Can AI help us live decades longer? How should businesses adapt to AI? I dug into that and more on a panel at the Single Family Office Summit in New York City.
Here are some of my favorite moments:
8:07: How AI plus biotech will change human health: KJ’s story.
11:56: My investment in Sent and why they will kill Twilio
If you live in New Jersey, I don’t have to tell you we’re choosing a new governor. But how should I vote? I used ChatGPT to help me learn about the candidates and make an informed decision.
The primary election is tomorrow, June 10. Incumbent Phil Murphy is term limited. While Mikie Sherrill leads, it’s anyone’s race at this point.
In local elections, it can be tough to find information about the candidates. But AI can do the research for me, scouring the internet for every piece of info on these pols.
I used a heavily “engineered” prompt to find info on the candidates. I told ChatGPT my priorities in terms of issues. Then I instructed it to give me the info in a table format. This will let me easily compare the candidates.
Here’s what ChatGPT found…
Crime, Bail Reform, and ICE
My #1 issue is crime. If we’re not safe, nothing else matters.
So first, I asked ChatGPT about crime and bail reform. The details here were slim, but Fulop stands out for his record of reducing crime in Jersey City. Score one for the mayor!
Cooperation with ICE has been controversial in New Jersey, with Newark mayor and gubernatorial candidate Ras Baraka getting arrested in a protest at a detention facility. I want to see full cooperation with ICE, removing dangerous illegal aliens from our state.
None of the Democratic candidates are taking a strong pro-ICE stand. Steve Sweeney seems a bit better than most, opposing a bill to stop ICE cooperation. But none are pledging to work closely with Trump on this issue.
Overall, none of the Democrats align with my views exactly. But given his great work reducing crime as mayor, I’m giving this round to Fulop.
The Housing Crisis
New Jersey is one of the most expensive states in America. Next, ChatGPT gave me an overview of how each candidate will address the housing crisis.
Most candidates don’t have a clear plan. Here’s my plan: build like crazy.
The only candidate who stood out in this round was Fulop. If you look at Jersey City’s skyline, you can see Fulop’s work — it’s grown enormously in his 12 years in office.
Fulop wants to double the rate of affordable housing statewide, which would do a lot to help the average man. All that construction also creates a ton of jobs.
I’m giving this round to Fulop as well.
Taxes
NJ has much higher taxes than most of America. That makes it hard to compete with no tax states like Texas and Florida.
Sweeney has pledged not to increase taxes. Josh Gottheimer goes further, with plans for a 15% property tax cut.
Fulop seems to be the only candidate who wants to raise taxes.
He wants to extend a 2.5% tax on business profits over $1 million. This discourages companies from doing business in our state and takes jobs away from our neighbors.
This round goes to Gottheimer.
Education
Only one candidate has made education a core message: Sean Spiller.
Spiller is the president of the state teacher’s union. It’s safe to assume he’d increase school funding. He also wants free pre-kindergarten statewide.
While Spiller’s plans could have some benefits, they’d also cost us a fortune. What’s more, a teacher’s union president is not going to advocate for needed reforms, like charter schools.
I’m calling this round a draw.
Wrap-Up
Given the lack of substantive info out there on these elections, ChatGPT did a great job. Having a tool to do research and put the results in a standard format in one place is incredibly valuable.
If you use AI to help you vote, I encourage you to read through the sources it cites and do some Googling as well. This helps you verify the info AI is giving you.
We need more reporting on these elections. We also need politicians to answer tough questions. This will make it easier for citizens to decide.
So, who did I vote for?
With strong stances on crime and housing, Steve Fulop got my vote. He may not be perfect, but if he can do for the state what he’s done for Jersey City, he’ll go down in history as one of our greatest governors.
I’m also looking forward to hearing from the Republicans in the general election. I just might switch teams come November!
Less than 4 miles from where I sit, UnitedHealth CEO Brian Thompson was murdered with a ghost gun. So I wondered, how big of a threat are 3D printed firearms?
A Rapidly Advancing Technology
Unfortunately, 3D printing firearms has become increasingly easy. There are entire YouTube channels devoted to it.
The guns they’re printing are increasingly sophisticated. Weapons that were once made of failure prone plastic parts are increasingly made out of much stronger aluminum.
In the video above, you can see the creator of the PSR channel firing a Czech-designed Nexus Evo 3 Gen 2.
