A reminder for new readers. That Was The Week collects the best writing on critical issues in tech, startups, and venture capital. I selected the articles because they are of interest. The selections often include things I entirely disagree with. But they express common opinions, or they provoke me to think. The articles are only snippets. Click on the headline to go to the original. I express my point of view in the editorial and the weekly video below.

This Week’s Video and Podcast:

Content this week from @kteare, @ajkeen, @MikeNayna, @HowardMarksBook, @Noahpinion, @readmaxread, @krishnanrohit, @Alex_Lazarow, @chudson, @ManuKumar, @gruber, @Jessicalessin, @KateClarkTweets, @ttunguz, @mslopatto, @waxeditorial, @VinodJason, @julipuli, @vkhosla


Editorial: Capitulation?

Essays of the Week

The Westminster Declaration

Fewer Losers, or More Winners?

Thoughts on techno-optimism

“Techno-optimism” is a sign of V.C. crisis

Slouching Towards Utopia

What does it take to become a unicorn founder? Insights from 200 startup journeys

The Big Reset in Seed to Series A Graduation Rates is Real and Permanent

State of Venture in Fall 2023

Video of the Week

Jamie Rhode on Venture Capital

AI of the Week

Apple and AI

Amazon’s Andy Jassy Plans to Crash the AI Party

Thrive Capital to Lead Purchase of OpenAI Employee Shares at $80 Billion-Plus Valuation

AI Reaccelerating Cloud Growth

Sam Bankman-Fried is going to talk himself right into jail

News Of the Week

Threads Is Closing in on 100M Users Says Zuckerberg

Apple TV+ Show ‘The Problem With Jon Stewart’ Reportedly Canceled

BRICS: Chinese Yuan Surpasses Euro, Becomes Second Main SWIFT Currency

Startup of the Week

Hands-On With the New Journal App in the iOS 17.2 Beta

X of the Week

Vinod Khosla on AI and Copyright

Editorial: Capitulation?

Charles Hudson and Manu Kumar are two of the best early-stage investors in the world. This week, they both penned meaningful essays declaring that the fundamentals of early-stage investing have changed, possibly permanently.

Charles states:

For the past 18 months, the Series A market has been very quiet. Outside of AI-related investments, it feels like deal volume is off 75%. The Series A investors I know don’t feel any pressure to make investments and don’t really seem that excited or interested in much these days. Unfortunately for seed-stage companies, the Series A market can remain on strike longer than most seed-stage companies can remain solvent.

He predicts a significant dip in the number of seed-stage companies doing a Series A, a drop in Series As, and fewer extensions or bridge rounds.

Manu Kumar of K9 Capital extends the theme and suggests only companies building real businesses with real customers and revenue will survive. He implores companies to understand the following:

Early stage venture, particularly Pre-Seed and Seed stage venture, is a different game today than it used to be 10 years ago. LPs and GPs should be aware of this dynamic as they make investment decisions.

This follows from Sam Lessin’s theme a couple of weeks ago, declaring that the era of larger checks coming in at later rounds is now over.

At Signalrank, we have always distinguished between organic unicorns, built over many years from the seed stage, and artificial unicorns created by a single large check in an early round. Organic is always best and can survive downturns.

In this week’s video of the week, Jaimie Rhode doubles down on the theme, explaining that in venture capital, only power-law companies matter. Those companies grow to valuations, enabling an entire fund to be returned or more. The venture capital industry relies on the few power law winners returning invested capital.

In the heady days of 2019-2022, it was possible for a power-law company to emerge fast due to a single round of financing. And then quickly go on to do two or three more rounds within a year or two. Charles and Manu correctly point out that this is unlikely, except perhaps in AI. A power law winner will have to be built organically.

The implication is that early-stage investing will need to become once again deliberate and patient. I use the word capitulation (with a question mark) in the title, but this is just a recognition of reality. Capitulation to reality is a good thing. So, no, it is not a capitulation.

But it begs the question, can venture capital support wealth creation if it is really a lottery for a power law winner? Is there a way to benefit from the growth in venture-backed companies without needing to play the lottery of picking individual companies?

Then I read Howard Marks, the CEO at Oaktree, in Fewer Losers, or More Winners?

I feel strongly that attempting to achieve a superior long-term record by stringing together a run of top-decile years is unlikely to succeed. Rather, striving to do a little better than average every year – and through discipline to have highly superior relative results in bad times – is:

less likely to produce extreme volatility,

less likely to produce huge losses which can’t be recouped and, most importantly,

more likely to work (given the fact that all of us are only human).

Simply put, what [General Mills’s] record tells me is that, in equities, if you can avoid losers (and losing years), the winners will take care of themselves. I believe most strongly that this holds true in my group’s opportunistic niches as well – that the best foundation for above-average long-term performance is an absence of disasters.

Marks is a long-term believer in consistent average returns. Usually, this requires indexing, where investors can invest across an asset class and enjoy returns at or close to the market norms. If, on top of that, you can avoid losers, then you will beat the market.

SignalRank does exactly that but in the venture capital market. Marks writes only about public markets, bonds, credit, and debt. If it is possible to deliver better-than-average returns from venture investments by de-risking them, then that would be a revelation. Series B rounds return about 3x over 5 years if it were possible to invest in all of them. Here are all venture-backed Series B companies for 2012 onwards, with performance (scroll right in the table to see all fields).

