixes: common problems I ran into and the workarounds that solved them

• Beyond image editing: Google’s ready-made Nano Banana app templates

• Building with AI Studio: how I vibe-coded my own app and how you can do the same

• Business potential: 10 business ideas you can build with Nano Banana

Unlock the full 4,000 word guide, deep dive with detailed use cases and visualized examples.

Subscribe now

Read more

Hey, I hope you had a good summer. I get so many questions and comments about our NotebookLM guides. So I wanted to double back today on some of their new features.

In the AI News 📰📢🚨

First some th…

Read more

How should we best think of the coolest AI tools in mid 2025? of Why Try AI and of Artificial Ignorance have a lot of experience with successful rundown Newsletters, packed with tangible insights and varied utility of applying AI.

Why Try AI

Why Try AI

Hype-free, hands-on AI for non-techies.

By Daniel Nest

Artificial Ignorance

Artificial Ignorance

At the intersection of engineering and intelligence. Essays, analysis, tactics, news, and more.

By Charlie Guo

Notable Articles

The latest from the GPT-5 rollout and recent AI News:

1. AI Roundup 131: GPT-5's rocky rollout (Charlie)

2. AI Catch-Up: GPT-5 vs. Everything Else (Daniel)

3. GPT-5: The Reverse DeepSeek Moment (Zvi)

4. OpenAI’s Waterloo? [with corrections] (Marcus)

The GPT-5 Reckoning gives momentum to Anthropic, Google and Qwen

I’m not going to lie, the OpenAI antics of the latest model launch pushes talent and consumer attention to competitors [and more funding].

AI Rundowns aren’t Dead 😵

AI rundown Newsletters are incredibly valuable to keep track of what’s going on in the least amount of time possible. Recently major beehiiv Newsletters switched over to Substack to do just that too:

Daniel’s Bio:

Daniel’s Why Try AI newsletter focuses on practical guides to AI for non-techies. He demystifies generative AI for the average person via humorous, beginner-friendly takes without hype. In addition to large language models, Daniel loves experimenting with creative GenAI tools like image generation, AI video, AI music platforms, and so on.

Charlie Bio:

Charlie Guo writes Artificial Ignorance, covering news, analysis, and deep dives at the intersection of AI and software. After graduating from Stanford in Computer Science, he founded a YCombinator-backed startup and authored the book Unscalable, featuring stories from prominent tech founders on "doing things that don't scale."

How this collab became possible:

Two of my favorite people doing AI publications: 🙏

0:41 - what’s in this deep dive?

Identifying the top AI Tools

Identifying the top AI tools means doing a rundown across interfaces capabilities including chatbots, reasoning models, research tools, coding, browser agents, image, video, music, voice and 3D Generation. This is a fairly comprehensive look into this topic, accurate for the summer of 2025.

Relevant Tool Explorations

Charlie and Daniel think a lot each week about the best AI tools for specific tasks. Daniel’s Sunday content is getting more valuable as the quality of AI News begins to diminish elsewhere on the whole:

Get Insider Details

• Here Are My Go-To AI Tools

• The Chatbot Trap

• SEO for AI: A look at Generative Engine Optimization

Charlie remains one of the most lucid and clear thinkers around AI’s utility with original and nuanced framing of what’s going on.

Don't be Artificially Ignorant

Recommended Reading

Daniel’s wit and Charlie’s agility of explanation aren’t just easy paths to AI learning, they entertain too.

• No-Prompt Prompting? So Lazy, It Just Might Work!

• LLMs are getting dumber and we have no idea why

• AI Is Simple. Just Ask the Experts.

Learn about Artificial Ignorance

• 5 Best LLMs You Can Use for Free

• The GPT Era

Learn more on Why Try AI

If you are unwilling to pay for this sweeping tour of AI in modalities of 2025, this can be a fair replacement:

My Go-to AI Tools

I’m also obsessed with how a few others cover AI in their notes:

This article distills what Charlie and Daniel have learned about AI tools for the past few years and actually offers up a few AI tools off the beaten track that are perfect for AI Enthusiasts. I think you’ll find a lot of value in it! You might even want to Bookmarket it. 👀

The Coolest AI Tools of 2025 (So Far)

Read more

Around 2.5 years after ChatGPT went live, I started to notice a new breed of AI sherpa accounts popping up on Substack that I had never seen previously that felt entirely different.

They weren’t technical or news orientated, they explored the human interface between psychology and cognition and the new interface of AI.

