Get instant access

Dear subscribers,

With the launch of Fable 5.2, Anthropic now has the two best models in the world (Opus 5.5 is the other one).

I've spent 500+ hours using both models in Claude Code. In this video and post, I’ll show you exactly how I use each across 5 real use cases.

Watch the tutorial now or read the written guide below. This post is sponsored by Anthropic, but opinions are my own.

I cover:

• (00:00) Fable 5.2 vs Opus 5.5: When to use each model

• (02:32) My Claude Code setup revealed

• (03:30) Set up a chief of staff for your inbox and calendar (Opus)

• (06:11) Create an advisor for your career and life (Fable)

• (09:14) Automate a complete workflow end-to-end (Opus)

• (14:01) Design and build a mobile app with Claude Design (Opus)

• (18:47) Help a loved one through a long-term illness (Fable)

• (20:54) The secret prompt that improves your AI skills over time

I’m proud to partner with Riverside

Riverside just launched four AI agents to help you create videos end to end. You can direct them to prep your recording, edit the episode, cut clips, and handle publishing to social, all in one place. Each agent specializes in one job, and the team learns how you like things done and only gets better over time. Try all four agents now and use my exclusive code PETERYANG to get one month free.

Get 1 Month Free

When to use Fable 5.2 vs Opus 5.5

Here’s the TL;DR from Anthropic’s Fable 5.2 post:

• Fable is smarter than Opus on most tasks, including coding, long-running agents, and more. I used it to build a 3D kart racer you can play for free.

• Fable costs 2x as much per token as Opus. On the Max plan, you can use Fable for up to half of your weekly limit.

Even though Fable is better, I don’t recommend using it for everything because Opus 5.5 is still an incredible model. Instead, my advice is to:

Use Opus as your daily driver and Fable for your hardest work.

Alex (AI researcher) at Anthropic shared some great tips for using Fable:

1. Ask Fable what’s wrong with your idea before it builds anything.

2. Use Fable on medium effort, which is what Anthropic employees do too.

3. Tell Fable upfront what needs your approval and what it can do on its own.

4. Give it a whole project instead of a single next step, because Fable can think through the steps on its own.

Personally, I use Opus for managing my emails, automating workflows, and building personal apps. I use Fable for getting strategic advice on my business, caring for my mom, and auditing all my AI skills for bugs or issues.

In the rest of this post, I’ll walk through each use case with exact prompts you can copy to set them up yourself:

1. Set up a chief of staff for your inbox and calendar

Read more

Dear subscribers,

Today, I want to share a new episode with Ashley.

Ashley is the VP of Product at OpenAI for health and this is a very special episode for me. We talked about how we both use AI to help our parents fight cancer — including preparing for doctor visits, researching treatments, and coping with our emotions as caregivers. Both of Ashley’s parents got diagnosed with cancer in a single week and it was really inspiring to hear how she took care of them.

Ashley and I talked about:

• (00:00) How her dad’s cancer led Ashley to OpenAI

• (07:15) Why I talk to AI more than my own mom

• (09:41) Are doctors annoyed when patients bring AI research?

• (14:17) Is AI really about to cure cancer?

• (19:12) Should you trust ChatGPT with your health records?

• (25:38) How ChatGPT caught a risky treatment before it was given

• (28:00) Why hospitals still use faxes and CDs

• (36:16) What Ashley types into ChatGPT during her dad’s appointments

• (41:09) Making time for your parents with a demanding job

I’m proud to partner with Granola

I use Granola to record my mom’s medical conversations, with her care team’s permission. Having notes we can revisit later helps us keep track of what was said, the decisions we made, and what to follow up on.

Try Granola for Free

Top takeaways I learned from this episode

1. “Both of my parents got diagnosed with cancer within a week of each other.” Ashley took a month off from work to take care of them after and doesn’t regret it. She also emphasized taking time off for yourself as the caregiver: “Even with AI, you have to make sure that you’re budgeting time for your own mental health.”

2. “Are you deluding yourself that you’re actually that busy?” Ashley had an always-on life until her parents got sick. Now she blocks time for her dad’s appointments, even as a busy executive at OpenAI. “I’m lucky that not once have I ever gotten pushback at work about this.” When she started making time for her parents, she realized, “Maybe some of those work meetings didn’t need me there.”

3. Share how you’re dealing with cancer and people will come out to help. For example, a friend’s father-in-law helped Ashley find the Stanford specialist who became her dad’s doctor. “You don’t know [who can help] until you share,” she said. I’ve found the same with my mom: Sharing her situation has helped me connect with oncologists and experts.

4. “I didn’t feel like I was in the driver’s seat of what was most important to me, which is my father’s life.” Ashley was weighing surgery and radiation for her dad, but his lung and heart specialists weren’t set up to share information and she had to piece it together herself. Give AI the patient’s full medical history and use it to have informed discussions with their medical team.

5. Give ChatGPT your medical records to flag possible conflicts before treatment. Ashley recounted how her former boss, Fidji Simo, caught a conflict using ChatGPT right before treatment. Her doctor hadn’t seen the relevant history elsewhere in her record. Ashley has seen multiple cases of this happening.

6. Do your homework before the appointment. Ashley uses ChatGPT to new therapies for his lung disease before doct appointments. She could then ask his doctor about the therapies to get his opinion on whether they’re viable options.

7. “What question should I be asking right now?” Ashley keeps ChatGPT open during her dad’s telehealth appointments. When the doctor mentioned “left wedge resection,” she looked up the unfamiliar lung-surgery term on the spot. “I’d sort of be educating myself, and then I could make better use of the time.”

8. Give your parent’s records their own space in ChatGPT. Ashley uses project-only memory to keep her dad’s records separate from her own health conversations. She also gets her dad’s consent before uploading records and redacts details he wants kept private.

9. Keep one care summary you can pull up under pressure. My /fuck-cancer skill keeps a one-page Google Doc on my mom’s care and Ashley described how she has a similar doc that keeps track of her parents’ latest medications, lab results, appointment notes, and more to discuss with their care team.

10. Have ChatGPT check for new research every week. Set up a recurring task in ChatGPT for your condition and ask it to search ClinicalTrials.gov and other research sources for new trials and treatments. My /fuck-cancer skill does this for free out of the box.

Watch the episode now and subscribe if you enjoyed the interview.