It is possible to remove likely losers (SignalRank only selects 2800 out of 20,000 B rounds since 2012). By doing so, the group averages over 5x return in five years and picks close to 30% unicorns compared to 10% in the market.

Avoiding losers produces a higher proportion of winners. Howard Marks is right. This has never before been possible with private venture-backed companies, but it now is. We can now make smart indexes in the venture capital asset class.

So, the capitulation implied in Sam Lessin, Manu Kumar, and Charles Hudson’s pieces is understandable. And correct. if you play the normal venture game then finding a power law winner is the meaning of success. Until then, dry powder should remain dry.

But if you can play the Howard Marks version of venture capital, then continuing to invest makes sense, across a broad range of likely winners and eliminating likely losers. Building an index-like portfolio of likely high performers produces predictable and excellent rewards.

Essays of the Week

The Westminster Declaration


OCT 19, 2023

I’ve joined 138 artists, journalists, and public intellectuals from around the world to warn of a shortsighted perspective on free speech and digital governance that’s mobilised social media companies, universities, governments, and NGOs to derange the free-flow of information.

Lost – Roswitha Schleicher-schwarz

The document I’ve signed, along with Matt Tiabbi, Tim Robins, and Julian Assange, among others, is called The Westminster Declaration and its text is as follows.

The Westminster Declaration

We write as journalists, artists, authors, activists, technologists, and academics to warn of increasing international censorship that threatens to erode centuries-old democratic norms.

Coming from the left, right, and centre, we are united by our commitment to universal human rights and freedom of speech, and we are all deeply concerned about attempts to label protected speech as ‘misinformation,’ ‘disinformation,’ and other ill-defined terms.

This abuse of these terms has resulted in the censorship of ordinary people, journalists, and dissidents in countries all over the world.

Such interference with the right to free speech suppresses valid discussion about matters of urgent public interest, and undermines the foundational principles of representative democracy.

Across the globe, government actors, social media companies, universities, and NGOs are increasingly working to monitor citizens and rob them of their voices. These large-scale coordinated efforts are sometimes referred to as the ‘Censorship-Industrial Complex.’

This complex often operates through direct government policies. Authorities in India1

and Turkey2 have seized the power to remove political content from social media. The legislature in Germany 3 and the Supreme Court in Brazil4 are criminalising political speech. In other countries, measures such as Ireland’s ‘Hate Speech’ Bill5

, Scotland’s Hate Crime Act6 , the UK’s Online Safety Bill7, and Australia’s ‘Misinformation’ Bill8 threaten to severely restrict expression and create a chilling effect.

But the Censorship Industrial Complex operates through more subtle methods. These include visibility filtering, labelling, and manipulation of search engine results. Through deplatforming and flagging, social media censors have already silenced lawful opinions on topics of national and geopolitical importance. They have done so with the full support of ‘disinformation experts’ and ‘fact-checkers’ in the mainstream media, who have abandoned the journalistic values of debate and intellectual inquiry.

As the Twitter Files revealed, tech companies often perform censorial ‘content moderation’ in coordination with government agencies and civil society. Soon, the European Union’s Digital Services Act will formalise this relationship by giving platform data to ‘vetted researchers’ from NGOs and academia, relegating our speech rights to the discretion of these unelected and unaccountable entities.

Some politicians and NGOs9are even aiming to target end-to-end encrypted messaging apps like WhatsApp, Signal, and Telegram.10

If end-to-end encryption is broken, we will have no remaining avenues for authentic private conversations in the digital sphere.

Although foreign disinformation between states is a real issue, agencies designed to combat these threats, such as the Cybersecurity and Infrastructure Security Agency in the United States, are increasingly being turned inward against the public. Under the guise of preventing harm and protecting truth, speech is being treated as a permitted activity rather than an inalienable right.

We recognize that words can sometimes cause offence, but we reject the idea that hurt feelings and discomfort, even if acute, are grounds for censorship. Open discourse is the central pillar of a free society, and is essential for holding governments accountable, empowering vulnerable groups, and reducing the risk of tyranny.

Speech protections are not just for views we agree with; we must strenuously protect speech for the views that we most strongly oppose. Only in the public square can these views be heard and properly challenged.

Fewer Losers, or More Winners?


SEP 12, 2023

My memos got their start in October 1990, inspired by an interesting juxtaposition between two events. One was a dinner in Minneapolis with David VanBenschoten, who was the head of the General Mills pension fund. Dave told me that, in his 14 years in the job, the fund’s equity return had never ranked above the 27th percentile of the pension fund universe or below the 47th percentile. And where did those solidly second-quartile annual returns place the fund for the 14 years overall? Fourth percentile! I was wowed. It turns out that most investors aiming for top-decile performance eventually shoot themselves in the foot, but Dave never did.

Around the same time, a prominent value investing firm reported terrible results, causing its president to issue an easy rationalization: “If you want to be in the top 5% of money managers, you have to be willing to be in the bottom 5%, too.” My reaction was immediate: “My clients don’t care whether I’m in the top 5% in any single year, and they (and I) have absolutely no interest in me ever being in the bottom 5%.”

