“The biggest understanding gap I see between founders and VCs today is this understanding of the relationship between the investor focus on terminal outcome and the founder focus on the microeconomics and unit economics….. The net result is a lot of of frustrated founders who don’t understand why they can’t raise with $1-2 million in ARR and investors who don’t understand why founders don’t realize they are in small markets, regardless of early traction.”
As Charles also notes, this is exacerbated by the rapid increase in venture fund size. Every dollar increase effectively needs another several dollars of startup exit value to justify the AUM growth.
Just articles, posts and thoughts that I’ve found interesting
falling asleep on the beach while reading a book, anime style art
Crooks’ Mistaken Bet on Encrypted Phones (New Yorker) – How European police have cracked “safe” encrypted phones often used by criminals, and the wealth of data it’s provided. Come for the tech story and stay for insight into why cocaine is huge in Europe, how smuggling logistics work, and the slang used to describe murder.
The Dangerous Rise of ‘Front-Yard Politics’ (The Atlantic) – Derek Thompson on why obsessing over slogans and words (and the performative display of them), is further distracting us from getting stuff done and creating artificial conflict.
“To get big-brained about it, something like La Sombrita could only happen in a high-regulation/low-trust society like the US. In every other variation (low regulation/high trust, high regulation/high trust, low regulation/low trust) you get either larger public works without fear of vandalism or misuse (a proper bus shelter), or like in Quito (a lower regulation society) you get natural ad hoc bottom-up solutions.”
Got a lot of reactions to my last post about acquihires (specifically, lack of them) and how that might change the approach startups take to pursuing ‘soft landings.’ Wanted to clarify and respond to some of those questions, backchannels, etc.
It’s often ok to just shutdown. I wasn’t suggesting that every founder/team/investors would prefer an acquihire to other forms of wind down, and definitely not that founders always “owe” their investors this attempt. If anything I was trying to emphasize to founders that it’s ok to try a different process, often against the common wisdom.
To founders (some in our portfolio) who assumed I was specifically subtweeting them. Nope. Of course the suggestion was spurred on by what I’m observing in the market – including a lot of stories from founders we haven’t backed – but I was commenting on the market at large.
I don’t want to fund an acquihire marketplace. If you want to build a marketplace for small acquihire transactions go right ahead but my post wasn’t a Request for Startup 🙂
Some people disagreed with me! “Nah, you give up all leverage when you do what you recommended,” was the feedback from a few readers. My POV is that it’s more nuanced than this. You have no leverage – or at best fake leverage aka bluffing – when you don’t have alternatives. Go have some private conversations, float some trial balloons, etc. But I truly believe the strategy I detailed is UNDERUSED in our industry.
Makes the Founder Seem Like a Loser? No way. I think it’s artful and thoughtful when done correctly. All of this is changing IMO – analogous might be layoff lists these days – no shame in being included in one when you know you did the best you could but the company had to make cuts. Obviously more falls on the CEO/founder shoulders but again, I think my strategy is going to be self-selecting for the type of leaders who authentically can articulate the situation and has some value to transact.
Appreciate the discussion – it’s why I enjoy blogging 🙂
“Worst case scenario we’ll sell to a larger startup or public company for about ~$1.5m per engineer.” Yes, this was the ‘fallback plan’ for many team in the web2 era and they weren’t wrong. Especially in the early days of mobile/iOS engineering, if you hired strong technical talent into your early stage company, you basically created an acquisition outcome floor. I was on both sides of these transactions – buying startups for Google/YouTube and angel investing in high quality technical founders. Sometimes you’d even get lucky and receive stock in the acquirer, which was how I gained pre-IPO equity in high growth stars like Pinterest and Facebook.
Starting our venture fund Homebrew professionalized and scaled my insights into soft landings. Acquihire potential absolutely isn’t enough in and of itself to justify venture funding (we play to win!), but in certain situations investors do talk about these things as positive optionality. And during our first few years we leaned in to help teams find the right home when it didn’t work out for them as an independent company. This produced two successful intra-portfolio acquisitions where one team joined a larger startup we previously seeded (Chime and Bowery Farming were the buyers) and a whole bunch of other transactions. The proverbial win-win-win: founders got to land their company often with some retention premium; employees got job offers; and we got capital back, that even if it wasn’t a power law return, allowed us to recycle into new investments or the existing portfolio. I’d say that for a small, two person fund we got pretty good at this motion when needed!
