The thesis for my latest investment is simple: Ev Williams is one of the founders. End of thesis.
Ev co-founded Twitter with Jack Dorsey back in 2006. Now, he’s working on a new startup called Mozi.
Mozi helps you connect with your actual friends in the real world. It makes social media actually social again, as opposed to a mad dash for followers.
When I go on X, I don’t see any of my close friends. None of them use it!
Increasingly, I don’t even see people I follow. X, like all social apps, is going the way of TikTok – showing you the most engaging content, regardless of where it came from.
Mozi is different. You only see your actual friends.
Let’s say my old friend Tim is up from Virginia for the day. He could easily forget to tell me. Mozi could alert us right away so we don’t miss that special moment to be together.
This is exactly the kind of social app I’ve been looking for. In fact, right before I got the deal memo for Mozi in my inbox, I was thinking to myself, “No one is authentic on social media anymore. It’s all about clout and followers.”
It’s funny isn’t it, life’s coincidences?
Mozi is the direction I want to see technology go in: real life connection, not clout chasing or AI girlfriends. We should be bringing people together, not further isolating them.
Together with his co-founder Molly, Ev is going to change the way we relate to technology. We’ll be happier for it.
Check out Mozi and reconnect with the folks who matter most!
Sorry for the lack of posts last week, guys! I got sick, but I’m 100% now and delighted to be back!
What Happens If You Invest Money You Can’t Afford to Lose
If you can invest in startups, you’re rich.
I know, you’re not Bill Gates. Neither am I.
But you’re well ahead of most humans. You’re not scratching to survive.
Do you like this comfortable life?
Do you like not having to work for anyone else? Or at least being able to withstand some time unemployed?
Do you like not having to worry about money, or whether you’ll eat?
I’d wager you do. So do I.
So don’t risk it. Not for anything.
Warren Buffett said it better than I can:
“It’s insane to risk what you have and need for what you don’t have and don’t need.”
Wrap-Up
Why do people invest in startups?
I’ll tell you why I do it.
It’s interesting. Simple as that.
And no, I’m not insensitive to money. If I make a bunch, I might buy something. Why not?
But I didn’t need money when I started this. If I did, how could I even do this full time?
So I have my fun. But I’m not going to imperil this free and secure existence.
That would be silly.
And I suggest you don’t either.
Note: I am not an investment advisor and this is not investment advice. It is merely my opinion derived from research and experience. Take it with a grain of salt.
“All the greats in business studied the greats in business.” Podcasts are a wonderful way to do it. Here are my favorites…
This Week in Startups — My #1 most listened to podcast for several years running. If you’re a founder or investor, TWIST will tell you everything you need to know.
You can stop reading the tech news. TWIST is the place.
I’ve learned so much from this show. Half my investment strategy comes from it!
A great episode to start with is their recent VC Roundtable, which prompted many hours of thinking, writing and strategy changes on my part.
All In — Any part of my investment strategy that didn’t come from TWIST or Jason’s book Angel probably came from All In.
It’s a rollicking tour through tech, politics and more. I look forward to it every week!
My favorite episode of the year was their show with the Collisons. If you’re new to All In, it’s a wonderful place to start!
Founders — This is the most underrated tech and business podcast in the world. The host, David Senra, is the source for the quote above.
This guy is a maniac.
He finds books on European founders that are only in Italian or French and has them translated just for him! Then he tells you about it.
This is unique content you can’t get anywhere else. I only wish I’d found out about it sooner!
20VC — Harry Stebbings might be the hardest working man in venture capital. I’m just glad he’s over there in Europe so I at least have a chance. 🙂
Harry has these awesome, unique interviews you won’t hear anywhere else. One I loved recently was with Edwin Chen, founder of Surge.
Edwin had never been on a podcast, as far as I know. I’d never even heard of the company.
And yeah, it’s at $1 billion a year in revenue with no funding. Wow.
If you don’t listen to 20VC, you’re missing out.
Wrap-Up
Podcasts are such a wonderful way to learn. You can listen while you do housework, get your steps in, you name it!
