Tremendous

This platform lets me diversify my real estate investments so I’m not too exposed to any one market. I’ve invested since 2018 with great returns.

More on Fundrise in this post.

If you decide to invest in Fundrise, you can use this link to get $100 in free bonus shares!

I used this app every single day to dictate to my computer, I’m even dictating this text using Wispr Flow! It’s way better than Apple’s native dictation.

My productivity is up about 25% since I started dictating rather than typing. I’m also less tired and stressed.

Get a free month of Wispr Flow Pro here!

My Top 3 YC S26 Startups So Far: Robot Maids, AI Health Insurance, and AI-Proof Security

July 2, 2026

YC S26 kicked off last week. 56 startups have already launched. Here are my top 3 so far, from robot maids to cheaper healthcare…

Prescience

Just this morning, my Christian health share emailed me. My discount for staying fit? Gone.

Traditional insurance is even worse. Premiums through the roof, no increase in quality.

Prescience is fixing this mess.

Prescience runs employee-funded health plans. They use AI to prevent illness and lower cost.

Prescience pulls data from wearables, gets you doctor visits same-day, and even prescribes GLP-1’s proactively.

If this company succeeds, the upside is unlimited.

Nori Robotics

What if you could buy a robot maid for $1,300? That’s the vision for Nori Robotics.

These are robots on wheels, rather than legs. On my recent trip to Tokyo, wheeled robots brought my breakfast at my favorite diner. Wheels are stable and practical.

Best of all, when you teach Nori a new skill like pouring a drink, that skill becomes available to everyone.

Rock-bottom price. Network effects. I love everything about this business.

Nebula Security

Ever since the Anthropic Mythos announcement, the technology industry has been terrified of hackers using AI to break into their systems. Nebula Security may be the company to stop it.

Founder Eten Zou and the team recently found a critical exploit of Android. These guys have serious cybersecurity chops.

As AI finds one security hole after another, there will be a mad race to patch them. Nebula is the right company at the right time.

Wrap-Up

I’ve got cold messages out to all three of these startups. I’m itching to learn more about what they’re doing!

We can whine about YC’s valuations. But who else is backing so many ambitious companies in so many areas?

I’m excited to see more awesome S26 startups launch. I just might find my next investment…

There will be no blog tomorrow. Have a wonderful July 4th and Happy 250th Birthday, America! 🇺🇸

More on Fundrise in this post.

If you decide to invest in Fundrise, you can use this link to get $100 in free bonus shares!

You’ve Got Your First Customer. Now, How Do You Raise Money?

I used this app every single day to dictate to my computer, I’m even dictating this text using Wispr Flow! It’s way better than Apple’s native dictation.

My productivity is up about 25% since I started dictating rather than typing. I’m also less tired and stressed.

Get a free month of Wispr Flow Pro here!

From Cold DM to Wire in 5 Days: Why I Move Fast on Great Startups

July 1, 2026

An amazing founder DM’ed me last week. I knew this deal wouldn’t last. Here’s the exact timeline — from cold X message to wire transfer in just five days…

Tracking a Recent Investment, Minute by Minute

Tuesday, June 23:

1:42pm: Cold X message received

4:49pm: I reply and ask for Calendly link

10:22pm: Calendly link arrives

11:00pm: I book first open slot

Wednesday, June 24:

10:30am: Meet with founder, promise a decision by EOD Thursday

Thursday, June 25:

3:47pm: I request allocation

4:17pm: Founder confirms allocation received

Saturday, June 27:

2:30pm: Docs and wiring instructions received

Sunday, June 28:

5:24pm: Docs signed, wire transfer initiated

Why Speed Is Critical for Investors

I’ve been at this for five years now. I know what I’m looking for.

When I see it, I move fast.

Some investors take 3 weeks to respond to an e-mail. At that point, you’re not competitive.

Is a great founder going to wait 3 weeks? By that time, the round will be full.

But What About Due Diligence?

Diligence still matters. I looked into this startup quite a bit, reading reviews, digging into team backgrounds, etc.

But for a small, early stage check, diligence can’t drag on forever.

When I see a great deal, I work late into the night doing diligence. Then, I sleep on it.