Get instant access

Dear subscribers,

Since I shared that I’m helping my mom through cancer, hundreds of you have replied with stories about your parents, partners, and friends. The sad reality is that:

Two in five people in the US will face cancer in their lifetime. When you include caregivers, almost no family is left untouched by this disease.

Watch my personal video about how I’ve been PMing the shit out of my mom’s care with AI’s help.

Timestamps:

• (00:00) My mom's story and how I've been helping

• (02:04) Step 1: Create a brief with my free AI skill

• (05:13) Step 2: Organize medical records

• (07:41) Step 3: Research treatment options

• (14:19) Step 4: Find experts to have in your corner

• (15:41) Step 5: Protect quality time with family

I’m proud to partner with Wispr Flow

Wispr Flow is the most accurate AI voice dictation tool I’ve tried, and now it supports meeting notes too. Wispr’s new speech model, Canto, also has best-in-class performance for dictation in real-world conditions. Get 1 month free with my link.

Get 1 Month Free

1. Create a one-page brief

When my mom’s cancer came back, I thought I’d mostly use AI to research her disease. Instead, it’s been far more useful for staying on top of paperwork and follow-ups across different doctors and hospital departments.

That’s why I built a /fuck-cancer skill that anyone can use to create a concise, fact-based brief with 5 sections:

1. Patient information: Essential details and medical contacts for quick reference.

2. What to do next: The top 3 actions to take or questions to ask your doctor.

3. What we know: Confirmed facts about the diagnosis and treatment.

4. Medical terms: Plain-English explanations of medical terms.

5. Care log: A compact, dated history of updates.

To install it, just paste this into ChatGPT or Claude:

Install this skill at https://github.com/petergyang/fuck-cancer

The first time you run the skill, it’ll ask who the brief is for and what’s happening before creating the brief as a Google Doc. Here’s a quick 40-second clip walking through how the skill works:

Having AI update my mom’s brief before every appointment has been critical in helping us ask her doctors the right questions.

As one reader wrote to me:

“You have to be a huge patient advocate. Push for answers. Push for biopsies. Push for proper testing. Ask questions.”

If you find value from my /fuck-cancer skill, please ⭐ the GitHub repo so more people can find it.

2. Organize medical records

Your brief is only as good as the data behind it, so start requesting medical records early to give AI a complete picture.

My mom has unfortunately had cancer three times before, so her history was scattered across different hospitals and portals. I wish we’d requested all these records right away instead of waiting:

1. Written records like scan reports, pathology reports, and treatment history.

2. Imaging files like CT, PET, and other scans, usually delivered in a format called DICOM. These are separate from the written records, and hospitals can take 2 or 3 weeks to hand them over on a usb drive or even a CD-ROM.

3. Pathology blocks or slides of the actual tumor. These are the most precious material for sequencing the tumor and second opinions (more on those later), and they take the longest to get. If the patient had cancer before, request them as early as possible. My mom is still waiting on a pathology block from her prior cancer.

Pointing ChatGPT at my mom’s brief and records has been critical in grounding its answers on her actual history instead of generic medical advice.

3. Research treatment options

My mom’s diagnosis followed a typical path:

1. Her doctors saw a suspicious tumor in a CT scan and did more follow-up scans.

2. A biopsy then confirmed it was cancerous.

3. Her doctors discussed her case and treatment in a tumor board review (a Zoom call with the oncologist, radiologist, surgeon, and other specialists).

4. Treatment usually starts with standard care based on NCCN guidelines in the US, which is typically some combination of chemotherapy, radiation, surgery, or immunotherapy.

It’s been 3 months since my mom’s first CT scan revealed a tumor, and she’s still waiting to find out whether this is a breast cancer recurrence or something else.

I have a lot of respect for doctors and have many friends who are physicians and nurses. But as an active caregiver, working with the medical team has sometimes felt like working with a coworker who doesn’t respond to emails or calls and only communicates through voicemails and faxes:

It’s easy for patients and caregivers to feel helpless during this whole process, but there are steps you can take during the long wait:

a) Understand standard care and get a second opinion

In the US, doctors typically follow NCCN guidelines for first-line cancer treatment, and you can read them with a free account. Alternatively, once you use my skill to create a brief, you can just ask it in ChatGPT or Claude:

For complex cases like my mom’s, it’s also worth asking for a second opinion at a major cancer center like MD Anderson (Texas), MSK (New York), Dana-Farber (Boston), or Stanford and UCSF (Bay Area).

We’re currently asking UCSF to provide a second opinion (my mom’s being treated in Canada), and having her records and scans organized from step 2 has made this process much easier.

It still takes a month or so to get a second opinion, so start early.

b) Sequence the tumor

Cancer is actually more than 100 different diseases, and each person’s tumor can behave differently.

That’s why we’re working with BostonGene to sequence my mom’s tumor and look for DNA, RNA, and other markers that could point to better treatments. If there’s one thing I learned in this whole process, it’s that:

Sequencing requires the pathology block or slides, takes at least a few weeks, and will cost us around $4,000, so ask about it before the biopsy if you can.

c) Research clinical trials

Clinical trials can open up new treatment options, sometimes even before standard care stops working. To find relevant trials, you can just ask my skill:

If you’re using ChatGPT or Claude, you can also ask it to rerun this search every week or month to keep you updated.

This doesn’t mean that you should blindly try to enroll in every trial. Some early phase I or II trials are unproven and could actually make things worse.

But it’s good to understand the options and eligibility early so you can discuss them with the care team at the right moment.

4. Find experts to have in your corner

One of the best things to come from sharing my mom’s story is the people who’ve offered to help. I’m especially grateful to:

1. Ari from OpenCancer, who made introductions to people who could help us navigate testing and pathology review.

2. Bayli from the Sijbrandij Foundation, who helped us prepare questions about the diagnosis, tissue testing, and treatment options.

3. Dr. Odabashian, Dr. Bernstein, and other oncologists who generously offered to provide their perspectives on my mom’s case.

To keep them in the loop, I work with ChatGPT to draft short updates to email my board of experts after every major appointment or finding.

Here’s an example update:

Many families facing cancer keep it to themselves. I’ve found that sharing the journey online or even just with friends empowers other people to help.

5. Protect quality time with family

When I found out my mom’s cancer was back, I took a week to stay with her. We had many walks, meals, and chats that I’ll cherish. I also call my parents every day so they can talk to my girls.