And now I’m telling you the world is different. Very different.
a busy city intersection with lots of giant billboards and people walking, at night, digital art [DALL-E]
In 2023 with few exceptions acquihires are dead as we knew them. The majority of typical acquirers (large and small) don’t have incremental headcount budget. Those who do, often believe they can hire from the open market without the hassle of an acquisition. Cash is at a premium so it’s not going to cap tables (preferred or common walk away from the deals with no dinero). In fact, sometimes acquirers are asking for the remaining cash on hand from the startup in order to ‘zero out’ the salary burden they’re taking on [HW note: 99.9% of the time my answer is no fucking way]. And when they’re giving stock to existing shareholders instead of cash it’s at high 2021 valuations, buried below a preference stack.
None of this means we’ve backed off helping founders in these situations, but we do try to set expectations with them and collaborate with the other investors. My personal rule of thumb is that to the extent there’s cash or valuable IP still in the company, we need to make sure that we’re good stewards of those assets (per above, why I balk at giving up cash in an acquisition where there’s little bidirectional value exchange). But when it comes down to the forward-looking time of the founders and team – eg do they actually want to go work at the potential acquirer – their opportunity cost and happiness is really important. No founder should feel compelled to sign up for four years of earn out misery just to get their venture investors a few cents on the dollar.
Times like these call for somewhat different strategies, perhaps shifting from the ‘companies are bought not sold’ mindset (which is very much true in situations where the startup has optionality or at least competitive offers). My counterintuitive suggestion is that more founders should publicly announce they need to find a home when seeking this outcome. Put together a great post or deck about the situation, quality of the team, what they know how to do better than anybody else, and why they’ve had trouble raising additional capital. Let potential acquirers find you (who knows you might even end up with some funding offers). It’s sort of a litmus test – if you can’t make the argument convincingly in public I’m suspect you’re going to somehow magically figure it out in a quiet, closed door process. Not in today’s market conditions.
Downsides? Emotional I guess. But really, “this didn’t work out the way we hoped” is the theme song of startups so join the chorus.
Giving up negotiating leverage with a potential acquirer? Again, not really in this market. The only way you get to negotiate is if you have a BATNA, and my POV is this will increase that likelihood for 80% of companies in this position. So go talk with a few of your most promising relationships first, but don’t hesitate to go wide when you’re not getting immediate traction.
Some VC with an operations team should go build out the template for this – make it easy for founders and normalize this process, removing any stigma. Instead of spending your last quarter of existence digging through haystacks for needles, build a magnet, and pull the needles towards you. If over the course of the next year you see any Homebrew portfolio company try this out, I’ll let you know! And good luck, it’s rough out there.
An AI Safe Harbor Provision Would Create Guidelines For Development & Safety Without Premature Regulations
The conversation around Artificial Intelligence has started to take on a binary quality, rather prematurely, as if we were debating the two sides of a coin rather than a more complex shape. “Let builders build as is” vs “Regulate.” Ironically, both positions are outputs of acknowledging the incredible early power and promise of the tipping point we’ve reached, but neither incorporate the ambiguity. Fortunately there’s some case law here which might help, and we only have to go back to earlier Internet days and the concept of safe harbor.
a 19th century sailing ship, with robot sailors on deck, docking in a harbor, cyberpunk [DALL-E]
‘Safe harbor’ is a regulatory framework which provides that certain conduct won’t break a rule so long as specific conditions are met. It’s used to provide clarity in an otherwise complex situation, or to provide the benefit of the doubt to a party so long as they abide by generally acceptable reasonable standards. Perhaps the most well-known example in our industry is the 1998 Digital Millennium Copyright Act (DMCA) which provided safe harbor to Internet businesses around copyright infringement performed by their end users so long as several preconditions were met (such as direct financial benefit, knowledge of infringing materials, and so on).