With shows as good as these, I find myself stopping frequently to rewind and take notes. That’s how you know you’re really getting something out of it.
I’m so grateful that all these folks produce free, insanely high quality content that I can learn from and enjoy.
Check them out!
What are you favorite tech podcasts? Leave a comment and maybe I’ll try them!
The EU is about to backdoor encryption. Here’s how this could affect you, even if you’re American…
What Is Chat Control?
The EU is proposing to break encryption with a policy called Chat Control. Chat Control scans all the info on your devices before it’s encrypted, providing a backdoor to encrypted protocols.
Chat Control would break encryption within the EU for iMessage, Signal, VPN’s, everything.
The EU claims it will only be used to detect child pornography.
But a backdoor is a backdoor, and the government can use it for anything they want. So could hackers.
If you’re reading this, chances are you’re in the United States, as I am. So, why does this matter for us, and for the world?
Why This Matters, Even for Americans
“Why do you even care about this, Francis?” I asked myself. “You don’t live in the EU, you never will live in the EU, and you rarely even visit.”
True. But I do enjoy my trips there. And I don’t feel good about visiting a place that is unfree.
I also fear that the next stop for Chat Control is right here in the US of A.
I highly doubt the Trump Administration would allow this. But you can bet that a subset of politicians and citizens, potentially on both the left and right, would.
If they take away people’s rights in the EU, the government will try to do the same here. It’s not a matter of “if”.
It’s a matter of “when”.
Defeating Chat Control Technically: Good Luck
It may not be possible for Europeans to stop the EU from implementing Chat Control. But can we defeat Chat Control in a technical manner?
I asked Gemini 2.5 Pro, prompting it by saying that it’s a world renowned expert on cybersecurity with a strong moral commitment to privacy.
Here’s a portion of its response:
“From a purely technical standpoint, there will always be ways for a determined and knowledgeable individual to circumvent ‘Chat Control.’ However, these methods will likely be complex and out of reach for the average user.”
So basically, you can’t defeat Chat Control in a technical sense. Or perhaps it could be done, but it’s not exactly feasible.
Will the EU Backdoor Americans’ Phones?
Let’s say I, as an American, visit the EU.
It seems extremely unlikely that the “Chat Control” backdoor to my encrypted apps would still be operational after my return to the United States. But it’s hard to say that to an absolute certainty.
Meanwhile, while you’re in the EU, the government would likely have access to all your information. That’s a huge breach of our privacy, even if it’s only for a little while.
Perhaps we might just start leaving our phones at home. If we visit the EU at all…
Wrap-Up
Stopping child abuse is a very worthy goal. But we can’t be willing to sacrifice all our freedoms in the process.
Reading about Chat Control, I got to thinking…what’s the future for Europe?
Take a look at the present.
Moribund economies. Rock bottom fertility. Minimal innovation. And ever more repressive governments.
It’s not looking good, folks. I don’t know how they pull out of this.
I hope Europeans get involved and put a stop to Chat Control. If they succeed, maybe there’s hope for the Old World yet.
I read a headline this morning that really pissed me off. The EU is planning to break encryption and scan everyone’s phone. Here’s why that’s a horrible idea…
What the Heck Is the EU Up To?
First, I’ll break down what Chat Control is, then why the EU’s excuse doesn’t hold water, and finally, what we can do about it.
“In Brussels, the European Commission, Danish presidency, and member states will be trying, once again, to revive the CSAM/chat control that would effectively break encryption online in the name of combating child sexual abuse material, a process more colloquially known as ‘chat control’.”
This would require breaking encryption within the EU for everyone. iMessage, Signal, VPN’s, everything.
This leaves a backdoor in every encrypted service which governments can exploit. Hackers can exploit it too.
Now that you know what Chat Control is, let’s talk about the flimsy excuse the EU is using to justify it.
Why “The Children” Is Just an Excuse
Whenever the government wants to take away your freedom, they use one of two excuses: terrorism or “the children.”
Lately, “the children” is more popular. Maybe that’s because it’s been a while since we’ve seen a major terrorist attack in the West. And as a guy who lives right next to New York City, I pray we don’t.