I’m no gun expert, but it looks a lot like a typical assault rifle. He puts round after round through the weapon at high speed with no apparent issues.
3D printed guns like these are difficult or impossible to trace. And laws meant to stop the spread of firearms don’t catch these “ghost guns.”
How Big Is the Risk?
Seeing a masked man on YouTube firing a ghost gun is disturbing, no question there. But how much worse are ghost guns than the normal ones?
In most of America, it’s laughably easy to get a firearm. Where I grew up in Wisconsin, you can buy a shotgun at Wal-Mart just like you were buying milk.
Most states and cities allow handguns and even assault rifles like the AR-15. The 3D printed weapons are no more dangerous than what’s for sale in countless stores nationwide with few if any controls.
Take Luigi Mangione, the accused killer of Mr. Thompson.
He used a 3D printed pistol. But if he had been almost anywhere else in America other than New York City, he could’ve bought a pistol legally.
Remove every 3D printed gun, and Mangione probably still would’ve committed this awful crime.
How to Regulate Ghost Guns
Although it’s easy enough to buy normal guns in this country, we still don’t want every Tom, Dick and Harry printing weapons with zero oversight.
The government should work with 3D printer manufacturers to ensure their devices cannot print firearm parts. The 3D printer could have software that recognizes key firearm components, such as the receiver or trigger group, and stops the printing process. It could even notify authorities of the attempt.
This could prevent guns from falling into the hands of criminals and minors.
Wrap-Up
We need better regulation to stop ghost guns. If government works with 3D printer manufacturers, we could make a big dent in the ghost gun problem.
But we should also keep in mind how easy it is to get the normal ones.
I’m not saying take everyone’s guns away. But there has to be some middle ground between “take everyone’s guns away” and “unlimited, military grade firearms for the criminally insane.”
If you trade on inside information in the public markets, you go to prison. In the private markets, all we do is trade on inside information. Here’s how to do it…
How to Get Exclusive Startup Data
Last year, I met with an amazing founder we’ll call Jim. The company was pre-revenue, but for a pre-revenue business, this was as good as it gets.
I wasn’t ready to invest yet. But I asked if he could put me on his investor update.
Fast forward a year…
New e-mail in my inbox. It’s Jim’s latest update.
Acting on Superior Information
When I opened that e-mail, my jaw dropped. Jim had hit several hundred thousand in ARR in just a couple of months since launch. He was growing over 40% month over month.
I replied, asking him to meet. We got on Google Meet and he gave me the latest on the company.
These guys were absolutely crushing it. But could I get a taste?
I asked Jim to invest, and he happened to have a bit more space in the round he was closing up. I quickly wired over the money, eager not to miss the opportunity.
I was able to place this bet because I had inside information.
Only a few dozen people got those updates. I was one of them.
The average investor had no idea how fast this company was growing. But I did, and I was able to profit from that knowledge.
A Scalable Model
Being on the investor update of promising companies gives me a proprietary information flow.
Every month, I get a bunch of e-mails with non-public financial info for awesome early stage startups. I pick the best and write them a check.
This is a repeatable and scalable model.
Just last week, I got an update from another company I met last year. They’ve begun to grow like crazy, and I’m setting up a meeting with them now.
If you want to outperform other investors, you need some advantage. Better information is one of the most powerful advantages you can have.
Because almost all information about private companies is private, there is nothing illegal or improper about acting based on private information. That’s the only information there is!
Take advantage of that. When you meet an awesome founder, ask to be added to their investor update.
And if you’re a founder, add investors you like to your update list. This gives them an easy way to follow your company. It also means you’re going to get crazy inbound if your growth up-shifts, which saves you a lot of time in fundraising.
Wrap-Up
“All I ever wanted in life was an unfair advantage.” That’s what Frank Costello said. I couldn’t agree more.
Getting on the update list for great startups is an easy, legal way to get an advantage. There’s no reason not to do it.
As I write this, we’re 5 days into June. Soon, a raft of new updates will hit my inbox.
Among them just might be my next investment. I can’t wait!
Are you sick of getting those automated messages on LinkedIn? So am I. AI Sales Development Reps is one of the 5 hot categories I’ve never invested in.
Each cycle has a hot idea that 100 startups pursue. Most of the time, they all flame out.
Here are the hot categories I’ve never touched…
AI SDR’s
This market is incredibly crowded. There are hundreds of these companies, and they seem to have no differentiation.
Worse yet, the products don’t work.