That’s a trick I learned from Hetty Green, the “Witch of Wall Street”. She never made an investment without sleeping on it first.

By the next day, I have an answer.

Wrap-Up

Founders are used to hearing “no.” What’s worse is hearing nothing.

If I’m not ready to invest, I give a quick “no” or “not yet”. I don’t want to waste their time.

But when that rare deal makes my heart race, I’m not waiting around.

More on tech:

Is Y Combinator — Gasp! — Actually a Bargain Now?

I Left $157,000 on the Table: How I Missed a Great Investment

Save Money on Stuff I Use:

This platform lets me diversify my real estate investments so I’m not too exposed to any one market. I’ve invested since 2018 with great returns.

More on Fundrise in this post.

If you decide to invest in Fundrise, you can use this link to get $100 in free bonus shares!

I Left $157,000 on the Table: How I Missed a Great Investment

I used this app every single day to dictate to my computer, I’m even dictating this text using Wispr Flow! It’s way better than Apple’s native dictation.

My productivity is up about 25% since I started dictating rather than typing. I’m also less tired and stressed.

Get a free month of Wispr Flow Pro here!

Is Y Combinator — Gasp! — Actually a Bargain Now?

June 30, 2026

We all love to whine about YC valuations. But lately, YC has quietly become one of the best deals for investors…

Getting More for Your Money

When I was meeting YC companies 2-3 years ago, this is what the best ones looked like:

$100k ARR
$20-25M valuation (200x to 250x ARR)

Here’s what I’m seeing now in the top startups:

$1-3M ARR
$30-40M valuation (10-40x ARR)

The price is up a little. But traction has exploded, and revenue mutliple has dropped around 90%.

Why I’m Looking More Closely at YC

You’re paying a little more, but getting a lot more in return. This makes investing in YC a better value than it’s ever been.

I’ve made a bet in each of the last two batches. Given the stronger value proposition, I’m going to be looking at future batches more closely.

When I Can’t Make the Numbers Work

There are still two types of deals I avoid, whether from YC or anywhere else:

< $200k ARR, $25-40M valuation (too early)
White hot startups raising at $200-500M valuations (not enough upside)

Angel investing is a business of outliers. So naturally, there are exceptions: a startup building in an area I’m obsessed with, or a founder with an incredible track record.

Why Are These Startup Growing So Fast?

How the heck have startups gone from $100K ARR to millions in the same short program?

You guessed it: AI.

The best startups are using agents to recruit a team, build a product, and find customers at warp speed.

Wrap-Up

“Long ago, Ben Graham taught me that ‘Price is what you pay; value is what you get.’ Whether we’re talking about socks or stocks, I like buying quality merchandise when it is marked down.” – Warren Buffett

I love a good bargain. And I’m starting to find them where I’d least expect: the world’s top accelerator.

If you’re investing and you’ve stepped away from YC due to valuations, I encourage you to take another look.

And if you’re in YC and you’re building something amazing, shoot me a message! I’d love to hear about it.

DMs open.

More on tech:

38 Words Got This Founder on My Calendar — Here’s the Cold Email Formula That Wins

Save Money on Stuff I Use:

This platform lets me diversify my real estate investments so I’m not too exposed to any one market. I’ve invested since 2018 with great returns.

More on Fundrise in this post.

If you decide to invest in Fundrise, you can use this link to get $100 in free bonus shares!

AI Agent Hype Meets Reality: The Product Doesn’t Work (And Churn Is Coming)

I used this app every single day to dictate to my computer, I’m even dictating this text using Wispr Flow! It’s way better than Apple’s native dictation.

My productivity is up about 25% since I started dictating rather than typing. I’m also less tired and stressed.

Get a free month of Wispr Flow Pro here!

China’s GLM 5.2: The Most Powerful Open-Source Model Yet — But Does It Deliver in Real Life?

June 29, 2026

China just dropped the first open source SOTA model: GLM 5.2. It crushes the benchmarks, but struggled in my real world testing.

Today, I ran GLM through a three-round test: building a pitch deck, teaching me about data centers, and designing a website.