Hey, I’m all for protecting children. Put the child abusers under the jail.
But there’s always some limit to the amount of freedom we’re willing to give up to accomplish any goal, no matter how worthy.
Balancing Security and Freedom
Here’s a great way to stop child abuse…
What if we required cameras in everyone’s house? We use AI to scan the video feed for signs of child abuse.
This would massively reduce child abuse. But are we willing to pay that price?
Yeah, didn’t think so!
And of course, the government would tell us that the cameras are only going to be used to stop child abuse.
Nothing else. Pinky swear.
This is what the EU is saying about Chat Control. Do you believe them?
I sure as heck don’t. They’ll use it against whomever they please, including anyone who says something the government doesn’t like.
The UK is already arresting people for tweets. Chat Control will allow even tighter repression of speech, this time across the entire EU. And I imagine the UK will adopt something like Chat Control in about a nanosecond.
What’s to stop the EU from using Chat Control to flag and punish political dissent in general?
Wrap-Up
Hey, I get it. Chat Control isn’t a camera in your bedroom.
But slopes get slippery…
I don’t know about you, but I’m not ready to hand over my phone to the EU’s cops and bureaucrats. It all comes down to what Martin Luther King said:
“Injustice anywhere is a threat to justice everywhere.”
Tomorrow, we’ll talk about why Chat Control isn’t just a threat to EU citizens, but a threat to Americans and the world at large.
Cyan Banister is one of the best investors ever, backing SpaceX, Uber, Flexport and more. To find the next great startup, Cyan asks an unusual question…
The Human Behavior Question
When she sees a new product, she doesn’t just calculate the Total Addressable Market (TAM). She asks herself “Does it shift human behavior?”
Cyan’s best investments changed human behavior in big ways.
Uber has us getting into stranger’s cars every day. We never would’ve done that 20 years ago!
It also let people turn on income whenever they need it. This wasn’t possible before.
SpaceX is shifting human behavior in even more profound ways.
Making high speed internet available worldwide is a massive change in human life. People in an African village can work remote jobs and make 10x what they’d make locally.
And if Elon gets us to Mars, that would be perhaps the biggest behavior change in the history of our species.
How Cyan’s Appproach Is Different From Other Investors
I’ve never heard any other investor ask this question.
Like most angels and VC’s, I try to calculate a TAM for startups I meet. But many of the companies I meet barely exist yet.
They’re just beginning to sign up customers. They may pivot to a completely new product and market several times before they hit product-market fit.
I can still do a TAM calculation. But all too often, it’s numbers in the air.
The real market size is unknowable. Any attempt to calculate it is an exercise in false precision.
That’s why Cyan’s method is better. Instead of trying to quantify that which cannot yet be quantified, she asks a qualitative question.
“Does it shift human behavior?” is an easier question to answer. And if a product gets us stubborn humans to change how we behave, it must have had a big impact.
I like to calculate the TAM as well, just for reference. But the exact number is less important than how the startup impacts our lives.
Wrap-Up
Cyan’s track record as an angel and now a VC is better than just about anyone’s. I can learn a lot from her, and so can every other investor.
I’m only grateful that people in our industry are so free with their secrets!
If you’re an investor, ask “Does it shift human behavior?” after every founder meeting. If you’re a founder, ask how your product can shift how we behave.
We all want to work on the most important problems. To get there, we have to change human life for the better.
P.S. I found out about Cyan’s method from a wonderful, vintage episode of the Angel podcast by Jason Calacanis. He does an in depth interview with Cyan during her time at Founders Fund. I highly recommend you watch the full hour interview, which you can find here.
You just finished pitching that big VC. Did you bomb? Did you nail it? Here’s how to tell…
The Weird Way to Gauge Investor Interest
There’s a way to tell, but it’s counterintuitive…
The more questions they ask, the more interested they are. Mr. Moneybags interrogating you like a murder suspect isn’t a sign of skepticism. It’s a sign of being engaged.
Is Moneybags giving you rapid fire questions, one after another? Is he staring at you as he does it?