I was just chatting with a founder yesterday and he showed me his LinkedIn messages. He got a note that said “Hi [blank], I see you’re Founder | CEO of…” It couldn’t find his name and it transcribed his job character-for-character from his LinkedIn bio.
No one’s responding to a message like that. So the AI SDR product doesn’t drive any sales.
Hello, churn! And sure enough, these products churn like you wouldn’t believe.
AI Note Takers
I use an AI note taker, and I love it. But there are 1,000 of these and one isn’t all that different from the next.
Also, an AI note taker is a feature, not a product. The product is the videoconferencing system.
You’re already seeing AI note taking embedded into Zoom. I’m sure Meet and Teams will have it soon.
Bye bye AI note taker.
AI Image Generators
I loved the image of Pope Francis in Balenciaga and ice. But I don’t love AI image generators as an investment.
We will probably do image generation using foundational models. Grok and OpenAI already do it well.
What’s more, there are about 1,000 AI image gen tools. None of them have any real differentiation.
It’s hard to build a $1 billion business that way.
NFT’s
I love art, but I don’t collect it. So why would I buy NFT’s?
I think owning digital art could become a thing. But the prices for Bored Apes and the like made no sense.
When I see incredible froth in a category, I like to take a step to the left and find another place to play.
Creator CRM’s
Like NFT’s, Creator CRM’s were all the rage in 2021. Now, most of those startups are dead or dying.
The problem is, only a full time creator is likely to buy SaaS products to manage their business. And there aren’t that many full time creators.
Even among full timers, most don’t make much money. This means LTV’s will be low and churn will be high.
Wrap-Up
What do all these categories have in common? They’re all obvious.
Sometimes, 100 people have the same idea and go build the same startup. Each product has a slightly different spin on the problem, but fundamentally they’re the same.
I go toward the non obvious ideas.
Take Prandtl Dynamics. They’re building an ultrasound system to stop drones. In the future, every soldier will carry one.
That’s transformative tech. And it’s not something 100 companies can do.
That’s the type of thing that gets me excited!
If you’re a founder, don’t make a me-too product. Make something transformative.
Do it even if everyone tells you it’s stupid. That could be the best sign that you’re on to something.
Did you know Amazon has its own AI models? Until yesterday, neither did I! This morning, I took them on a test drive.
Amazon’s models are called Nova. They can output text or images and even run workflows for you.
The Nova models are designed to use within applications running on AWS. Nova models are much cheaper than the more advanced OpenAI models, which is great for developers.
You can use Nova as a consumer too. Let’s see what this tool can do!
Stop Francis From Getting Fat
This afternoon, I’ll be meeting one of my founders who’s in town for New York Tech Week. We’re headed to Dominique Ansel Bakery, home of the famous cronut! And no it’s not hype — they really are great.
I wondered, can my figure afford one of these greasy treats? Let’s ask Nova how many calories they have…
Nova gives us a wide range, from 240 all the way up to a shocking 1,300 calories. I think I like the 240 number better. 🙂
Nova also provides 3 citations. I clicked through on each. The first two are relevant, but the third contains the calorie count for croutons, not cronuts.
Still, Nova gave me some good and useful info. I’ll give this round a B+.
Spicing Up My Exercise Routine
I find myself doing the same lifts over and over at the gym. Maybe Nova can help…
Nova had some interesting ideas, including dumbbell rows. I’m not quite sure how to do those, so I’m going to watch a few YouTube videos.
What I’d really like to see Nova do is include a video in its output, showing me how to perform each exercise. That would be incredible!
But still, this was helpful info. I’m giving this round an A-.
Researching a Startup
Last night, I met with a startup creating a tool for laboratories. After the meeting, I wondered…how many labs are there in America? Let’s ask Nova…
My first 2 questions were pretty simple, and I ran them on Nova Pro. For this last one, I’m using Nova Premier, Amazon’s best in-house model.
Nova finds that there are around 60-80,000 research laboratories across academia, industry and government. That’s a big number, but the citations Nova provides are pretty shaky.
They have some basic data for the government. But when I click through on the other sources, they don’t have any relevant info.
It’s hard to trust this number. But since the government lab count seems plausible, I’ll give Nova a C on this round.
Wrap-Up
I was surprised how well Nova performed!
Overall, I’m giving it a B. If Amazon improved the citations, it could reach an A.