GLM showed some powerful capabilities. But it also made surprising mistakes…

Round #1: Making a Startup Pitch Deck

For our first round, I had GLM 5.2 make a pitch deck for Uber.

I turned on Wispr Flow and rambled on what I wanted the deck to cover: team, traction, and vision. I gave GLM some style pointers, like putting high-contrast text on a black background.

GLM produced a beautiful deck!

I’ve seen the original UberCab deck. It didn’t look nearly as slick. If only Travis had had GLM!

This round gets an A+. Every founder should make their deck with GLM.

Round #2: GLM Teaches Me About Data Centers

I was excited to use GLM’s Knowledge/Teaching Material mode to learn about AI data centers. But it made me switch models 4 times due to high usage, landing on the weaker GLM-5-Turbo.

Rotten UX!

Eventually, GLM produced some solid info. I learned that five 1 gigawatt data centers consume as much electricity as all of New York City.

Power really is the bottleneck!

The report was great, but getting it was a pain. I’m giving this round a C.

Round #3: Designing a Website

GLM 5.2 claims to be strong in coding and design. Can it make a better design for this blog?

I gave GLM a vague prompt for a DOS-like interface, but cleaner and fresher. It switched to GLM-5-Turbo again, ran slowly, then outputted a blank page.

I told GLM to redo it. This time, I got a very nice homepage.

But I shouldn’t have to ask twice. Lovable nails this on the first try.

I’m giving this round a B-.

Wrap-Up

Overall, GLM earned a B- in my testing.

It excels on pitch decks, but struggles to make a basic webpage. Overall, it did not feel nearly as powerful as Claude or Grok.

Still, GLM 5.2 makes a lot of sense for startups.

Download it, post-train it on your data, and run it dirt cheap. The outputs will be good enough for most products.

If token bills are killing you, give GLM 5.2 a try!

More on tech:

No Crew, No Budget: How Grok Imagine 1.5 Helped Me Create Slick Startup Marketing Videos

Save Money on Stuff I Use:

This platform lets me diversify my real estate investments so I’m not too exposed to any one market. I’ve invested since 2018 with great returns.

More on Fundrise in this post.

If you decide to invest in Fundrise, you can use this link to get $100 in free bonus shares!

38 Words Got This Founder on My Calendar — Here’s the Cold Email Formula That Wins

I used this app every single day to dictate to my computer, I’m even dictating this text using Wispr Flow! It’s way better than Apple’s native dictation.

My productivity is up about 25% since I started dictating rather than typing. I’m also less tired and stressed.

Get a free month of Wispr Flow Pro here!

I Left $157,000 on the Table: How I Missed a Great Investment

June 26, 2026

I just missed out on $157,000. Last year, I passed on a startup that is now raising at 25x the valuation. Here’s how I blew this deal, and what I’ll be doing differently next time…

Why I Passed

We’ll call this startup Supercorp. In April 2025, I had an opportunity to invest in their seed round.

They had solid growth of 3.5x year over year. The founder had a good technical background, and tthey were nearing $2 million ARR.

What worried me was the market. Their segment had fewer than 2,000 potential customers, according to my research.

Even at Supercorp’s healthy ACV, that produced a TAM of just $500 million. What’s more, their growth was a little below what I was seeing in some other startups.

What I Missed

The biggest thing I missed here: focusing too much on their initial customer segment.

They soon expanded way beyond it. Now they’re building for many industries, shipping new products, and massively expanding their TAM.

I also didn’t pay enough attention to the founder’s background. His decades of experience in the market he’s serving gave him a huge advantage.

Unconscious Bias?

There may be another reason I missed Supercorp. The team doesn’t look like your typical startup.

They’re not 22. They don’t live in San Francisco or New York.

Did that bias me unconsciously? Hard to say.

But it’s possible. When I meet less traditional teams in the future, I’ll have to look at them extra closely.

What This Miss Cost Me

The seed stage deal was done around $20 million. Now, they’re raising at $500 million.

This was a syndicate deal, and it had some significant fees on it. But after fees and dilution, my typical $10,000 check would still be worth $157,000. That’s a 16x return in 14 months.

Instead, I made nothing. Yeah, it sucks.