Then you’ve struck a chord. And now, he’s considering an investment, but he needs to check a lot of boxes first.
No two human beings are exactly alike. But all else being equal, more questions means more interest.
First, let’s go through what to do if Moneybags gives you tons of questions. Next, we’ll cover what to do if he’s silent…
How to Handle Investors That Interrogate You
Use this moment to maximum advantage. Answer each question clearly and concisely.
Here’s a simple rule: your answer should take the same amount of time as it took him to ask the question. Let’s go through an example…
Good answer:
“Moneybags: What was your MRR last month?”
“You: $12,000.”
Bad answer:
“Moneybags: What was your MRR last month?”
“You: Last month was crazy! We almost got there with Megacorp but our champion got reassigned and the deal got pushed to Q3. We’re almost at the line. And we have a huge pipeline that we’re working through. I want to hire another couple sales guys that are real killers to…”
You see what you did? Instead of a number, you gave him “blah blah blah.” Not good.
Give a clear answer. That takes Moneybags’ interest and keeps it high.
How to Handle Investors That Are Silent
Now, what if Moneybags doesn’t ask many questions? What if he doesn’t ask any at all?
First, you want to prevent this situation.
Start the meeting with a brief pitch. It should last around 5 minutes in a 30 minute meeting, and no more than 10.
Some investors, like me, want to skip straight to Q&A. Be prepared for that as well.
After you’re done pitching, never say “Any questions?” Say “What questions do you have?”
This assumes Moneybags has questions. Even if he doesn’t, it will push him to come up with some and ask them.
That gets him more engaged. When he’s asking you questions, he’s not answering emails, playing Candy Crush or staring out the window.
If he still can’t come up with any questions for you, pose some common questions you’ve gotten during your fundraise. Then answer them.
This could help overcome objections he has that he may not feel comfortable telling you.
If you keep getting a question from multiple investors, address it proactively in your pitch. This makes you more persuasive right off the bat.
Wrap-Up
Never fear an interrogation. Every question is a gift.
It means the investor is interested. And it’s an opportunity to persuade.
For each question you answer clearly and concisely, you’re that much closer to getting a check.
GPT-5 sucks. I tested it Friday, bad. I tested it again today, even worse. Stick a fork in it.
Let me show you how flawed this model is.
When I tested it on Friday, OpenAI was having routing issues that may have caused GPT-5 to underperform. So this morning, I ran it through 3 more tests to see if GPT-5 can finally live up to the hype.
Here’s what I found…
Round #1: Teach Me About Medical Tourism
This afternoon, I’m meeting with an awesome startup that helps Americans arrange medical procedures overseas.
I love medical tourism! I got a full physical in Japan with a battery of tests for only $100!
So I asked GPT-5 to size up the market and its key players. I gave it a very detailed prompt. I asked for info on market size, leading companies, and more.
GPT-5 gave me a weird error, asking me to choose between a report with current data or one that stops in mid 2024. Uh, who’s choosing an outdated report?
So, I told it to give me current data. It threw the same error again. Again, I told it to find me current data.
A third error! I give up.
Any decent model can do this. I ran the exact same prompt through Gemini 2.5 Pro and got a beautiful report in minutes. But GPT-5 just flails helplessly.
I’m giving this round an F.
Round #2: Why Is Dollar General So Successful?
If there’s one store I love, it’s Dollar General.
It’s so cheap, with such great staff and selection, that I find myself there several times a week. And it’s not just for folks who are struggling: one of the richest guys I know is a dedicated DG man.
This afternoon, I’m meeting a founder who used to be a top exec at DG. So I wondered, why is this company so darned successful?
Let’s ask GPT-5…
GPT-5 gave an okay response, citing factors like a strong staff training program and small store sizes. But it didn’t cite a single source, nor did it explain why DG does better than other dollar stores.
This is a pretty basic, surface-level report. It feels like something GPT-3.5 could’ve written 2.5 years ago.
Compare that to Gemini 2.5 Pro. Gemini gave me a beautiful, carefully footnoted report in seconds.
Gemini explained that DG owns its own distribution centers, which gives it a huge advantage in logistics and cost. GPT-5 never mentioned that.