If you’re in your early 40s like me, you probably have young kids and aging parents too. This is also when it’s easiest to get caught up chasing more money and titles for your career. I’ve found that:

You know when your next performance review is, and you can measure your salary and follower count. It’s harder to see what you lose when work takes another night from your kids, or when you keep putting off a trip with your parents.

If you already earn enough to support your family, ask how much more you really need and what you’re trading for it. This comic from Dan Dougherty is my constant reminder that time with our loved ones is precious:

My girls are somewhere in those early panels, and it’s only a matter of time before they grow up. As for my mom, her care is still my #1 priority, and I won’t stop until she gets the right treatment to beat this evil disease.

You can get my free /fuck-cancer skill here. Let me know in the comments if you have any questions.

Dear subscribers,

Today, I want to share a new episode with Jaclyn.

Jaclyn is a product director at Google Labs and we had a great chat about how Gemini 4 changed her product job, how she keeps AI slop out of her team’s docs and products, and how she’s raising three kids with AI’s help. Gemini 4 Carbon is currently the best model in the world based on benchmarks, and it was fun to get a sneak peek into how Google’s own product leaders use it internally.

Jaclyn and I talked about:

• (00:00) How Gemini 4 changed the PM job at Google

• (02:11) Should it even exist? How Google Labs picks what to build

• (09:16) The doc rule that keeps AI slop out of Google Labs

• (10:45) Which Labs experiments graduated into the Gemini app

• (12:53) Why Google Labs never treats a killed experiment as failure

• (17:43) Code is free, so what’s actually worth building?

• (24:00) Her 5-year-old is the creative director of their AI podcast

• (26:30) Best screen time for kids: Creation, not consumption

I’m proud to partner with Wispr Flow

Wispr Flow is the most accurate AI voice dictation tool I’ve tried, and now it supports meeting notes too. Wispr’s new speech model, Canto, also has best-in-class performance for dictation in real-world conditions. Get 1 month free with my link.

Get 1 Month Free

Top takeaways I learned from this episode

1. “It shouldn’t take someone longer to read your doc than it took for you to write it.” Writing is thinking and Jaclyn believes that outsourcing writing to AI is a big mistake. AI can be a thought partner, but you should firmly be in the driver’s seat. AI-written docs are a waste of everyone’s time and Jaclyn expects her PMs to be able to defend every statement themselves.

2. “It’s no longer about can you build it, it’s about should it even exist?” When code is free, deciding what to build is the job. You can rebuild many products if you “throw a couple of agents at it over a weekend” but that doesn’t mean you should. Instead, aim higher to build products, not features, that make a real difference in customers’ lives.

Read more

Get instant access

Dear subscribers,

The internet is being hacked right now, and I want to help you protect yourself.

People’s bank accounts, health records, and API keys were all targets last week because AI is finding security holes faster than anyone can patch them.

Watch my tutorial to learn how to work with AI to secure your accounts in 6 steps.

Timestamps:

• (00:00) 12 security fires in one week and 6 steps to take now

• (03:16) One AI prompt that audits all 6 for you

• (05:11) Step 1: Disconnect apps you don't use

• (06:55) Step 2: Use a password manager and check Watchtower

• (09:57) Step 3: Turn on two-factor with an authenticator app

• (10:48) Step 4: Rotate or retire your API keys

• (11:50) Step 5: Ask AI to scan and update your software

• (12:51) Step 6: Don't trust breach emails or calls

I’m proud to partner with Granola

The world’s best teams, including Ramp, Linear, Replit, Brex, and more, use Granola to take notes and focus more in meetings. I’ve tried many meeting apps myself but keep coming back to Granola. Get 1 month free via my link below.

Try Granola for Free

Why the internet is being hacked

Just this past week, we’ve seen critical security issues in:

• PHP: The NSA told organizations to patch critical bugs or take systems offline. PHP is used by millions of websites on the internet.

• WordPress: Attackers used stolen keys to plant backdoors on websites.

• Quest Diagnostics: Hackers accessed patient records.

• Banks: Unauthorized AI agents were found inside customer accounts, and JPMorgan, Bank of America, and Wells Fargo are now restricting them.

Jason Clinton, Anthropic’s deputy chief security officer, tweeted:

This is all happening because AI can now find security vulnerabilities in legacy software that human audits missed for years. So it’s an arms race between attackers looking to exploit these issues and defenders trying to patch them.

Personally, what prompted me to act was receiving this critical security alert from Google that said a Linux machine had deleted some emails. The problem is, every agent — from Grok Bot to Muse to ChatGPT Dots — uses Linux machines!

Enough is enough 🫠

6 steps to secure your accounts right now

I spent two hours securing all of my accounts after receiving the alert above. Here’s my simple 6-step security checklist that you can follow to do the same:

1. Disconnect apps you don't use. Go to myaccount.google.com/connections and remove any apps that shouldn’t have access to your Gmail and Google data. I had 47 apps connected (!) so it would have taken forever to manually remove each one. Instead, I asked ChatGPT and Claude, “On myaccount.google.com/connections, review my connected apps and give me a list of apps you think I should disconnect.” Then I reviewed the list and had AI disconnect them for me.

2. Use a password manager with unique passwords. I was using just a few passwords for all my accounts, which is crazy in hindsight. So I finally decided to stop being cheap and pay ~$48/year for 1Password to set up unique passwords for each website. It has a handy Watchtower page (below) that told me which websites had reused or weak passwords:

3. Turn on two-factor with an authenticator app. Text message codes aren't safe because attackers can trick your phone carrier into moving your number to their SIM card. Instead, use an app like Google Authenticator and turn on two-factor authentication for your accounts.

4. Rotate or retire your API keys. If you've vibe coded a bunch of apps like me, ask ChatGPT or Claude: “Audit all my projects for any leaked keys and help me rotate my active keys and archive ones that I no longer use.” AI found two keys in Anthropic Console that were still active for apps I stopped using.

5. Ask AI to scan and update your software. Ask ChatGPT or Claude: “Check my computer and apps for any updates. Run a dependency audit on all my projects. Don't install anything, just list everything that you found.” Vendors are shipping security fixes fast right now, so it's worth running this every few days.