These two events had a strong influence on me and helped define my – and what five years later became Oaktree’s – investment philosophy, which emphasizes risk control and consistency above all. Here’s how I put it 33 years ago in that first memo, titled The Route to Performance:

I feel strongly that attempting to achieve a superior long-term record by stringing together a run of top-decile years is unlikely to succeed. Rather, striving to do a little better than average every year – and through discipline to have highly superior relative results in bad times – is:

less likely to produce extreme volatility,

less likely to produce huge losses which can’t be recouped and, most importantly,

more likely to work (given the fact that all of us are only human).

Simply put, what [General Mills’s] record tells me is that, in equities, if you can avoid losers (and losing years), the winners will take care of themselves. I believe most strongly that this holds true in my group’s opportunistic niches as well – that the best foundation for above-average long-term performance is an absence of disasters.

As you can see, my dinner with Dave was a seminal event; his approach was clearly the one for me. (Incidentally, I want to share that after decades of not having been in touch, Dave was among the many kind people who wrote in recent months to encourage me vis-à-vis my health issue. This is a great example of the many personal dividends my career has paid.)

Putting It in Brief

That first memo, and the bit cited above, include a phrase you’ve likely heard from Oaktree: If we avoid the losers, the winners will take care of themselves. My partners and I considered this phrase so fitting that we adopted it as our motto when Oaktree was formed in 1995. Our reasoning was simple: If we invest in a diversified portfolio of bonds and are able to avoid the ones that default, some of the non-defaulters we buy will benefit from positive events, such as upgrades and takeovers. That is, the winners will materialize without our having explicitly sought them out.

We thought that phrase was innovative. But in 2005, while working with Seth Klarman to update the 1940 edition of Benjamin Graham and David Dodd’s Security Analysis – the “bible of value investing” – I read something that indicated we were late by about 50 years. In the section Seth asked me to edit, I came across Graham and Dodd’s description of “fixed-value” (or fixed-income) investing as “a negative art.” What did they mean?

At first, I found their observation cynical, but then I realized what they were saying. Let’s assume there are one hundred 8% bonds outstanding. Let’s further assume that ninety will pay interest and principal as promised and ten will default. Since they’re all 8% bonds, all the ones that pay will deliver the same 8% return – it doesn’t matter which ones you bought. The only thing that matters is whether you bought any of the ten that defaulted. In other words, bond investors improve their performance not through what they buy, but through what they exclude – not by finding winners, but by avoiding losers. There it is: a negative art.

One more anecdote concerning the origin of the phrase: I’ve always been interested in old books. A few years ago, while walking through a Las Vegas convention center on the way to meet with a client, I came upon a rare book fair. I stopped at the booth of a book dealer I know, and my eye immediately fell on a book he had for sale: How to Trade in Stocks, by Jesse Livermore. Here’s the quote the dealer had highlighted: “Winners take care of themselves; losers never do.” You may be tempted to believe Livermore borrowed my idea . . . until you realize that, like Graham and Dodd, he published these lines in 1940. So much for my innovation.

At the time I adopted that saying, my partners and I were primarily high yield bond investors. And since non-convertible bonds have little upside potential beyond their promised yield to maturity, it truly was the case that our main job was to avoid the non-payers, with the assumption that some subset of the payers would likely give us exposure to positive developments that occurred. It was an appropriate way to sum up our approach as bond investors.

But fortunately, I joined up with Bruce Karsh in 1987, and in 1988 we organized our first distressed debt fund. Now we were investing in bonds that had defaulted or seemed likely to do so. We thought we might be able to buy them at bargain prices because of the cloud they were under, giving us the possibility of capital appreciation. Bruce has since become well known for his investing acumen, and, certainly, his returns since 1988 can’t be attributed to the mere avoidance of losses. When you aspire to returns well above those available on bonds, it’s not enough to avoid losers; you actually have to find (or create) winners from time to time. The returns generated by Bruce and his group show that they’ve done so.

Oaktree now has a number of what I call “aspirational strategies,” meaning they need winners. So why do we still use the above phrase as our motto, and why is “the primacy of risk control” still the first tenet of our investment philosophy? The answer is we want the concept of risk control to always be top of mind for our investment professionals. When they review a security, we want them to ask not only “How much money can I make if things go well?” but also “What will happen if events don’t go as planned? How much could I lose if things get bad? And how bad would things have to get?”

Thoughts on techno-optimism

What it means to me, and why I support it.

NOAH SMITH, OCT 20, 2023

By Cory Doctorow from Beautiful Downtown Burbank, USA – The Sphere as Mars, view from my hotel room at Harrah’s, Las Vegas, Nevada, USA, CC BY-SA 2.0,

“No! Quite wrong! Open the gates!” — Baron von Munchausen

In my roundup this week, I endorsed Marc Andreessen’s “Techno-Optimist Manifesto”. I noted a couple points of disagreement, but overall I think this sort of uncompromising blast of techno-accelerationism is exactly the kind of extropian enthusiasm we need to shake us out of the doldrums of the gloomy 2010s.

But still, Marc’s manifesto doesn’t quite articulate how I think about techno-optimism; there’s a good deal of overlap, sure, but I think there’s a lot to be gained from reformulating the idea in my own words, so that when I say I’m a “techno-optimist”, people know exactly what that means.

Well, here is what I mean.

Different types of optimism: Positive vs. normative, active vs. passive

When people use the word “optimism”, they can mean various things, so I think it’s good to start with a taxonomy. First, when it comes to technological progress, there are basically two things you can be optimistic about:

You can think that there’s a lot of useful technology out there left to discover.