Wrap-Up

In the end, I got hung up on TAM math. What I didn’t see was a strong team with a great product who’d find someone to buy it.

Every investor misses deals. This one stings, but there’s always another.

Hopefully I’ll nail the next winner.

I’m off to the woods for a camping trip. ⛺ Have a great weekend, everybody!

More on tech:

Fundraising Friction is Killing Your Round – Here’s How to Remove It

Save Money on Stuff I Use:

This platform lets me diversify my real estate investments so I’m not too exposed to any one market. I’ve invested since 2018 with great returns.

More on Fundrise in this post.

If you decide to invest in Fundrise, you can use this link to get $100 in free bonus shares!

Fundraising Friction is Killing Your Round – Here’s How to Remove It

I used this app every single day to dictate to my computer, I’m even dictating this text using Wispr Flow! It’s way better than Apple’s native dictation.

My productivity is up about 25% since I started dictating rather than typing. I’m also less tired and stressed.

Get a free month of Wispr Flow Pro here!

38 Words Got This Founder on My Calendar — Here’s the Cold Email Formula That Wins

June 24, 2026

Yesterday, a founder nailed a cold message in just 38 words. I put him on my calendar first thing the next morning. Here’s why short messages win…

Investor Time Is Very Limited

You only have 60 seconds to make an impression.

Investors, big or small, get tons of inbound. I look at a couple thousand deals a year.

I simply cannot spend very long looking at most deals.

That founder who sent me the 38-word message? I put him on the calendar first thing.

The guy who writes me a novel? Probably not going to get the response he wants.

Writing a Brief Cold Message That Actually Converts

A good cold message looks like this:

One sentence vision
Team info
Traction

Here’s a fictional example:

“I’m Jim, founder of Megastartup. I built my last startup to $5 million in annual revenue.

Megastartup combines every YC company into one startup and automate it with agents.

We’re at $100k ARR.”

Why This Message Will Get You Meetings

See how short that was? Just 33 words.

I just timed how long it took me to read it aloud: just 11 seconds. You can hold someone’s attention for 11 seconds, even these days.

In that short moment, you’ve shown you have a strong background. You’ve clearly communicated what you’re doing. And you’ve shown traction.

This should be enough to get you a lot of meetings.

Long Messages Get Deleted

What if I had written a cold message that goes on for four paragraphs and a couple of hundred words?

The investor sees that wall of text and hits delete. There’s 75 other emails in his inbox.

Your longer message didn’t make a better impression. It made no impression at all.

Cut your message to the bone. I want to see 50 words, tops.

What’s True of Cold Messages Is True of Decks

So you’ve cut the fat from your cold message. But are you still sending investors a deck with 42 slides?

Then you’re doing it wrong.

Seven slides. That’s it.

Title
One sentence vision
Team
Traction
Market size
Competitors
Ask

Each slide should have just one or two sentences. Make the text enormous and easy to read.

People flip through pitch decks. You want to enable that.

Wrap-Up

A lot of the problems founders have in fundraising come down to this: they think investors spend hours pouring over their message and deck.

We don’t. You get a minute, maybe two.

In those brief moments, we decide whether it’s worth meeting you. Give us some compelling facts, fast!

Cut your cold message. Throw out your deck.

And watch the number of investor meetings you get go through the roof.

More on tech:

AI Agent Hype Meets Reality: The Product Doesn’t Work (And Churn Is Coming)

No, You’re Not Getting a $250k Angel Check — Here’s Why

Save Money on Stuff I Use:

This platform lets me diversify my real estate investments so I’m not too exposed to any one market. I’ve invested since 2018 with great returns.

More on Fundrise in this post.

If you decide to invest in Fundrise, you can use this link to get $100 in free bonus shares!

AI Agent Hype Meets Reality: The Product Doesn’t Work (And Churn Is Coming)

I used this app every single day to dictate to my computer, I’m even dictating this text using Wispr Flow! It’s way better than Apple’s native dictation.

My productivity is up about 25% since I started dictating rather than typing. I’m also less tired and stressed.

Get a free month of Wispr Flow Pro here!