6. Don't trust “breach” emails or calls. You should expect a wave of phishing emails that claim a service got hacked and ask you to act now. Check the sender's address and right-click the button and choose “Copy link address” to see where it actually goes. I basically try to avoid clicking any links in supposedly “urgent” emails.

Get your personal security audit with my AI prompt

I turned all 6 steps above into one prompt that you can paste into ChatGPT, Claude, or your favorite AI agent to walk you through everything and even build a personal security audit report like this:

It's read-only by default, so it won't change anything without your approval. Upgrade now to get the full prompt below, plus access on behindthecraft.com along with 16+ other AI skills, my Personal AI course, and more.

Upgrade to get instant access

Here's the full prompt for you to copy and paste into your favorite AI agent:

Read more

Dear subscribers,

Today, I want to share a new episode with Jake Moran.

Jake is my go-to expert for creating and editing videos with AI. In our episode, he demoed how he prompts Claude and HyperFrames to make incredible videos for product launches, long-form tutorials, and shorts. We also covered the key terms that you need to learn to guide AI to make the right edits.

Jake and I talked about:

• (00:00) Video editing 101 to get Claude to make what you want

• (06:14) Parakeet: A free local model for accurate cuts

• (08:40) Hook viewers with zooms, B-roll, and whip pans

• (13:37) Demo: A launch video built from three screenshots

• (18:41) Launch videos in 4 parts: Hook, context, payoff, brand

• (20:39) Claude added Jake’s AI avatar without being asked

• (23:28) Jake’s AI edit of my raw YouTube video

• (29:25) How to copy any video’s style using AI

• (36:17) Draw feedback on the frame instead of typing

I’m proud to partner with Wispr Flow.

Wispr Flow is the most accurate AI voice dictation tool I’ve tried, and now it supports meeting notes too. Wispr’s new speech model, Canto, also has best-in-class performance for dictation in real-world conditions. Get 1 month free with my link.

Try Wispr Flow for Free

Top takeaways I learned from this episode

First, watch this 43 second video that Jake made with AI in our episode:

Here are my top takeaways on exactly how he made this:

Read more

Today, I’m excited to launch /video-short, my new AI skill that lets anyone turn a raw recording into a compelling short video. Best of all, it’s 100% free.

/video-short can add B-roll, motion graphics, captions, music, and more using your favorite AI agents. It even comes with a review page to let you give feedback to AI on any frame until the short feels just right.

Watch my tutorial now to see exactly how the skill works and get the skill on GitHub.

Get /video-short for free on GitHub

If you find the skill useful, please star the GitHub repo so more people can find it. You can also upgrade to paid to get 16+ other quality AI skills, $600+ in Grok Bot and other discounts, my personal AI course, and an invite to my private Slack community.

Below’s an explainer on what the skill can do and how you can install it for free.

I’m proud to partner with Google Gemini

I use Gemini 4 Argon in my video short tutorial because I think it’s the best multimodal model with great APIs for image, video, and music generation.

Get a Gemini API key in AI Studio and try it with my skill to create your next short.

Try Gemini 4 Argon in AI Studio

What /video-short can do

/video-short makes it easy for you to iterate with AI to create compelling shorts. The best part of the skill is it gives you a review page to give feedback to your AI agent on any frame until the short feels right:

The skill supports making shorts with 4 different layouts:

1. A-roll: You, full screen.

2. Punch: A tighter shot of you for emphasis.

3. Split: A visual on top, you below.

4. B-roll: A visual full screen while you keep talking.

It can also add the following automatically:

• Captions that match what you say, word for word.

• Motion graphics written in HTML to bring the short to life.

• B-roll footage found online that’s relevant to what you’re saying.

• AI video and music (optional) generated with Gemini Omni Flash and Lyria. This step needs a Gemini API key with billing enabled.

How to install /video-short

To install the skill, just paste this into Claude Code, ChatGPT Desktop, Antigravity, or your favorite AI agent:

The first time you run the skill by typing “/video-short”, it’ll check if you have ffmpeg (to cut and encode video), yt-dlp (to pull B-roll), Parakeet (to transcribe the video), HyperFrames (to render the video), plus Python and Node.js.

If you don’t have the above, it’ll offer to install what’s missing. Plan for around 1 GB of disk space, mostly for the Parakeet speech model (smaller models aren’t as accurate).

How to use /video-short

I’ve made this skill as simple as possible to use, just follow these 4 steps:

1. Install. Paste the install prompt above into your AI agent. You only do this once.

2. Upload. Record yourself talking, then run the skill and point it at your raw video: /video-short edit my video at [path/to/video.mp4]. It’ll transcribe the recording and cut retakes, false starts, and dead air.

3. Plan. It researches the topic, finds B-roll from the web, and shows you a scene-by-scene plan in chat: what’s on screen, the words spoken, and anything it wants to generate. Reply “go” or ask for changes.

4. Iterate. Once your agent is done with a draft edit, it’ll open the review page for you to play the short, pause on any frame, and leave feedback. The agent will then apply your feedback before reloading the review page with a new version.

Here’s an example short that I created using the skill:

Get /video-short for free on GitHub and star the repo if you find it useful!

Dear subscribers,

Today, I want to share a new episode with Anthony Pompliano.

Pomp and I worked together at Facebook a decade ago before he became an investor, best-selling author, and public company CEO with an audience of 3.3M. We had a great chat about how to build an extraordinary career and life and how he’s building Silvia, an AI CFO for independent investors, to take on the AI labs.

Pomp and I talked about:

• (00:00) Why Pomp never had a career plan

• (04:43) Why an audience is the best way to win

• (10:33) Paid work trials instead of job interviews

• (18:35) Hire for revenue, not margins

• (20:32) Silvia: An AI CFO for investors

• (25:00) How AI found a great tax strategy

• (32:05) How a team of under 10 plans to compete with the AI labs

• (35:58) Pomp’s latest take on Bitcoin

• (39:21) How Pomp teaches his four kids to solve problems

I’m proud to partner with Riverside

Riverside’s new MCP lets me prep episodes, edit recordings, create clips, and schedule posts to multiple video platforms, all from ChatGPT and Claude. You can also use it to host webinars, create brand videos, and much more.

Try Riverside for 1 Month Free

Top takeaways I learned from this episode

Pomp’s career playbook

1. Here’s a brief history of Pomp’s fascinating career. He never had a 10-year plan. Instead, each move solved his biggest problem and set up the next one:

   • Learn the skill you’re missing: His first two startups failed to scale, so he joined Facebook to learn growth.