You can think that new technology will make the world better.

Borrowing a term from economics, I could call these “positive” and “normative” techno-optimism.

It’s obvious that these aren’t the same thing. For example, I often say I’m “optimistic” that there will be huge advances in autonomous drone technology over the next decade. But I also think that this will mostly be used for war, rather than for civilian applications like delivering packages. Now that isn’t necessarily bad news for humanity — a future where war evolves into robots fighting other robots in the skies could mean a lot fewer dead human beings. I’m just not very confident that powerful new military technologies will lead to improvements in human welfare. What I am confident about is my prediction that these technologies are possible.

Positive techno-optimism is basically the opposite of stagnationism. In recent years, a number of thinkers have started to say that humanity has basically picked the low-hanging fruit of what the Universe can do, and that future innovation will get a lot more expensive — or even slow to a trickle. Positive techno-optimism says that no, there’s still a lot of low-hanging fruit, and relative to our total GDP it’s still not that expensive to find it.

Normative techno-optimism is different. It says that more technology will make the world a better place for humanity. In fact, this kind of techno-optimism is surprisingly rare — even some of my favorite science fiction stories are implicitly or explicitly built around the idea that no matter how our capabilities improve, humans’ fundamental brutish nature will never change. As far as I can tell, this very prevalent attitude comes from the World Wars in the 20th century, in which the industrial technologies that were deployed to improve living standards before 1914 were turned to destructive purposes. There are a few people who argue that more technology makes us better people — Steve Pinker made this argument in The Better Angels of Our Nature, and I think it’s the implicit premise of Lois McMaster Bujold’s science fiction.

So that’s one distinction. For both of these, you can then define both active and passive forms of optimism. Passive optimism is basically just the belief that technology either will keep progressing at a rapid clip, and/or that it will improve human life, no matter what we do. Active optimism is the idea that technology will keep progressing and/or improve human life only if we humans take the appropriate actions to make sure this happens. Active optimism is what Gramsci and others have called “optimism of the will”. You could also call these “unconditional” and “conditional” optimism.

So you could define a little two-by-two chart:

Obviously these aren’t all mutually exclusive. For example, you could think that technological progress is inevitable, but that it’ll only benefit humanity if we choose the right policies. Or you could believe that technological progress will only continue if we choose the right policies, but that if it does continue, it’ll automatically be good for the world. Etc.

“Techno-optimism” is a sign of V.C. crisis

It’s not funny, V.C.s only do this when they’re in extreme distress.


OCT 20, 2023

I would say my main question about “The Techno-Optimist Manifesto,” an essay by the prominent venture-capitalist Marc Andreessen, is whether or not the goatse was intentional. The manifesto, published to the website of Andreessen’s fund a16z on Monday (it currently occupies the entire front page of the fund’s site), was originally accompanied by the following digital illustration, which many readers will recognize immediately:

Look familiar?

Does this image of a cyber-sphincter, pulled apart by two strong hands, revealing a pixellated heaven inside, look familiar? Did Marc Andreessen–a man who certainly would like to be known as a funny online guy, familiar with the folkways of the internet, able to cut it up with regular shitposters on the e-streets–mean for his grand statement of purpose to be illustrated by an abstract version of the internet’s most famous and beloved shock image?

Sadly, probably not. The image has since been changed, the hands removed, the resemblance gone. It’s too bad, because before it was swapped out the goatse was about the only surprising or intriguing component of Andreessen’s manifesto–the only aspect that suggested a sense of irony or self-awareness.

Read Max is a newsletter for techno-optimists, techno-pessimists, techno-agnostics, and techno-perverts. It is supported entirely by paying readers, in their capacity as agents of the techno-capital machine.


The rest of it is, well … you can read it yourself, though I can’t really recommend that. The manifesto is about 5,000 words; nearly every sentence is treated as a single paragraph, which makes it feel (at best) like Andreessen just dumping a notebook in which he’s jotted down some quotes he liked and lines that came to him in the shower, or (at worst, and more often) a LinkedIn post straining for profundity. He repeats himself and over-uses italics; he quotes Nietzsche at tedious length, and also Carrie Fisher. At the end he assembles a list of “Patron Saints of Techno-Optimism,” starting with three Twitter accounts.

The basic thrust of “The Techno-Optimist Manifesto” is that technology is the key to human thriving, and that certain malign elements in society–Andreessen names “experts,” “bureaucracy,” “sustainability” and “social responsibility” as “enemies”–have convinced us otherwise. These “enemies,” who “are suffering from ressentiment” must be escorted “out of their self-imposed labyrinth of pain,” for the good of humanity, and convinced of the error of their ways. Once their path is cleared, techno-optimists can make “everyone rich, everything cheap, and everything abundant.” Eventually, “our descendents will live in the stars.” He concludes: “We owe the past, and the future. It’s time to be a Techno-Optimist. It’s time to build.”1

If that sounds particularly familiar, it may be because “It’s Time to Build” is the title of a post Andreessen wrote in April 2020, in which he offers a broadly similar, if somewhat less lofty, argument: “The problem is inertia… The problem is regulatory capture… Every step of the way, to everyone around us, we should be asking the question, what are you building?”