I’ll give GPT-5 a C for this round.
Round #3: My Unpaid Research Assistant
I’m starting to feel sorry for ChatGPT. So, I’ll give it an easy one…
I’m always looking up the funding history of startups. I want to meet companies before their seed round, so I need to make sure they’re early stage before I contact the founder.
This morning, I wanted to know the funding history of a startup called Rid. Rid helps you sell your unwanted junk online.
Can GPT-5 help?
GPT-5 gave a solid response, finding that Rid is part of the current YC cohort. That means they probably haven’t raised a seed round yet.
But when I ran the same question through Gemini 2.5 Pro, it gave me a better answer. Gemini explained that YC gives companies $500k and explained how that financing works. This is a fuller and more direct answer.
GPT-5 gave me what I needed, but it wasn’t as informative as I’d like. I’ll give this round a B.
A B isn’t bad. But for a supposedly frontier model, it’s not great.
Wrap-Up
Overall, I’m giving GPT-5 a grade of C- on this re-test.
That’s even lower than on Friday, when I gave it a B-. Maybe these updates are actually making it worse?
I want to get excited about GPT-5. But between weird errors, incomplete answers and generic reports, I walked away disappointed.
GPT-5 is the worst product launch I’ve seen in years. For a company valued at $500 billion that is supposedly the market leader, OpenAI did a rotten job here.
Sam is a smart guy and OpenAI has a lot of great people. They also have incredible distribution.
I’m confident they can come out with a great model. But this isn’t it.
Can GPT-5 beat the reigning champion, Elon’s Grok 4? I ran them through 3 tests to find out.
When Grok 4 came out last month, I asked it 3 real world questions that I actually needed the answer to. Grok 4 crushed it, scoring an A-.
Can GPT-5 do even better? Let’s find out…
Round 1: GPT-5 As Investment Analyst
For our first round, I asked GPT-5 to help me manage my investments. How should I allocate across index funds, bonds and cash?
GPT-5 gave a solid response, recommending a portfolio of mostly equities. This makes sense at my age, 39.
But GPT-5 also recommended something unusual: a special allocation to small cap/value stocks.
I’ve asked numerous top AI models about financial planning, and none of them recommended this.
Since total stock market index funds already contain those type of companies, this allocation seems duplicative. GPT-5’s sourcing was also weak here, pointing to a Wikipedia article rather than a reliable, primary source.
GPT-5’s recommendation wasn’t bad, but it was overly complex without much additional benefit. For this round, I’m going with Grok 4.
Round 2: GPT the VC
For its next assignment, I asked GPT-5 to find me some startups.
People are having kids later in life. This means the need for IVF is going to grow enormously.
What startups could help make IVF cheaper and more reliable? Let’s ask GPT…
I had GPT focus on pre-seed companies with under $750k in funding. It could only find one startup that met my criteria: a company called Microgenesis.
The company looks interesting, but Crunchbase shows it has raised over $2 million in funding. So, the one startup GPT could find doesn’t even meet my criteria.
Grok 4 found several startups that fit my criteria when I tested it last month. In this round, Grok blows GPT-5 away.
Round 3: Make Me a Better Angel Investor
I meet with a lot of founders. I often wonder, “how could I do a better job in these meetings?”
Let’s ask GPT-5…
GPT gave some solid ideas, like making sure to show interest in what the founder is doing. But its response was much more limited than Grok 4’s.
Grok 4 added helpful advice about how to build dealflow and diversify my portfolio. Those are great things to keep in mind, especially for newer angels.
GPT-5 gave an adequate response, but Grok 4 was much more informative. I’m giving this 3rd and final round to Grok.
Wrap-Up
Grok 4 crushed GPT-5 in my testing, winning 3/3 rounds. Overall, I’m giving GPT-5 a B-.
Compared to Grok, GPT-5’s answers were very basic. The information was sometimes inaccurate and sourcing was spotty at best.
GPT-5 is a disappointment. After a 2 year wait from GPT-4 to GPT-5, I was expecting a lot more.