   • Get an audience first to get deal flow: He wanted to invest but didn’t know the next generation of founders, so he built a 2.4M following on X by going deep on Bitcoin.

   • Show up every day: In May 2018, he started a daily newsletter that began as crypto news. He still writes it himself every morning, and it now has 270,000+ subscribers.

   • Start small to test fit: He and his partner started with a $3.5M fund and took no management fee. Pomp has since invested in 300+ private companies.

   • Solve your own problem: Pomp was tired of updating his personal finance spreadsheet by hand every week. As AI took off, his team turned it into Silvia, an AI CFO for independent investors.

Read more

Get instant access

Dear subscribers,

It’s easier than ever to build software, but harder than ever to make money from it.

When I left my job, I thought I’d vibe code a SaaS that made money from users while I slept. Instead, I built open source tools like No AI Slop and Human Review for agents.

Now, many builders are saying that “Software is dead.” (Pieter Levels) and “Pretty soon all software will be de facto open-source.” (Amjad, CEO Replit).

So what the hell is going on? In this post, I’ll steelman both sides and give you my take on why I think:

1. It’s harder to make money from traditional software.

2. Money is shifting to software that’s built for agents first.

I’m proud to partner with Riverside

Riverside’s new MCP lets me prep episodes, edit recordings, create clips, and schedule posts to multiple video platforms, all from ChatGPT and Claude. You can also use it to host webinars, create brand videos, and much more.

Try Riverside for 1 Month Free

Why the “software is dead” debate is raging again

Here’s what set off the “software is dead” debate again:

1. Expensive software is being cloned and open sourced. A small team used Opus to rebuild seven Adobe apps, including PhotoCraft, a Photoshop clone that hit 45K+ GitHub stars. The same team is now cloning Word, Excel, and PowerPoint.

2. Classic games are being decompiled and remixed. Modders are using AI to mash together classic games like Mario in Halo and a free browser version of Super Smash Bros 64 that has 1,000+ characters.

These developments have led people like Naval to tweet that “models are the last moat in software,” and Chamath to say that software IP has been “rendered worthless.”

Let’s explore the argument from both sides:

4 reasons why traditional software is in trouble

1. AI can now reverse engineer many apps and games

If you can download it, AI can decompile it. REA, a toolkit built specifically to reverse engineer any software, is currently the #1 trending repo on GitHub, and people are decompiling all kinds of things with it.

For example, here are some classic games that you can now play for free on your browser thanks to AI (at least until the lawyers get involved):

• Unreal Tournament

• Call of Duty Black Ops Zombies

• Diablo

2. Builders are replacing SaaS with personal apps

Let’s face it, many traditional SaaS apps have become incredibly bloated.

For example, I was paying $200 / year for 1of10, a YouTube research tool. The product kept adding features I didn’t need, so I asked Claude to build a replacement. It took just 5 minutes to build the feature that I actually wanted: a Chrome extension that showed me which videos performed well on any YouTube channel.

Many SaaS apps have fallen into the trap of catering only to power users and adding features that most people won’t use. This happened to Adobe and I’d argue it’s happening to ChatGPT too.

As another example, instead of paying a monthly subscription for a bloated fitness app, I just built my own with AI:

I can add or remove features to this app whenever I want. It’s personalized just for me, so I’ll probably never pay for another fitness app again.

The counter-argument here is that most people won’t want to build their own software even with AI’s help. But I believe that building with agents won’t feel like coding at all. Instead, building will happen naturally in your chats with AI. For example, you might ask Claude what to cook tonight and it’ll build you a meal planner with recipes for the whole week.

3. Enterprises are replacing SaaS with internal apps

The personal app trend isn’t limited to individuals. Enterprises are also replacing SaaS with their own internal tools and agents. For example:

• Retool found that 35% of enterprises have replaced SaaS tools with internal apps.

• Ramp built its own coding agent, Inspect, which now writes 75% of merged PRs.

• Curative canceled a $600K/year Salesforce contract after building an internal CRM.

In enterprise, the biggest competitors to your SaaS might just be internal apps from your customers themselves.

4. Agents are replacing humans as software’s main users

If your product is still charging a subscription for humans to click through screens without proper API or MCP support, you’re probably in trouble.

For example, I’ve been using Blender to make beautiful 3D assets like this video of Himeji Castle during cherry blossoms. But I have no idea how to use Blender’s UI:

Instead, I rely entirely on Claude to use it through Blender MCP. As I wrote before, if a piece of software doesn’t have an API or MCP that my agent can use, I probably just won’t use it period.

That’s why I think a good chunk of traditional SaaS won’t make it to the agentic era.

Ok so to recap, I believe that it’s harder than ever to make money from traditional software because:

1. AI can now reverse engineer many apps and games

2. Builders are replacing SaaS with personal apps

3. Enterprises are replacing SaaS with internal apps

4. Agents are replacing humans as software’s main users

Thankfully, none of this means that “software is dead.” Here’s where I think the money is moving to in software instead:

Read more

Dear subscribers,

Today, I want to share a new episode with Micky.

Micky is a full-stack engineer who has taught 100K+ people how to build with AI. In our episode, he walked me through his 4-step agentic engineering workflow to avoid pumping out slop: Isolate every task, build to a clear structure, prove via validation loops, and ship after agents review your PRs or escalate to you for a manual look.

Micky and I talked about:

• (00:00) Why most software factories pump out slop

• (02:06) Step 1 (Isolate): How to run multiple agents at once

• (07:18) Step 2 (Build): Give agents a structure to build on

• (10:44) Step 3 (Prove): Build agentic validation loops

• (16:48) The one step Micky never hands to agents

• (21:30) Step 4 (Ship): Getting agents to review PRs reliably

• (28:20) Why Micky doesn’t write specs

• (33:08) Comparing Astra and Fable for coding

• (43:08) Are AI skills obsolete with today’s models?

I’m proud to partner with Wispr Flow

There are many AI voice dictation tools out there, but Wispr Flow has been the most reliable one I’ve used. My kids also love using it to build games and apps with Codex and Claude Code. Get a month free with my link below.