As Nova improves, Amazon may no longer need to partner with Anthropic. That could leave the Anthropic team out in the cold as OpenAI and xAI raise mountains of cash.
Give Nova a try! You may be surprised how good this unknown model is.
DeepSeek just dropped a new model, R1-0528. Has China leapfrogged the US? This morning, I tested the model to find out.
DeepSeek shocked the world with its R1 release in January. But when I put it up against the best from Gemini, DeepSeek lost.
Let’s see how DeepSeek’s latest release performs…
Timing My Workouts
I need to get a brief HIIT bike workout in this morning. But I’m not sure of the best time to do it.
I ate a light breakfast about an hour ago — is it too soon to work out? Let’s see what DeepSeek says…
DeepSeek gives me a variety of options in a nice table format. My breakfast was on the larger side in the context of their table, although I’d still consider it light for me. I’m an eater! 🙂
So, it sounds like I need to wait around 2 hours. Guess I’ll blog instead!
This was a pretty good response, although I wish they had quantified the breakfasts in terms of calories. I’d also like to see some citations to experts on exercise.
Overall, I’m giving this answer a B.
SaaS Companies Making It Big
In 4 years investing, I’ve had a few companies go bust. But luckily, I also have a few that are beginning to break out.
Two have cracked $20 million ARR so far. But how rare is that, really?
Let’s ask DeepSeek how many companies make it to that level…
DeepSeek estimates that only 1 out of every 200-500 SaaS startups ever cracks $20 million.
DeepSeek gives some citations, but there’s no clickable link to verify the info myself. That makes me wonder if it’s even real.
So, I tried to verify the data on my own. And guess what that involves?
A trip to Google! I may as well have not even bothered with DeepSeek.
This query demands a data driven response. The whole question involves statistics. But DeepSeek couldn’t get it done.
I’m giving DeepSeek a C here.
Startup Research
This evening, I’m meeting with a company that’s using robots to automate scientific research. I’ve never seen anything like it.
But I wondered, is anyone else doing this? Let’s ask DeepSeek…
DeepSeek gave a pretty strong answer here! It turns out that although automated, robotic research is new to me, it’s not exactly new.
A company called Strateos is doing it, as are several major drug manufacturers. I’ll look into their platforms more this afternoon before my meeting.
I would’ve liked to see some citations here. Instead, I have to go hunting on my own.
But overall, DeepSeek gave me a pretty good answer. I’ll give this round a B+.
Wrap-Up
DeepSeek’s answers would’ve been impressive 2 years ago. Not anymore.
The responses were informative and seemed reasonably accurate. But without any citations, it’s hard to know how accurate.
I can get similar responses from many models. Llama, Claude, Mistral, they could all easily do what DeepSeek just did here.
DeepSeek is behind the leaders, OpenAI, xAI, and Google. Those products do much deeper research with better citations.
China is building bridges with massive robots. Is America hopelessly behind?
This weekend, I heard about China’s bridge building robot on The All-In Podcast. The idea of building a bridge at warp speed with robots intrigued me.
So this morning, I decided to learn more…
The Iron Monster
Many Chinese bridges are built with a massive robot called the SLJ900/32, also known as the Iron Monster. It can move big sections of pre-built bridge into place in a matter of hours.
Looking at the Iron Monster, it’s easy to view America as pathetically behind. We use cranes that take weeks to do what the Iron Monster does in a few hours.
But actually, each country is using the method that best suits its needs.
Is America Behind?
The Iron Monster is great for building the same type of bridge over and over, especially for high speed rail. China is building a ton of bridges right now, so they need that type of construction.
America has been a developed nation for a long time. We already have tons of bridges, so we rarely need to build a new one.
We also have little need for high speed rail. Our country’s population density is low and we have tons of airports.
What’s more, new technologies on the way may make high speed rail obsolete. Self-driving cars and eVTOLs can satisfy the same need at a fraction of the cost.
What we need isn’t a bunch of new bridges. It’s maintenance of the bridges we have. For that, the flexible, crane-based methods we use make more sense.
Wrap-Up
It’s easy to see a viral video out of China and think “Wow, we’re screwed.” It fits the meme that China is way ahead of us.
The truth is usually more complicated. There’s a reason America is still much richer than China — we’re doing a lot of things right.
Dig into the facts behind these viral clips. You’ll often find that things aren’t quite what they seem.
People buy your product for 2 reasons: you make them money or you save them money. If you want to grow fast, make that value proposition crystal clear.