Well, yes, indeed. What are you building? One reason that, three and a half years later, Andreessen is reiterating that “it’s time to build” instead of writing posts called “Here’s What I Built During the Building Time I Previously Announced Was Commencing” is that Marc Andreessen has not really built much of anything. In the years since he determined that it was time to build, his fund invested tens of millions of dollars in a video-game Ponzi scheme that immiserated its players and a company that sells blockchain transaction records said to reflect ownership of ape cartoons. That’s not just not building; it’s so not-building it’s not even the opposite of building, which would be “destroying,” and has the benefit of relating to the real world in some way. It’s ”venture investing in crypto companies,” which is its own little onanistic universe, conceptually and practically unrelated to “building” entirely.

What Andreessen has been doing in lieu of building, and besides signing off on bad cryptocurrency investments, is spending a lot of time on Twitter.

Slouching Towards Utopia

by Brad DeLong


OCT 22, 2023

I wrote a book on AI development, exploring how it works, its history, and its future. Order it here, ideally in triplicate!

That in the long 20th century, for the very first time in human history, the principal axis of history was economic—rather than cultural, ideological, religious, political, military-imperial, or what have you


Slouching towards utopia is a wonderful, magisterial book by Brad DeLong

that shows how humanity is moving towards utopia, even if under duress, bedraggled, one step at a time. It is a grand narrative and travels through the long 20th century, from 1870 to 2010. A quick summary of my impressions, though fair warning since the book defies easy categorisation it’s only a soupçon and not the whole meal.

The book puts forth an interesting but ultimately unconvincing grand narrative that globalization, industrial research, and modern corporations were the key drivers of economic progress in the “long 20th century.” However, these factors alone don’t fully explain the complex and multi-faceted forces behind innovation and growth, which are better seen as stacked S curves that evolve and compete within our demandscape, These are necessary lenses, but perhaps not sufficient.

While markets and capitalism played a crucial role, DeLong somewhat overstates their impact and underplays the role of technology itself as well as social and political factors. The interplay of ideas, knowledge, resources, and talent seems more important. It feels weird to call this utopian book one that doesn’t fully deal with “technology” but it does feel pertinent.

I feel there is a companion biography within the same time period which solely focuses on technology and explains the change in humanity as a way we deal with the powers it deals us. I hope someone’s written it. If not, take this as inspiration and I can be persuaded to write one.

And so, the core question of what enabled the rapid technological progress in the late 19th and 20th centuries remains open. The reviewer suggests factors like the connectivity between inventors, the size of the potential market/reward, and the societal structures shaping individual potential over time may offer more insight.

Overall, the book is intellectually stimulating and highly insightful, and the dichotomy between markets and redistribution is a convenient scaffolding to show DeLong’s incredible erudition. I wish I could read the other half of the book that unfortunately met the editor’s scissor!

This one’s a long essay, as befits a magisterial work. But still readable, as the book deserves. Still, here’s a quick guide. The first three sections are about the book itself and its themes. The next three sections are exploring how we should understand those themes. The final section is my conclusion of how to answer the challenges the book raises. Now, the book.

What does it take to become a unicorn founder? Insights from 200 startup journeys

The future of innovation is global. We discuss it here.

OCT 21, 2023

Is there a proven road to become a unicorn founder? What does it take to scale a startup, and is it different by market?

One of the most interesting pieces of research in the field recently came out by Endeavor, which charted the journeys of unicorn founders in the U.S., China and Emerging Markets, leveraging their rich global dataset.

What they found in some cases confirmed what I already knew or assumed and in other ways challenged me.

A few highlights that stood out for me:

Global citizens: the best, particularly emerging market founders leveraged multiple global experiences

Elite markers mattered less than is publicized: only 1/3 of founders graduated from an elite university and 20% at an elite employer

What did matter was startup experience at incredible fast growing startups. 50% had previous startup/scale-up experience. Remarkably, 50% were serial entrepreneurs. The average had 10 years experience

In my book I called them creators, but emerging market founders solve real mass market problems

Without further adieu, the below are excerpts from the Behind The Scenes and the Original Endeavor Report published here and here.

The Big Reset in Seed to Series A Graduation Rates is Real and Permanent

We are in the middle of a big reset where the percentage of Seed stage companies that graduate to Series A rounds will drop dramatically, with implications for founders and VCs.


I wrote this memo for the investing team at Precursor earlier this quarter, as I wanted to distill a bunch of my thoughts and share them with our team. I decided to publish it to a larger audience with a few minor edits – I hope people find it helpful.

A Fundamental Shift in the Series A Venture Landscape

For the past ten years, the near-term goal for most seed-stage companies has been to raise a Series A round of investment. Good venture capital firms saw 50-75% of their seed investments graduate to Series A rounds during this time. The ability to get successful seed-stage companies financed at the Series A stage was taken as a given by most seed-stage firms and their limited partners. In many ways, graduation rates became a key part of the marketing story around seed-stage firm performance.

Thanks for reading Venture Reflections! Subscribe for free to receive new posts and support my work.

For the past 18 months, the Series A market has been very quiet. Outside of AI-related investments, it feels like deal volume is off 75%. The Series A investors I know don’t feel any pressure to make investments and don’t really seem that excited or interested in much these days. Unfortunately for seed-stage companies, the Series A market can remain on strike longer than most seed-stage companies can remain solvent.