Try Wispr Flow for Free

Top takeaways I learned from this episode

Micky’s 4-step agentic engineering workflow

1. Use Micky’s 4-step workflow that makes agents prove their work before you merge. Micky’s “software factory” is his AGENTS.md file plus a handful of skills:

   • Isolate. The new-feature skill gives each agent its own copy of the code on a separate worktree, so that you can run multiple agents in parallel.

   • Build. The code-structure skill writes code based on service-layer architecture with explicit inputs and structured returns.

   • Prove. The evidence-driven-testing skill has the agent screen-record itself testing the app to compare the before and after states for a change.

   • Ship. The greploop skill uses Greptile, an AI reviewer, to test and improve the PR until it scores 5/5 with zero unresolved comments. Micky then usually reviews the PR himself before shipping.

   • Everything Micky ships to real users moves through these four steps.

How Micky makes agents earn his trust

Read more

Get instant access

Here’s the recording and key takeaways from our recent live workshop. Paid subscribers can find all workshop recordings here.

Start with real failures, not generic scores

Read more

Dear subscribers,

Today, I want to share a new episode with Sam.

Sam is the co-founder of Granola, the AI meeting notes app now valued at $1.5B. Granola has helped me survive (and even enjoy) countless meetings, so I was excited to learn how Sam uses AI to show up prepared. We also talked about building an AI company without the 996 grind and what design means when agents skip your app.

Sam and I talked about:

• (00:00) The AI brief Sam now relies on before every new meeting

• (01:27) “What did I miss?”: Granola’s most used recipe

• (03:45) Why Granola hides the transcript despite user complaints

• (09:37) A Granola recipe to get feedback from founders

• (13:54) Why Sam feels a pang of sadness when agents skip your UI

• (19:30) Why Sam’s Figma use fell to 20% and how Granola stops slop

• (26:14) Nacho: The agent every Granola employee asks first

• (30:58) Why most people stop improving their product way too early

I’m proud to partner with Google AI Studio

Google AI Studio was my go-to prototyping tool when I was a product manager. You can use it to generate three versions of your product, add visuals with Google’s Nano Banana model, and publish them to a custom URL for customer feedback. It’s an easy, free way to move from prompt to something your team or customers can try.

Try Google AI Studio for Free

Top 10 takeaways I learned from this episode

Design principles for the AI era

1. Design for your user’s most stressful moment. Granola is built for the space between back-to-back meetings, when you’re late and forget who you’re meeting next. One of Sam’s principles is “Be calming: Granola should lower your blood pressure.” This has led to decisions like explicitly hiding the live transcript.

2. Design for agents using your app or product first. Sam felt “a pang of sadness” as a UI designer when people defaulted to using Granola through MCP and the API instead of the app. He resisted at first but now recognizes that many enterprises simply won’t adopt a tool that doesn’t plug into their internal agents. Some users will stick with the UI while others go through MCP, so design for both.

Read more

Dear subscribers,

Today, I want to share a new episode with Peng and Lauren.

Peng and Lauren are the lead designer and engineer for Grok Bot at SpaceXAI. In our interview, they showed me 14 bots that they use for work and life, from a design bot that helps Peng turn a single design into a user flow to an entire team of engineering bots that help Lauren ship faster.

Peng, Lauren, and I talked about:

• (00:00) The bots that the Grok Bot team actually uses

• (01:23) Peng’s chief of staff bot buys supplies and lists gear

• (06:31) Putting PM, design, and eng bots in the same room

• (10:07) A bot that turns photos into dioramas on Peng’s website

• (14:16) How Lauren’s bots test their own work before merging

• (17:59) Dr. Eggbot, a bot that designs and audits your bots

• (21:27) The assistant bot that booked Lauren’s multi-city trip

• (25:22) Eng bots that sometimes land PRs before Lauren reads them

• (38:59) How to trust your bots with more work, one skill at a time

I’m proud to partner with Granola

The world’s best teams, including Ramp, Linear, Replit, Brex, and more, use Granola to take notes and focus more in meetings. I’ve tried many meeting apps myself but keep coming back to Granola. It literally helps you take time back in your day.

Try Granola for Free

Top takeaways I learned from this episode

How the Grok Bot team uses Grok Bot

1. Build the design system and one keyframe, then let your design bot scale it. Peng’s design bot connects to Figma through Figma MCP and uses a skill that describes how his files are set up, plus his design system’s colors, type, and spacing. “I will do the first 5%,” Peng said. He builds the system and one keyframe, then asks the bot to extend it into every screen of the flow.

2. Make an eng lead bot to manage your other eng bots. Lauren hands big projects to her eng lead bot, Matcha. Matcha doesn’t do work itself. Instead, it breaks down the project for other eng bots to execute on. Each eng bot then spins up coding agents in the cloud, letting Lauren run “really massive agent swarms.”

Read more

I know what it feels like to come home after a busy day at work and try to learn AI at night. Every week, there’s a new model or tool that you’re supposedly behind on.

Since going independent, the best thing that I’ve built with AI isn’t a new tool or app. It’s a personal AI system that proactively takes work off my plate.

Today, I’m making the biggest upgrade yet to the paid membership by making it easy for you to copy this system through:

1. My full AI course with 25+ bite-sized lessons, 40+ prompts, and 16 ready-to-install skills.

2. $600+ in free AI tool access for Grok Bot, Cursor, Replit, Lovable, Granola, Wispr Flow, and more (valid for new users of each tool).

3. Monthly live workshops in a private Slack community where I’ll help you put everything into practice.

Join by October 7 to lock in $150/year for as long as you stay subscribed. On October 8, the price goes up to $200/year for new members.

Upgrade Now for Instant Access

If you’re already one of the 1,000+ paid members, you can keep your current price and sign in at behindthecraft.com to get all the benefits above.

Here’s a 3-minute video walking through what’s included in your new membership:

Get my personal AI course with 25+ lessons

ChatGPT, Claude, Grok Bot, Muse — you could spend all day just trying the latest AI tools.

Instead, I believe that everyone should invest in a system of personal AI skills and workflows that they can use with any new tool or model.

My AI course is the fast track to copy my system so that you can start taking recurring work off your plate this afternoon.