The DMCA allowed for billions of people globally to express themselves online, prompted new business model experiments, and created guardrails for any entrepreneur to stay legal. It’s not perfect, and it can be abused, but it met the reality of the moment in a meaningful way. And it made my career possible, working with user generated content (UGC) at Second Life, AdSense, and YouTube. During my time at the world’s largest video site, I coined the ongoing public metric ‘# hours of video uploaded every minute” to help put YouTube’s growth in perspective and frame for regulators how unfathomable and unreliable it would be to ask human beings to screen 100% of content manually.
Now 25 years later we have a new tidal wave but it’s not UGC, it’s AI and, uh, User Generated Computer Content (UGCC), or something like that. And from my point of view it’s a potential shift in capabilities as significant as anything I’ve experienced so far in my life. It’s the evolution of what I hoped — not software eating the world, but software enabling it. And it’s moving very very quickly. So much so that it’s perfectly reasonable to suggest the industry slow itself, specifically stop training new models while we all digest the impact of the change. But it’s not what I’d advocate. Instead let’s speed up creating a temporary safe harbor for AI, so our best engineers and companies can continue their innovation while being incentivized to support guardrails and openness.
What would an AI Safe Harbor look like? Start with something like, “For the next 12 months any developer of AI models would be protected from legal liability so long as they abide by certain evolving standards.” For example, model owners must:
Transparency: for a given publicly available URL or submitted piece of media, to query whether the top level domain is included in the training set of the model. Simply visibility is the first step — all the ‘do not train on my data’ (aka robots.txt for AI) is going to take more thinking and tradeoffs from a regulatory perspective.
Prompt Logs for Research: Providing some amount of statistically significant prompt/input logs (no information on the originator of the prompt, just the prompt itself) on a regular basis for researchers to understand, analyze, etc. So long as you’re not knowingly, willfully and exclusively targeting and exploiting particular copyrighted sources, you will have infringement safe harbor.
Responsibility: Documented Trust and Safety protocols to allow for escalation around violations of your Terms of Service. And some sort of transparency statistics on these issues in aggregate.
Observability: Auditable, but not public, frameworks for measuring ‘quality’ of results.
In order to prevent a burden that means only the largest, well-funded companies are able to comply, AI Safe Harbor would also exempt all startups and researchers who have not released public base models yet and/or have fewer than, for example, 100,000 queries/prompts per day. Those folks are just plain ‘safe’ so long as they are acting in good faith.
Within the 12 month period, AI Safe Harbor would be extended as is; modified and renewed; or eliminated for general regulations. But the goal is to remove ambiguity + start directing companies towards common standards (and common good), while maintaining their competitive advantages locally and globally (China!).
Why Tech Companies Avoid Customer Service and the Opportunity That Comes With Actually Engaging Your Users
Thirty years from now when you’re reading my memoir pay attention to Chapter 8 because that’s when I became President of the United States. The populist momentum that resulted in an unprecedented third-party ascension was all based on a single premise: the large tech companies should staff competent, responsive and empathetic customer service departments.
a telephone locked inside a clear plastic box, digital art [DALL-E]
My YouTube video went viral. Where I picked up a Yellow Pages, looked straight into the camera, and ranted about how we can get a locksmith on the phone, the local supermarket will pick up when it rings, even (with a little bit of effort) my doctor. But try to call Google. Try to call Facebook. Try to call most of the tech companies no one is there to pick up. Send them an email or file a ticket? Good luck. We are dependent on them for our lives and our businesses, and we make them billions of dollars, and their employees wealthy. But they won’t help us navigate through this new world they’re creating. And that’s why I’m challenging our government to regulate them. Not about monopolies or privacy or copyright, but customer service. They want DMCA safe harbor? They want Section 230? Well I want someone to answer the GD phone!!!!