Part of the premise for the rise of seed VC funds in the late 2000s was that cloud computing, other technical innovations, and lean startup thinking meant that startups could get to validation points earlier. Many articles were written about how we would see a generation of capital-efficient startups that would do more with less as a result. This capital-efficient ethos never really took root as it happened to coincide with an unprecedented explosion in the venture ecosystem’s access to capital and historically low interest rates. There was no need to be capital-efficient in an era of capital abundance.

With the era of capital abundance coming to an end, the stage might finally be set for the era of the capital-efficient seed stage company. Seed-stage startups are likely to confront a world in which raising Series A rounds of investment remains difficult for quite some time. There will be a premium on execution and doing more with less capital. The necessary pieces might finally be in place to push companies back toward the hoped-for levels of capital efficiency.

There is a natural tendency to assume the current depressed level of Series A funding is just a blip, and things will return back to previous levels. While that is possible, we are far more likely in a protracted era where Series A financing will remain challenging. The level of activity that we saw in 2020 and 2021 was an anomaly, and we are unlikely to see a return to those levels anytime soon (with the exception of investing in AI). To the degree that the Series A market comes back, it will be far more muted than it was at the peak. If not for the bridges and extensions done in 2022, we would already have more data on the tightening Series A market as more companies would have likely tried and failed to raise.

There is value in thinking about what life looks like from the vantage point of a Series A investor in today’s market. From their vantage point, things look challenging:

Many Series A and multi-stage funds are managing the largest funds they’ve ever raised. Many firms are still finding a way to deploy the largest funds they’ve ever raised, and it feels like there is some level of indigestion in the ecosystem. There is a lot of pressure to find companies that will produce big enough outcomes to impact their fund returns. The chart below (courtesy of Pitchbook) highlights this; it shows the amount of money raised by all US VCs, including the growth in the number of funds that were raised:

Lower public market revenue multiples means companies have to do more to go public. The public markets care about financial performance, and becoming a public company requires at least $200 million in ARR to be taken seriously. The multiple contraction in public markets impacts the way VCs think about the value of their private holdings and what those companies need to achieve to be IPO-ready.

Exits need to be $5 billion or more to matter – Related to the fund size comment above, there is no way to make the math work for large funds without big outcomes. If you own 20% of a $1 billion company, it only returns $200 million. That just doesn’t matter for a $1 billion or larger venture fund. I suspect that much of the enthusiasm for AI investing is driven by a desire to chase what feels like a really big market opportunity in the making.

Series B through pre-IPO rounds are very hard to get done – The people who do the follow-on investments for Series A investors are also not doing a ton of deals. Like seed investors, Series A investors can only be the point person on so many investments; they also need people who can finance the next round and help them manage their portfolio support load.

Series A firms do their own seed investing and are not relying on seed and pre-seed companies to show them deals – Most multi-stage firms have created their own dedicated seed funds or seed programs to do early-stage deals; they are not wholly reliant on seed funds and others in the ecosystem to bring them deals. This in-house pipeline of deals will compete with opportunities that we and other seed firms bring them.

With that backdrop, here’s what we will see in the remainder of 2023 and through 2024 and possibly beyond:

State of Venture in Fall 2023


Earlier this week, I held the annual meeting for K9 Ventures. As I mentioned in the meeting, I am grateful to the K9 Founders for allowing me to be along for the ride in the companies, and I am also thankful to the K9 LPs for allowing me to do what I love to do: build companies.

I present my view of the startup ecosystem in each annual meeting to the LPs. Unlike other funds that may have teams to crunch numbers and data and look at micro and macro trends, my view is simply that: my view. I describe this to K9 LPs as this is my opinion and feelings based on what I see from my vantage point as a pre-seed/seed stage investor.

Here are the five slides I’ve excerpted from the 2023 K9 Ventures LP Zoom deck that cover this:

Early stage venture, particularly Pre-Seed and Seed stage venture, is a different game today than it used to be 10 years ago. LPs and GPs should be aware of this dynamic as they make investment decisions.

For founders, my advice is consistent: ‘The best time to start a company is when you have an idea that won’t leave you alone.‘ You have to be a little bit crazy and naive to start a company. It’s never going to be easy. Even though the “venture market” might be in a tough spot, there are still incredible opportunities to innovate and create value. That will never stop. And the company you build today will be healthier. Ad astra

Video of the Week

AI of the Week

Apple and AI

Monday, 23 October 2023


Speaking of Mark Gurman, he has an interesting bit about Apple’s AI efforts in his latest Power On column:

The company has some catching up to do. Apple largely sat on the sidelines when OpenAI’s ChatGPT took off like a rocket last year. It watched as Google and Microsoft Corp. rolled out generative AI versions of their search engines, which spit out convincingly human-like responses to users’ queries. Microsoft also updated its Windows apps with smarter assistants, and Inc. unveiled an AI-enhanced overhaul of Alexa.

All the while, the only noteworthy AI release from Apple was an improved auto-correct system in iOS 17.

I would argue that the improved autocorrect in iOS 17 is a major feature — in my use it’s clearly an improvement, and autocorrect is a feature used every day, in almost every app, by almost every iOS user. It’s one of the most used and most important features in the entire OS. I’d also argue that Apple has done some terrific work with AI features in Photos. The search feature in Photos works really well.