Across 25+ bite-sized lessons, you’ll:

1. Build an AI chief of staff to manage your inbox, calendar, and other tasks.

2. Set up an AI advisor to make better decisions about your life and career.

3. Find work to automate by building a plan and mapping out the steps.

4. Create reusable skills and chain them together to automate entire workflows.

5. Make AI proactive by copying my best work and personal AI workflows.

Here are 3 things that make my course unique:

1. It includes 16 AI skills and 40+ prompts that you can simply copy and paste.

2. All lessons are handcrafted and tested by me to be AI slop free.

3. I plan to update this course monthly as AI changes. It’s not a static course.

Here’s a quick snapshot of a lesson in the course:

Get $600+ in AI tool credits and discounts

There are many AI tool discount packs out there. But if you sign up for all the tools and aren’t careful about renewals, you could face a $1,000+ bill at the end of the year.

I chose a smaller set of AI tools that I personally use and can recommend. Together, the offers add up to more than $600 in credits, including:

• $200 in Grok Bot and Cursor credits

• 3 months free for Replit

• 3 months free for Lovable

• 3 months free for Granola

• 6 months free for Wispr Flow

• and more…

These discounts are valid for new users of each tool only. I plan to add more over time including from the frontier labs.

Upgrade Now for Instant Access

Join my private Slack community with monthly live workshops

After years of building online communities, I’ve found that a small community where everyone knows each other is better than a large one with thousands of strangers.

That’s why I’ve been sharing all my honest takes and memes in my private Slack for paid members. My goal is to get to know every paid member so we’ve also been hosting monthly live workshops and, soon, in-person events.

My next workshop is a free live lesson on October 2, where I’ll demo my AI chief of staff and show how my advisor, automations, and chained skills work together.

Reserve your spot here.

As a reminder, the price for new members will go up from $150 to $200/year on October 8 (in 2 weeks). Upgrade now to lock in the current rate.

As always, thank you for supporting me and this newsletter.

I hope to see you inside the community!
Peter

Dear subscribers,

Today, I want to share a new episode with Ethan.

Ethan is the product lead for ChatGPT Finances at OpenAI and he previously co-founded Digit, which helped more than 10 million people save over $9 billion. In our interview, he shared 6 practical ways to use ChatGPT to find wasted spending, prep for taxes, maximize credit card rewards, and model major financial decisions. I’m confident at least one of these use cases will save you money today.

Ethan and I talked about:

• (00:00) ChatGPT Finances can save you money right now

• (03:37) Use case 1: Cancel recurring subscriptions

• (09:37) How ChatGPT protects your financial data

• (18:03) Use case 2: Find tax savings proactively

• (24:58) Use case 3: Cut ETF and mutual fund fees

• (26:20) Use case 4: Maximize credit card rewards

• (28:20) Use case 5: Model income and taking time off

• (31:05) Use case 6: Decide whether to rent or buy

• (34:00) How to set up recurring finance checks

• (36:58) Shipping 47 product updates in one week

I’m proud to partner with Wispr Flow

There are many AI voice dictation tools out there, but Wispr Flow has been the most accurate one I’ve used. My kids also love using it to dictate prompts to Codex and Claude Code so they can build games and apps. Try it out with my free link below.

Try Wispr Flow for Free

6 ways to use ChatGPT Finance to save money now

1. Negotiate or cancel recurring subscriptions. Use this prompt to ask ChatGPT to review your recent subscriptions and suggest ways to help you negotiate or cancel them:

2. Find tax savings proactively. After you connect your bank and investment accounts to ChatGPT, you can ask it to proactively audit and optimize your taxes. Personally, it has been a massive timer saver for me in my conversations with my accountant. Start with this prompt:

Read more

Get instant access

Dear subscribers,

Today, I want to share my 3-step system to automate any manual workflow with AI.

I use this system to produce my podcast and it saves me at least 5 hours a week by turning 15 manual steps into 8 AI skills that work together seamlessly.

Watch my tutorial now to see the full system and learn how to apply it to your own workflows.

Timestamps:

• (00:00) How I save 5 hours a week with AI

• (01:34) Step 1: Map each step of your manual workflow

• (03:28) Step 2: Build skills to teach AI how to do the steps

• (09:33) Step 3: Chain skills together to automate a workflow

• (13:09) Improve your AI skills with every edit

• (14:25) Turn one episode into 6 different content pieces

• (17:08) How to apply my system to automate your own work

I’m proud to partner with Riverside

Riverside’s new MCP lets me prep episodes, edit recordings, create clips, and schedule posts to multiple video platforms, all from ChatGPT and Claude. You can also use it to host webinars, create brand videos, and much more.

Try Riverside for 1 Month Free

1. Map the steps of your manual workflow

To automate a workflow, you have to first give AI context about all the manual steps.

For example, here are 15 steps that I used to do manually to produce my podcast:

It’s clear that there are 3 phases just from making this list:

1. Podcast prep: Researching the guest and preparing the interview guide.

2. Podcast edit: Drafting instructions for my video editor after recording.

3. Podcast production: Packaging the episode for YouTube, newsletter, social, etc.

You probably don’t have a podcast, so pick a workflow you’d like AI to take off your plate and document every step. This prompt can get AI to help you do this:

2. Build skills to teach AI how to do the steps

I use the 8 AI skills above to produce my entire podcast while still applying my taste and judgment along the way. Here’s how they work and how you can set up a similar system for your own manual workflow:

Read more

Dear subscribers,

Today, I want to share a new episode with Pedro, the co-founder and CEO of Brex.

Pedro is widely respected as one of the most AI-pilled CEOs. In our interview, he explained why most people are building agents wrong and how to build them as AI employees instead, each with a specific job, skill, manager, and budget. He also gave a live demo of how he still uses OpenClaw to stay on top of his $5B business and life.

Timestamps:

• (00:00) Build AI employees, not AI agents

• (05:23) Why Pedro spends half his time reviewing work

• (12:32) Live demo: Jim, Brex’s AI recruiting employee

• (21:24) Why Brex makes engineers build with AI in interviews

• (22:33) CrabTrap: How to stop AI employees from leaking data

• (27:30) The tokenmaxxing bill is coming. Here’s how to control it

• (40:45) Live demo: Pedro’s OpenClaw system runs his life

I’m proud to partner with Riverside

A webinar shouldn’t be a one-and-done event. Riverside lets you plan, stream, record, edit, and repurpose everything in one tool. You can host live Q&As and polls, capture every speaker in 4K, then turn the recording into clips, recaps, newsletters, and social posts with AI’s help. Use my code PeterYang to get one month free.