Snapping out of my daydream where our fractured country is united behind the idea of 1–800–4GOOGLE (by the way, in the Presidential fantasy my VP led a grassroots uprising for standardized charging plugs — their logo was a Guy Fawkes and USB-C plug), I do want to seriously suggest that one way for our industry to improve its standing with average consumers and small business owners is to be more user friendly when those folks have questions. Through my years at Google and YouTube I heard from lots of people about how much they loved our software but when something went wrong (locked out of an account, wrong information on their business listing, confusion around advertising) they went down a rabbit hole trying to get an answer from our company. And didn’t understand why these powerful corporations couldn’t afford to try and help their customers/users figure out this new world together. It was, and still is, a good question! I think there are four answers:
I. Software Margins
It costs money to staff support team (duh) and if you’re not showing a high margin structure you risk being penalized from an enterprise value multiple. The pernicious impact of striving for ‘software margins’ means that support is typically a cost center to be minimized, rather than a point of excellence that’s invested in and rewarded.
II. Humans Don’t Scale
Sure you can help human workers be more productive over time but they’ll never be as efficient as software automation or customer self-service. “Won’t scale” is historically a way to kill any idea, even if it, for the meantime would make a situation better (obviously there are exceptions to this when the stakes are really high). I’m sure AI-driven chat, etc will be a boon here too. But sometimes it’s not just about an answer, it’s about feeling respected and served.
III. Engineering Stereotypes Create Permission Structure
How do you tell an extroverted engineer from an introverted one? The extroverted engineer looks at *your* shoes when he talks.
While many of the engineers I’ve know are perfectly sociable, well-adjusted, highly conversant people, the ‘sullen hacker in a hoodie on the spectrum’ is anywhere from an antiquated stereotype to a true segment of our community. And either way it lets too many of us get away with not having to deal with the actual implications of the products we build. Because we’re not asked to serve on the front lines of our businesses hearing the challenges real users are facing. Rotate everyone through the support queues periodically is my solution.
IV. Elites Get Special Treatment
Maybe the real reason these issues don’t get solved is that the 1% have their backdoors into these companies. You’re a big enough advertiser or business partner to have an account manager. You went to grad school with the COO. And so on.
That’s why one of my periodic troll tweets was something to the extent of “I don’t know why everyone says [Instagram, YouTube, Google, etc] has such terrible customer support. Whenever I have a question I just email the VP of the product and they respond really quickly.”
And there you are, Chapter 8 of the autobiography.
Besides my random power fantasies, this post was prompted by a discussion with a very smart marketing and comms exec in the wake of the SVB bank run. The conversation evolved to one about how our industry (venture, startups, tech in general) could better message about the positive role we play in the economy. I quipped that besides good phrasing we needed actions too. When she asked what would I recommend my response wasn’t about getting rid of carried interest or breaking up the big companies but about customer support. Why?
There are going to be people who believe capitalism is flawed — we won’t win them.
There are going to be people who long for a world where things moved slower and they didn’t have to deal with disruption and could keep the status quo because it serves them better — shrug emoji.
There are going to be people who use tech as a punching bag when convenient to further their own goals — we should stand up to them.
But there’s an even bigger percentage of average Americans, who *like* technology and find many of the companies aspirational. These citizens, these business owners, these leaders — we have the opportunity to show them we can help them navigate the new world that we are helping them build. If takes a few margin points and some empathy it might be healthier and more sustainable than just lobbying and tweeting. The bigger structural issues need examination and *some* regulation, but there’s lots we can do on our own.
A Ticket From Michael Jordan’s First NBA Game Auctions Big $$$, Adam Sandler Gets Respect (Finally!), and Why Germany’s Envy of Silicon Valley Helped Wirecard Commit Massive Fraud
“The nature of comedy is you get the audience, you get the money,” says Michaels, SNL’s creator and executive producer. “Respect is the last thing you get.”
The ticket made him famous with his neighbors. They already liked Mike Cole anyway for his odd place in the equilibrium of the cul-de-sac, the tall, bald guy who waited too long to shovel his driveway when it snowed and then slipped on the ice while coming back from retrieving the mail.
When the country’s mining industry collapsed, a criminal economy grew in its place, with thousands of men climbing into some of the deepest shafts in the world, searching for leftover gold.
A hybrid of Craigslist, eBay, Reddit and Sotheby’s that facilitated $1.37 billion in sales last year was not quite what Randy Nonnenberg had in mind when he started a car blog with a college buddy as a hobby.