But I think Gurman’s summary does get at an essential truth. If I asked you “Which companies are at the forefront of AI-powered products?”, I doubt you’d put Apple on the list. And AI is proving so useful — and yet is a nascent field — that Apple needs to soon be on that list, lest their products begin to fall behind competitively. “Which companies are best at integrating AI into products?” is going to be like “Which companies are best at creating hardware at scale?” and “Which companies are best at human interface design?”

Gurman continues:

Now, Chief Executive Officer Tim Cook says that Apple has been working on generative AI technology for years. But I can tell you in no uncertain terms that Apple executives were caught off guard by the industry’s sudden AI fever and have been scrambling since late last year to make up for lost time.

“There’s a lot of anxiety about this and it’s considered a pretty big miss internally,” a person with knowledge of the matter told Power On.

As I first reported in July, the company built its own large language model called Ajax and rolled out an internal chatbot dubbed “Apple GPT” to test out the functionality. The critical next step is determining if the technology is up to snuff with the competition and how Apple will actually apply it to its products.

What I have heard from little birdies in Cupertino is not that there was a miss on this already. Apple is almost never at the forefront of stuff like this. They’re a deliberate company. Their goal, as with any new technology, is to integrate it into products in meaningful ways best, not first. That’s why there’s no internal anxiety that they’ve already missed anything related to AI.

The anxiety inside Apple is that many people inside do not believe Apple’s own AI/ML team can deliver, but that the company — if only for privacy reasons — will only use what comes from its own AI/ML team.

Amazon’s Andy Jassy Plans to Crash the AI Party

Two years after he replaced Jeff Bezos, Amazon’s CEO has been tightening everything from belts to culture. But he’s not letting a once-in-a-generation technology shift pass him by.

Amazon CEO Andy Jassy believes things are far from settled in the AI race: “We are like three steps into this marathon.” Art by Clark Miller; photo Getty.

By Jessica E. Lessin, Oct. 20, 2023, 7:00 AM PDT

I’m nine minutes into a two-and-a-half-hour conversation with Andy Jassy when it becomes clear why the CEO has invited me to Amazon’s Seattle headquarters. Like everyone else in business these days, he wants to talk about artificial intelligence.

Unless you’ve been under a rock this year, you know that the breakthroughs in large language models are upending the balance in tech. Microsoft is ascendant again after a strategic alliance with hotshot large language model developer OpenAI. Chipmaker Nvidia has skyrocketed to becoming one of the most valuable companies in the world with a lockdown on the chips that businesses everywhere are using to train their models.

But Jassy—who stepped into Amazon’s top job two years ago—thinks things are far from settled. “We are like three steps into this marathon,” he told me, while a Foo Fighters album from his vinyl collection played in the hallway for his staff.

As Jassy sees it, Amazon is focused on three layers of the proverbial AI stack. The first is the models themselves, which costs billions of dollars a year to build from scratch and are getting more expensive. The second is the applications built on top of the models—some of these, like ChatGPT, have gotten the lion’s share of the hype, but Jassy says many more—with different uses—will emerge.

And then there is the third layer in the stack where Amazon is making a big investment: that’s the chips that will power those ever-more-advanced applications. “We think a lot of the future training and inference is gonna be done on our chips,” he said.

His wonky description of the “three layers of the stack” isn’t exactly a talking point that screams tech visionary. And you could find a graveyard of companies that, when asked about some new disruptive technology, brushed off the competition by claiming it was “early days.”

But, like many things with Jassy, he presents his views with a logic—and supporting examples from 25 years at Amazon—that make them hard to ignore. And there is an intensity to his delivery that builds confidence—even if you aren’t sure he’s right. I left our lengthy conversations earlier this month more convinced than ever that the AI race has just begun—and that nobody should count Amazon out.

“Externally they haven’t done as much splash as some others,” Hans Vestberg, the CEO of Verizon, tells me of Amazon’s AI strategy. “But that’s typical of the organization. They think it through.”

Vestberg has known Jassy since before both got their CEO titles; his telecom giant most recently partnered with Amazon Web Services to provide accelerated 5G network speeds.

Like the dozens of business associates, board members and colleagues I spoke with about the Amazon chief, Vestberg attested to how little being CEO has changed Jassy. He and his wife still live in the same house they have for 14 years in Seattle’s Capitol Hill neighborhood, a far cry from the lakeside palaces of Bill Gates and former boss Jeff Bezos. He still prefers downtown Seattle restaurants to private executive dining rooms, and even knows local waiters’ names. He can still recite stats about the reading levels of the students at the local public charter school whose board of directors he chairs. He reads through a nightly packet of documents and data on Amazon’s businesses while exercising on an elliptical.

Thrive Capital to Lead Purchase of OpenAI Employee Shares at $80 Billion-Plus Valuation

By Kate Clark, Oct. 19, 2023 3:10 PM PDT

Investment firm Thrive Capital is leading a deal to buy OpenAI shares from employees through  a tender offer that would give the company a paper valuation of at least $80 billion, according to a person with direct knowledge. The deal would boost OpenAI’s valuation by at least three times from a similar transaction the startup made six months ago.

The deal would expand Thrive’s investment in the artificial intelligence developer, which sold employee shares to Thrive and other investors around April at a valuation of $27 billion. OpenAI’s new valuation would be more than 60 times its annualized revenue and make it one of the most highly valued firms backed by venture capital.