Try Riverside for Free

Top takeaways I learned from this episode

Build AI employees, not AI agents

1. Build an AI agent like you’re onboarding an employee. Give your agents the same operating constraints as a good employee:

   • Job: Own one clear outcome.

   • Skills: Include specific instructions and constraints.

   • Manager: Escalate to the human manager if unclear.

   • Budget: Spend only tokens based on what the outcome is worth.

2. Jim (Brex’s AI recruiter) is an example of what an AI employee looks like. Jim’s job is to help recruiters find quality job candidates. His skills include sourcing candidates, filtering inbound applications, and running analytics for any job. He can access Greenhouse and LinkedIn and escalates to a human when he’s unsure.

Read more

Jensen Huang claims that “AGI has arrived” with OpenAI’s GPT-6 Astra.

But I’ve found the model both incredible and frustrating to work with.

In this post, I want to share three things that I love about Astra and two problems that hold it back from becoming my daily driver.

I’m proud to partner with Riverside

I’ve used Riverside for years to record my podcast thanks to its reliable, high-quality recording tools. But now I use it for much more than recording. Riverside has evolved into an end-to-end platform with AI editing, clip generation, streaming, and publishing tools. Try Riverside free for one month with code PETERYANG.

Try Riverside for Free

Astra is incredible at games, high-level thinking, and computer use

1. Games and 3D models

Astra has gone viral on social because it’s incredible for building 3D models and games. I made a 24-minute tutorial that walks through how I built 4 games (ok, more like tech demos) that look like Star Fox, Call of Duty, StarCraft, and Slay the Spire:

I’m especially fond of my roguelike deckbuilder, No Moat, where you’re an AI founder who has to hire teammates and play your cards right to grow your company. You can play it for free here (best on desktop).

Astra is great at games because it can use tools like Blender to create 3D models and Godot to turn those models into playable games. It’s wild that I can just ask AI to create models in Blender without learning how to use Blender’s complicated interface:

You can build a game with Astra even on the $20/month ChatGPT plan. Just follow these 4 steps that I cover in my tutorial:

1. Ask AI to “Install Blender and Godot MCP”.

2. Discuss your game idea with AI.

3. Ask ChatGPT Images to “Mock-up what this game would look like.”

4. Ask AI to build the game from the mock-up. Play it, then give feedback.

2. High-level thinking and strategy

Astra is also good at high-level thinking and strategy. My favorite way to test a new model’s thinking is to share my plans and goals, then ask for brutally honest feedback:

This prompt usually gives me something new and uncomfortable to think about. Astra was especially blunt. 😅

3. Computer and browser use

Finally, Astra is good at computer and browser use. A fun prompt to try is to give ChatGPT a photo of you and ask:

Then Astra will draw a masterpiece that looks like this:

All jokes aside, I do have one big caveat with Astra’s computer use that I’ll cover next.

Astra hesitates to do the work and doesn’t follow my AI skills reliably

I’m less than a week into testing Astra, but two problems have stood out:

1. It’s often hesitant to “just do it”

When I ask Astra to do a task, it often explains how to do it instead of just doing it.

OpenAI’s official Astra prompting guide acknowledges that their new model can “stop when the user may expect it to make reasonable assumptions and persist.” To discourage this behavior, consider adding this to your custom instructions:

Overall, I find that Astra doesn’t quite “have the dog in it” to just get something done like GPT-5.6 Sol does.

2. It doesn’t follow the instructions in my AI skills well

This is more of a deal breaker for me than the previous drawback.

Astra tends to take liberties when following the instructions in my AI skills. For example, I have a very specific format that I want AI to follow for podcast production, but Astra just decided to ignore some of my rules:

This continued to happen even after I simplified my skills to remove outdated and conflicting instructions. I didn’t have this problem with GPT-5.6 Sol or Claude Fable.

My verdict

When Astra launched, my plan was to switch every ChatGPT task to use Astra Medium by default. But instead, I think I’m going to:

1. Use Astra Medium for building games, strategic thinking, and one-off browser and computer use. When I want a deeper perspective on my plans or help creating something I couldn’t build before, it’s a great model to work with.

2. Stick to Sol Medium for most execution work using my AI skills. It’s more predictable at following my rules and an established workflow.

Let me know what you think about OpenAI’s new model in the comments!

Today, I want to share a new episode with Sue Khim.

As the co-founder of Brilliant, Sue is building an AI tutor to help kids think for themselves. In our interview, she shared why better grades at school don’t always reflect learning and gave a demo of how AI can guide kids to understand problems instead of getting the answer. This episode is a must-watch if you’re a parent.

Watch now on YouTube, Apple, and Spotify.

Sue and I talked about:

• (00:00) Why AI should make kids think harder

• (01:34) Why test scores are down while grades are up

• (09:41) Koji: An AI tutor that won’t give you the answer

• (18:05) Why humans design lessons and AI builds them

• (22:25) Primitives: The building blocks behind Brilliant’s AI

• (29:30) Testing AI lessons with simulated students

• (32:10) Measure learning, not lesson completion

• (34:39) Sue’s rule for competing with the big AI labs

• (39:10) How to raise kids who enjoy doing hard things

This episode is brought to you by… Linear

Linear is the tool of choice for over 40,000 AI-native teams ranging from startups to enterprises. Many of today’s leading AI companies, including OpenAI, Cursor, and Cognition, use Linear to run product development end-to-end with agents. Try it for free with my link below.

Try Linear for Free

Top 10 takeaways I learned from this episode

How AI can help kids learn

1. “Using AI to skip learning is like bringing a robotic arm to the gym to lift weights for you.” I told Sue I feel this temptation too: It’s easier to ask AI to solve a problem than to understand how to solve it myself. When teaching kids to use AI, teach them to ask it for questions and hints that help them work out the answer instead of just asking for the answer.

2. Even a good explanation can do the thinking for your child. From Sue: “Cheating and explanations are closer than you’d think.” Both can let a child skip working through a problem themselves. When your child gets stuck, ask them to explain their thinking before you explain the answer. Give them a chance to discover what they’re missing.

3. Help your kids get absorbed in their interests. Sue pays attention to how deeply her kids explore their interests. It could be monkey bars, chess, or magic tricks. The aim is to experience the effort and satisfaction of improving, without treating learning as something you do to please adults. “It should feel hard to learn something new.”

Read more