A clear value prop is one of the biggest things I look for when I meet with founders.
It needs to be as simple as one sentence. And it needs to be measurable.
Let’s look at two strong value propositions, using case studies from my portfolio…
Save Them Money
North is one of my most successful investments. They plug into your cloud stack to optimize your spending. North can cut your cloud bill by 50%.
That could save a company hundreds of thousands or even millions of dollars a year. All that money flows straight to the bottom line.
“We’ll take one of your biggest costs and cut it in half.” That’s a clear value prop!
Make Them Money
Saving customers money is great. But you can also help them make more money.
That’s what LedgerUp, another great investment of mine, does. LedgerUp automates your invoicing to make sure you collect all the revenue you’re owed.
Founder Joseph Johnson recently worked with a startup that was missing $900,000 in revenue. They had booked the sales but never collected!
Believe it or not, this is super common. Joseph helped them collect that extra $900,000, a huge boost for the startup.
“Pay us a modest SaaS fee and we’ll bring in tons of revenue.” I’m reaching for my wallet right now!
Weak Value Props
Saving you money or making you money are strong value propositions. They’re clear, measurable, and compelling.
For a lot of startups, the value prop is much less clear.
“We enable better team collaboration.” Okay, but how do I measure that?
“We help improve employee satisfaction.” Great, but how does that impact my bottom line?
If your product gives the customer a vague benefit they can’t measure, you’ve got a weak value prop. And you’re going to struggle in sales.
The problem is, the cost of your product is always measurable! I know it’s $1,000 a month, but I don’t know what benefit I’m going to get from it.
At that point, I’m out.
Wrap-Up
Startups win when they provide value to customers. You want that value to be clear, measurable, and large.
Save the customer money. Make them more money. Pick one and lean into it.
I have 34 investments so far. Over and over, the startups with clear value props have grown fast. The rest haven’t.
Practice getting your value prop down to a single, clear sentence. Then work hard on delivering exactly what you promised.
If you can do that, your only problem will be keeping up with the inbound!
There will be no blog tomorrow. I’m speaking at the Single Family Office Summit in New York City. If you’re there, come say hello!
They wanted $100,000, minimum. I politely bowed out. This is how startups sabotage their funding rounds.
How Investment Minimums Work Against You
Some startups insist on a minimum check size. This excludes helpful investors and makes it harder to close the round.
At $100,000, this startup was excluding most angels.
A typical angel check is around $25,000. Smaller checks are common as well. I write $5,000 checks direct on the cap table all the time.
I recently was in a deal with a very high profile, billionaire angel investor. Even that investor only put in $250,000!
That gives you an idea of how rare the big angel checks really are.
Small Investors Are the Key to Closing Your Round
With most angels excluded from the round, this founder wasn’t just losing out on one small angel check. He was losing crucial help to close his funding round.
Freshpaint, a YC startup that raised a seed round recently, is a great example.
The smallest check they took was $5,000. That small investor did more than any other person to help close their round.
The small fry made a series of introductions. This chain of intros led to $700,000 in investments!
How I Help Startups Raise Money
Okay, time for me to brag…
I can actually beat that $700,000 number. I introduced a founder to a VC fund that gave him an $800,000 term sheet.
This is the impact of meeting small investors and taking their money: you get access to their network. That network is crucial to closing your funding round.
Without those helpful angels, you’re stuck cold messaging VC funds.
Guess what? They get 1,000 of those messages a day. They probably won’t even read yours.
But What About My Cap Table?
I can imagine what you’re saying on the other side of this screen: but Francis, it’s bad to take small checks! What about my cap table?
The reality is, there are dozens of names on the typical startup’s cap table, even at Seed or Series A. If they’re helpful, one more won’t hurt you.
See if that small investor makes some useful introductions for you. If so, let him into the round.
And if you’re really worried about your cap table, you can pop up a Roll-Up Vehicle (RUV) for angels. This is a simple LLC that bundles all those angel checks into a single line on the cap table. You even get a free RUV if you use AngelList Stack, a cap table product.
Wrap-Up
Some founders have gotten into their heads that they should only take big checks. Nothing could be further from the truth.
Excluding the small checks cuts you off from helpful people. Those helpful little angels are the key to closing your funding round.
Meet with all sorts of investors, big and small. And if someone’s helpful, take their check, even if it’s tiny.
Pretty soon, you may find yourself raising millions.
Tremendous 