Fundraising Friction is Killing Your Round – Here’s How to Remove It

June 24, 2026

Friction is killing your fundraise. Here’s how to remove it and close out your round, fast…

Take the example of an awesome founder I spoke with recently.

“An investor told me to require an e-mail address to see the deck,” he said. “What do you think?”

“Don’t require anything. Remove all friction,” I said. “Don’t give them any reason to click off your deck and ghost you.”

Common Friction Points to Remove From Your Fundraising Process

Every bit of friction in your fundraise is costing you money. Here are some common forms of friction to avoid:

Asking for an e-mail address to see a deck
E-mailing back and forth to schedule a meeting (just send a Calendly link)
Requiring anything more than an e-mail to book a Calendly appointment
Requiring passwords or 2FA to see a deck or book an appointment (yes, people do this)
Asking for NDA’s

Make it as easy as possible for investors to learn about your company, get excited, and book a meeting. Those meetings turn into checks.

Any Benefits from Tracking Aren’t Worth the Risks

Some founders think they should require an email address to see the deck so they can track who’s viewing it.

But you already know who you sent it to, right? What’s the point of tracking further?

Any possible benefit is hugely outweighed by the risk they get annoyed and move on to the next deal.

Wrap-Up

Fundraising is sales. If you want to sell more, make your product easy to buy.

Amazon makes it extremely easy to buy from them. You can get just about anything in 24 hours with a couple of clicks.

They’ve removed all friction. And who’s selling more than them?

Learn from Amazon. Make it easy for investors to see what makes your startup special.

What’s one piece of friction you’re removing this week?

More on tech:

No, You’re Not Getting a $250k Angel Check — Here’s Why

Save Money on Stuff I Use:

This platform lets me diversify my real estate investments so I’m not too exposed to any one market. I’ve invested since 2018 with great returns.

More on Fundrise in this post.

If you decide to invest in Fundrise, you can use this link to get $100 in free bonus shares!

The Secret AI Weapon That Could Close Your Round: VC Boom

I used this app every single day to dictate to my computer, I’m even dictating this text using Wispr Flow! It’s way better than Apple’s native dictation.

My productivity is up about 25% since I started dictating rather than typing. I’m also less tired and stressed.

Get a free month of Wispr Flow Pro here!

AI Agent Hype Meets Reality: The Product Doesn’t Work (And Churn Is Coming)

June 23, 2026

Pattern I keep seeing in AI agent startups:

> announce super powerful agent
> ship agent
> tons of sign-ups
> insane revenue growth
> massive funding round
> product simply does not work
> what now?
> heavy churn likely

Customers bought the dream of an agent that does everything.

But the product fundamentally does not work.

Worse yet, it produces junk outputs. Useless e-mails. Robocalls that scream “robot” and never convert a customer.

This AI slop means more work for humans, not less. So they cancel.

When the Sugar High of Growth Wears Off

As churn gets worse and worse, growth will become impossible to maintain. Eventually, revenue will trend down.

If the startup can make the agent work before runway runs out, it could still be a success. But for a lot of agents, they’re not even close to living up to the promises.

And because these startups have raised at such high valuations, raising more money will be hard. They’ll have to continue the crazy growth.

The likely outcome for these startups: death.

The problem with these companies is they have overpromised and underdelivered.

Lessons From a $100 Billion SaaS Company You’ve Never Heard of

20 years ago, I sat in the all staff meeting at Epic.

“What’s the best way to make sure we meet customer expectations?” founder Judy Faulkner asked.

“Work really hard!”

“Always put the customer first.”

People continued giving these Boy Scout answers. Judy shook her head.

“The key is to set expectations very low.”

That’s how Judy built a company that the Acquired guys estimate is worth over $100 billion.

Nail It, Then Scale It

AI agent startups love to claim they automate an entire job. Very few pull it off.

A better approach: nail one small part of a job. Scale that across lots of customers.

Then add more capabilities.

People will happily pay for a product that does one important task well. But if you try to do everything and screw it up, they’re going to churn.

Wrap-Up

Instead of building some superagent to do all things for all people, make your focus as narrow as possible. Test the heck out of your product and be sure it works in real-world conditions.