How the Biggest Fraud in German History Unravelled. (Ben Taub, New Yorker) —Wirecard was a massive multibillion dollar fintech fraud. And at times it seems like German authorities were willing to look the other way, just out of hopeful pride that their country had finally launched a unicorn startup.
On February 18, 2019, Germany’s financial regulator, known as BaFin, issued a ban on creating new short bets against Wirecard, citing the company’s “importance for the economy.” “It was at that moment that they sided with criminals,” a German parliamentarian later said. The same day, prosecutors in Munich confirmed to a German newspaper that they had opened a criminal investigation. But they weren’t going after Wirecard — they were going after the F.T. [for an investigative takedown piece about Wirecard]
Why Lack of Easy Options Made Me Focus On Who I Wanted To Be, Which Paid Off Over Time
a very sad businessman holding a popped balloon, digital art [DALL-E]
In retrospect the fact we were all day trading tech stocks from Stanford’s computer labs probably suggested it was a bit of a bubble, although eToys options did pay for two consecutive Spring Break vacations. I was getting my MBA at the time which in some ways wasn’t just part of the DotCom storyline but an epicenter. Our professors were literally rewriting the case studies in real time and my participation in the very first Internet Marketing class the GSB ever offered is a form of carbon dating that conclusively proves my old age. But by my graduation in June of 2000, the party had ended. As became clear quickly: the Stanford Business School Class of 1998 had founded the good Internet 1.0 companies; the Class of 1999 had founded the bad Internet 1.0 companies; and the Class of 2000 was just plain unemployed. And so I left the campus with student debt and limited prospects. But it turned out to be exactly what I needed.
My decision to attend business school wasn’t really about getting the credential. I was there to get a MBA, not be an MBA. In fact, as a classic liberal arts major, I found myself more attracted to the PhD students studying theory than the curriculums built around understanding practice. And while 25 years ago a business school campus was a compelling way to build a professional network (there are many other methods now), I had other reasons for being there: to figure out who I was. Or maybe, more specifically, to give myself confidence to be who I wanted to be.
Right brain mother and left brain father left me confused… which one was I??? Pre-Stanford that meant trying out jobs to seeing what fit. Feed the left by working on Late Night with Conan O’Brien and be the only one geeking out about an Access database to track guests. Pivot to the right with a few years of management consulting and use Powerpoint on cross country flights to make stop motion animations. Neither felt perfect but my work ethic wouldn’t allow me to just stop and figure it out. But business school? Maybe that was a chance to pause and examine myself while still ‘moving forward’ in my career. So I applied to the one school that felt like a good fit and crossed my fingers.
tldr IT WORKED! Came out of Stanford with a mission: to work on products which mattered to me, with people I could learn from, and to feel like I was making a real difference to the outcome. All I needed was a job. Then this happened:
NASDAQ 100
It wasn’t even the lowest moment of the crash as the stock market dead cat bounced a few times before heading lower for the early 2000s. Graduating into the bubble burst was no fun but had one very important benefit for me: there was no temptation to take a ‘get rich quick’ job. Over the previous years you could spend 12–24 month at a startup and become an IPO millionaire. Give a shit about what you were actually building? Is it sustainable? Who cares! You could justify getting in and then out and doing the important work later, with a swollen bank account. Now to be clear, this wasn’t everyone. There were absolutely founders and teams who selected areas that meant a lot to them personally, but the rose colored glasses were structural.
Would I have been tempted by this path? Absolutely! I was broke and needed money (had been working part-time at an enterprise tech startup along the way but otherwise savings were depleted). Man plans, God laughs as my wife says. No more jobs. Like overnight lots and lots of contraction, hiring freezes, and pulled offers. Sort of like, well, 2023.
The void of temptations meant I could stay the course and stick to first principles. It wasn’t always easy but I had a roommate, a girlfriend, a little bit of the income mentioned above, and some resolve. When six months later my networking connected me with Philip Rosedale and Linden Lab (the startup building virtual world Second Life), I knew I’d found the product I wanted to build, the people I wanted to build it with, and somewhere I though I could make a difference. And that’s where I spent almost three years. Linden Lab wasn’t an overnight success, not did it fulfill its potential, but it started me in the right direction. Which led to Google, then YouTube, and finally Homebrew.