OpenAI had something special in 2022. But now, they’re falling behind.
I get better results from Grok 4 and Gemini 2.5 Pro. And for coding, Claude rules the roost.
“$800k ARR…is that contracted or live?” Most investors never ask this question. They’re making a big mistake.
Whether you’re an investor or a founder, you have to understand the difference between contracted and live revenue.
Tl,dr: Contracted ARR is signed revenue, live ARR is actual money you’re billing. Focus on live ARR.
Let’s break down the differences and how big gaps can creep up on you…
Contracted Revenue
This is the total value of all the contracts with customers you’ve signed. You may see this written as “cARR” for short.
Let’s say that you signed an $800k ARR contract with Megacorp. They’re your only customer. That means your contracted revenue is $800k ARR.
Sounds awesome right? Not so fast…
Let’s say that Megacorp manufactures bobbleheads of top VCs. They have 8 plants where they plan to take your product live.
But right now, you’re only live in their Pittsburgh plant. You’ll go live at the other 7 over the next year.
The contract states they’re only billed for plants that are live. This is typical in SaaS contracts, although there is the rare contract that lets you bill the customer immediately even if they haven’t rolled out your product.
Live Revenue
Here’s where we get down to real money…
You’re charging them $100k per plant. So right now, with just the Pittsburgh plant live on your product, you’re only getting $100k a year coming into your bank account.
That means your live revenue is $100k ARR. Each month, Megacorp sends you a check for $8,333.33. I’m assuming you actually remembered to bill them (a surprising number of startups don’t), the contract stipulates monthly payments, and you’re past their likely net-30 payment terms.
When will you get the remaining $700k/yr coming into your bank account? Well, the deployment schedule says that the other 7 sites will go live over the next 12 months.
The Dangers of Relying on Contracted Revenue
You’re expecting to see that remaining $700k soon. But what if Megacorp decides to delay the project?
That $700k you were counting as “revenue” suddenly isn’t coming any time soon.
Maybe they cancel the other sites entirely, and you only ever get the Pittsburgh plant. That would mean the $700k evaporates for good.
If you were hiring thinking you had $800k in revenue, you’ve overhired. You have people sitting around doing nothing and you can’t afford to pay them.
You also told your investors that you’re at $800k in revenue. Now, they find out that number is a little squishy…
They should’ve asked what your live revenue is. If they didn’t, that’s on them.
But there’s no guarantee they’ll see it that way. The investors could blame you for not being clear.
Where Contracted vs. Live Revenue Disparities Are Worst
Not all startups have big contracted vs. live disparities. But in certain industries, they’re everywhere.
If there’s one type of company particularly prone to huge disparities between contracted and live revenue, it’s healthcare companies.
Healthcare moves slow. Really slow.
They sign a contract today. But the rollout takes an eternity.
Don’t I know it! I used to roll out medical software products for hospitals all over America. Those projects took years!
So if you’re a healthcare investor or founder, you have to pay particularly close attention to the contracted vs. live distinction.
Enterprise SaaS startups in general tend to have big differences between contracted and live revenue. Again, it’s those long deployment periods.
Consumer/SMB SaaS companies don’t see this problem as much. Their customers usually just plunk down their credit card and start using the product.
Avoiding the Mess
If you’re an investor, always ask what live revenue is. Explain exactly what you mean: “How much cash is actually hitting your bank account now?”
If you’re a founder, always tell investors what your contracted and live revenue is. Be extremely clear which is which.
Meanwhile, don’t hire and expand based on some pie-in-the-sky contracted revenue number. You may never get that money!
Hire, expand, and spend based on what you can actually afford now. If you cannot pay those folks with your live revenue and cash in bank today, you don’t want to hire them.
Wrap-Up
Bottom line: focus on live revenue.
Part of not being a schmuck as an investor is knowing these little nuances with big implications. The same goes for founders.
I didn’t know any of this stuff when I started investing 4 years ago. But bit by bit, you start to learn the pitfalls and how to avoid them.
Contracted revenue is a promise. Promises are broken.
Live revenue is real. This is the number to watch.
Tremendous 