I urge you to try to wrap your mind around what this means at scale. There are then likely hundreds of thousands to millions of global citizens and people exploring this, in a tremendous collective experiment, but not in a lucid way:

Among this unusual and ill-defined genre the most notable examples in my humble opinion are in terms of writers here are:

• (read his Notes)

• (from the popular X account: @signull )

Read Notes of Wyndo

Mind Melders 🛕

• To differentiate them from the AI sherpas that I talk about (typically AI tool guides makers), I coin these the Mind Melders of the new AI Interface. So what am I even talking about? 🧐

• These writers teach you how to think differently in the new era of Generative AI for maximum gains.

• They point to a cognitive shift that takes place when using AI tools natively.

The insights of those like Signull and Wyndo are an entirely new niche of learning how to be AI-native from the inside and not just via using empirically objective context engineering techniques. They deal with self-awareness, leverage, agency and augmentation. They fascinate as case studies of a new psychology of being with AI.

🗝️ A New genre of AI Educators

• Self-awareness at the AI interface

• Leverage

• Agency

• Augmentation

• (Embodied) Problem Solving [SLAAP, the acronym]

I asked about how can we use Generative AI to overcome our “blind spots” and augment our self-awareness. I’m hoping this is a good teaser to his publication. I’ll have you know he’s a fairly engaging and viral writer.

🔭 The AI Maker 🏝️

Wyndo runs this newsletter: The AI Maker, a publication dedicated to helping knowledge workers implement AI effectively without technical expertise. Through his hands-on experience building AI-powered tools, Wyndo helps readers develop AI-augmented workflows that enhance their capabilities rather than simply automating tasks, making his newsletter a trusted resource for professionals navigating the rapidly evolving AI landscape.

1. Implement AI effectively without technical expertise

2. Develop AI-augmented workflows that that enhance capabilities

The AI Maker

Making AI accessible for everyday life. Simple strategies to build smarter, work faster, and live better.

By Wyndo

• What became clear to me is it’s not about what they are building, but how they are self-engineering their consciousness augmented by AI. (An AI-Yoga of symbiosis if you will)

• Maybe you will recall in the Dune novels, “Mentats” are people who are highly skilled in logic, computing, and cognitive thinking. In the Dune universe, instead of creating computers or machines to solve logic issues, society trains selected humans to be Mentats.

• My question to all of you is what the new “Mentats” of Generative AI will look like both professionally and personally in 5-10 years? Will they just be context engineers skilled at AI agent orchestrators or will it also come with a kind of meta-cognitive and philosophical set-up? What I’m talking about isn’t just another kind of AI literacy or work-flow famility.

Who are the Mentats of Generative AI? ✨

About

Access to the internet is no longer the primary commodity of the digital age, now it’s usage and fluency with AI. It’s only normal some are ahead of the curve.

New mindsets and baseline expectations of what we can do with AI are emerging. In 2025 and after this date, AI is a core attribute of life-management and cognition for some folk, not just an accessory of tools or time-saving techniques. What becomes possible for you when you internalize and believe in AI through daily use?

It’s hard for me to convey how radical of a change this happens to be. It’s more than simply functional, a literacy skill, academic or an industrial output of work, it’s how AI changes you after prolonged and in-depth use. Some of these first “AI-natives” appear to interact with technology and the world differently due to their constant proximity to AI interfaces.

An AI Maker’s Articles

1. How I Used James Clear's Atomic Habits to Build 15 AI Systems That Run My Life

2. How I Finally Turned AI Into My Personal Operating System for Work

3. How I Learned Complex Topics 10x Faster with NotebookLM

4. My AI Therapy Workflow: Turn Claude/ChatGPT and NotebookLM Into Your Self-Discovery Tool

5. Forget Prompting Techniques: How to Make AI Your Thinking Partner

Learn to Build with AI

While I personally have no inclination to “meld into the machine”, I think these frames of thinking and ways of interacting with Generative AI have a lot of potential value:

Why Smart People Make Predictably Bad Decisions (And How AI Can Help You See What You Can't See)

Share

• This is the second guest post by to read the first one go here.

Get Secret Access

• How to use AI to overcome the expertise paradox and cognitive biases.

• How to use AI to fight human biases and overcome blind spots.

• How to audit workflows with strategic AI use.

• How to become more resilient and or adaptable with the help of AI tools at important crossroads and with decision-science.

Let’s take a step through the Wyndo of AI ⊞ into the inner chambers of the AI interface. Implementing AI workflows in tasks and making crucial life decisions in with AI in life, is also about a mindset.

⚫ Join the AI Vanguard Society, a place for deeper discussions, polls and interactions about AI, an (Alpha) community by AI Supremacy. Hosted on Mighty Networks, download the app (Android).

Read more