The point here isn’t to sugarcoat the difficult time of a downturn. Or to be blind to the multiple advantages I had going into the DotCom failure. But I do believe that manic bull markets tend to cause people to make decisions which are clouded by lure of easy wealth, or herd momentum. If there’s a silver lining to downturns it’s they provide all sort of clarity, even at the individual level.
Riffing Off VC Charles Hudson’s Blog Post, Here’s What I’m Trying to Answer
a group of different robots running a race, digital art [DALL-E]
If startup founders sometimes ‘Build in Public,’ is the analogou sventure capitalist motto to ‘Think in Public?’ Anyway, there’s no doubt that the story of the trailing months has been Artificial Intelligence. Over Homebrew’s first decade we’ve always been interested in what we’ve called ‘Applied AI’ (along with Applied CV, Applied ML)— opportunities where the technology itself was being extended and commercialized for a specific purpose (contrasted with core R&D or base model development). Companies such as Shield.ai, Kettle, and MasterfulAI, among many others, were Homebrew investments which fit this definition. But it’s also clear we’re at a new inflection point where our previous hypotheses needed to be updated. So like a stone in a polishing tumbler, ‘what are our principles here’ had been tossing around my head for a handful of quarters. And then I read Charles Hudson’s post, which prompted me [AI PUN] to just write this down.
Given team, data, and compute costs, will the ‘price of entry’ and ‘price of innovation’ on base models increase or decrease over time
Will different data types produce/require their own base models, and under what conditions are these base models likely to be produced by different companies/sources vs under a single corporate umbrella
How does one measure ‘quality’ and what characteristics will base model owners compete on besides ‘quality’ [price, latency, privacy, etc]
II. AI ‘Middleware’
In a multi-base model world, won’t there be some value created by dynamically switching between models depending on the use case? Won’t most application owners who seek to integrate ‘AI’ be interested in “best results” more so than having to choose a model upfront
Will this middleware layer have access to enough model attributes to even know when/how to manage between models
Can these companies protect their margins or will they be subject to either (a) intense competition pushing margins down to ‘base model query price + a few basis points or (b) the base model companies behaving like the record labels and basically being very deliberate about taking the majority of revenue created by a service built on top of their IP
Will middleware be able to augment the base models with new proprietary data in order to create a differentiated product
Will middleware companies seek to aggregate proprietary data sources in order to improve base models in unique ways
III. AI ‘Native’ Applications
What are the conditions by which the addition of AI catalyzes new product offerings built around this technology versus ‘AI’ being a feature that the market leading applications can build into their platforms. Will Zendesk be replaced by an AI Customer Support startup or does Zendesk integrate AI. Repeat this question for everything B2B.
OpenAI is a for-profit, running a venture fund, etc — what types of ‘partnership risk’ is there in backing alternatives who are competing with OpenAI funded startups. Are all the base models doing to use their cash to try and develop their own ecosystems and implicitly/explicitly try to pick winning apps?
What will the engineering teams at ‘non-native’ adopters need to look like in order to successfully integrate, manage and compete with the native apps
Will businesses who believe they have proprietary data to may help improve base models be able to sell that data and/or ‘pay it in’ to the model in return for discounted usage? Will they seek to create improved layers atop the base models
If you have POVs here I’m always happy to hear from you [hunter at homebrew dot co]! Remember, we invest our personal capital (typically a $100k-$500k initial investment, although ability to go larger when appropriate) in your companies and then get to work supporting you.
That’s why I’m participating in SeedChecks, an experience by which you can submit your pitch deck to a group of early stage investors without having to know any of us already. Everyone looks at your information separately and follows-up if there’s potential for mutual fit. It’s not a groupthink exercise and there’s no collusion, although hopefully maybe some collaboration as I’ve personally co-invested with many of the folks in the group.
Will opportunities like SeedChecks be the death of the warm intro? Of course not, but we’re all committed to the idea that now more than ever, great founders can be anywhere and anyone. So please consider submitting your materials if you’re looking to raise your first round of capital.