It’s exciting to have a buzzy launch and get to millions of revenue overnight.

But if you want to build an enduring company, try doing it Judy’s way.

More on tech:

No, You’re Not Getting a $250k Angel Check — Here’s Why

“Don’t Be Afraid to Fail” — Three Lessons from Elon That Will Make You a Better Founder, Leader, and Investor

Save Money on Stuff I Use:

This platform lets me diversify my real estate investments so I’m not too exposed to any one market. I’ve invested since 2018 with great returns.

More on Fundrise in this post.

If you decide to invest in Fundrise, you can use this link to get $100 in free bonus shares!

The Secret AI Weapon That Could Close Your Round: VC Boom

I used this app every single day to dictate to my computer, I’m even dictating this text using Wispr Flow! It’s way better than Apple’s native dictation.

My productivity is up about 25% since I started dictating rather than typing. I’m also less tired and stressed.

Get a free month of Wispr Flow Pro here!

No Crew, No Budget: How Grok Imagine 1.5 Helped Me Create Slick Startup Marketing Videos

June 22, 2026

Grok Imagine Video 1.5 is here

Our new image-to-video model with sharper realism, better physics and faster generations 🧵https://t.co/zGhs9czkC5 pic.twitter.com/9X4YicpMH8
— xAI (@xai) June 17, 2026

Elon just dropped Grok Imagine 1.5, his most powerful video model yet. Instead of hiring influencers and rigging cameras, I used Grok to create marketing videos out of nothing…

I tested the new model in 3 realistic startup marketing scenarios. Here’s what works and where Imagine still needs some polish…

Round #1: UGC-Style Marketing Videos, No Influencer Needed

Say you’re launching a new line of smoothies. You want to produce some UGC-style ads to show off your product.

Two years ago, you would’ve paid a gang of influencers. Today, startups are creating influencers out of thin air.

I had Grok make a video of an athletic girl showing off our new smoothie. Grok produced a decent video, but it doesn’t look as natural as I’d like.

Also, the maximum length is 12 seconds. That may be okay for TikTok or Reels, but it’s a little short.

I’m giving this round a B-.

Round #2: Making a Launch Video Without a Camera or Mic

Have you ever wanted to make your own buzzy launch video and get millions of views? Grok Imagine makes it easier than ever.

I could futz around with a fancy camera and mic setup. Or, I could just seed the prompt with a picture of me and have Grok do the rest.

I used my X profile pic and gave it a basic prompt to produce a launch video for an app. The video was realistic and engaging…maybe more engaging than I am in real life. 😂

Seeding the prompt with an image made a huge difference. The video was much more realistic.

The screens for the app are generic. No surprise there — I didn’t give it much specifics. Had I seeded the prompt with some screens from a real app, it would’ve come out even better.

For a short launch video, this works well. I’m giving Grok an A for this round.

Round #3: Product Concept Video, $0 Budget

You’re building a robot to sort packages. You want to show how awesome your robot will be when it’s finished!

But right now, your prototype is a janky mess of chips and wires. Can Grok help show everyone your vision?

I gave Grok a simple prompt, telling it to show a robot on wheels sorting packages. It came out with a great little video showing what the finished product could look like.

You can use videos like this in fundraising and recruitment. It helps get people excited about your startup!

Grok earns an A+ on this round.

Wrap-Up

Overall, Grok Imagine 1.5 earned a solid A-.

It did a great job of producing marketing videos for my imaginary startups, but the humans need a little more work to be realistic.

If you write your prompt carefully and seed it with images, Grok can produce awesome videos for $0 budget.

The next thing I want to see is longer-form video. We need videos of at least 2-3 minutes — well beyond the capabilities of any AI video generation model today.

Give the new Grok Imagine a try and get more eyeballs on your product!

More on tech:

You’ve Got Your First Customer. Now, How Do You Raise Money?

Save Money on Stuff I Use:

This platform lets me diversify my real estate investments so I’m not too exposed to any one market. I’ve invested since 2018 with great returns.

More on Fundrise in this post.

If you decide to invest in Fundrise, you can use this link to get $100 in free bonus shares!