• Josh Kushner’s investment firm is investing in OpenAI for the second time
• The deal would value the startup at more than 60 times annualized revenue
• Investors expect OpenAI to raise more capital after the tender offer

While news of OpenAI’s planned employee tender has been previously reported by various outlets, Thrive Capital’s involvement hasn’t previously been reported.

Investors expect OpenAI to use the new valuation from the tender offer to set a floor price for an outside capital raising. OpenAI CEO Sam Altman has privately said he expected the company to raise $100 billion in its quest to develop artificial general intelligence. Microsoft already has committed more than $10 billion to OpenAI in exchange for a substantial cut of the startup’s future profits and access to its technology.

The startup’s business has experienced meteoric growth this year. After generating $28 million in revenue last year, Altman has told staff the company is generating $1.3 billion in annualized revenue—a measure of the prior month’s sales multiplied by 12. The business largely consists of paid subscriptions to its ChatGPT chatbot and selling access to GPT-4, a large language model that powers ChatGPT, along with compute capacity to power it. The company’s losses couldn’t be learned. Its staff as more than doubled since the end of last year to at least 700.

The steep valuation of OpenAI in the new tender offer has turned off some potential investors, according to people who have looked at the deal. Tiger Global Management, Bedrock Capital, Sequoia Capital and Andreessen Horowitz bought OpenAI shares in 2021 from other shareholders.

AI Reaccelerating Cloud Growth

Tomasz Tunguz, Venture Capitalist at Theory

Yesterday, both Google and Microsoft announced their earnings for their cloud businesses. I’ve been tracking the growth rates of these companies and product lines for the last 18 months to develop a broad gauge of enterprise buying patterns after the downturn.

In the chart, it’s immediately evident that Mongo has seen dramatic increase in growth rate last quarter. We await the earnings from everyone aside from Google and Microsoft in the next few weeks.

Notably, Microsoft has also re-accelerated – like Mongo, although to a lesser extent. This growth is fueled by AI workloads.

Google, on the other hand, declined after having plateaued last quarter.

Listening to the Microsoft earnings call, There are some interesting insights to be gleaned.

First, Microsoft is diversifying its algorithms portfolio from primarily OpenAI to open source models.

“Azure AI provides access to best-in-class frontier models from OpenAI and open-source models, including our own as well as from Meta and HuggingFace”

Second, the adoption of Azure OpenAI has grown from 11,000 last quarter to 18,000, a 67% growth, demonstrating the broad wave of adoption that has & likely will continue to Azure revenue growth.

“Because of our overall differentiation, more than 18,000 organizations now use Azure OpenAI service, including new-to-Azure customers.” “Higher-than-expected AI consumption contributed to revenue growth in Azure.”

Also, the growth rate should sustain next quarter, indicating some durability in the growth rate.

“In Azure, we expect revenue growth to be 26% to 27% in constant currency with an increasing contribution from AI. Growth continues to be driven by Azure consumption business, and we expect the trends from Q1 to continue into Q2.”

Third, the paid Copilot business – the assistant across Microsoft products is now roughly a $360m ARR business, assuming every customer pays list price. The average customer purchases 38 seats of Copilot (though this is very likely skewed by larger customers).

“We have over 1 million paid Copilot users and more than 37,000 organizations that subscribe to Copilot for business, up 40% quarter-over-quarter, with significant traction outside the United States.”

Microsoft’s presence in the developer ecosystem continues to astound. At the time of acquisition, Github had about 25m users & projected to reach 100m in 2025. They’ve crossed the mark 2 years early.

“All of the number of developers using GitHub has increased 4x since our acquisition 5 years ago.”

Both Google & Microsoft have announced Copilot for security products, features that other security vendors including Cisco are developing, uggesting this may be the next vertical to grow from AI after content generation & legal software.

Microsoft: “We see high demand for Security Copilot, the industry’s first and most advanced generative AI product, which is now seamlessly integrated with Microsoft 365 Defender. And our SIEM Microsoft Sentinel now has more than 25,000 customers and revenues past $1 billion annual run rate.”

Google: “We also integrated Duet AI across our cybersecurity portfolio to differentiate in the marketplace, providing generative AI-powered assistance in Mandiant Threat Intelligence, Chronicle Security Operations and Security Command Center. This reduces the time security teams spend writing, running and refining searches by 7x. “

Google has benefitted from some significant interest in Duet, their Copilot product which has driven customer expansion.

“In Workspace, thousands of companies and more than 1 million trusted testers have used Duet AI…Google Workspace also delivered strong revenue growth, primarily driven by increases in average revenue per seat. “

At hyperscale, the companies with the deepest AI exposure are enjoying faster growth rates as enterprise demand for these products accelerates.

Sam Bankman-Fried is going to talk himself right into jail

Somehow, the least suspicious parts of his defense are the 288 auto-deleting Signal conversations.

By Elizabeth Lopatto, a reporter who writes about tech, money, and human behavior. She joined The Verge in 2014 as science editor. Previously, she was a reporter at Bloomberg.

Sam Bankman-Fried is so fucked.

I have come to court every day since opening arguments thinking, Surely things cannot get worse for this man. Surely we have reached the bottom. Unfortunately, there is no bottom — in the prosecution’s telling, FTX and Alameda Research, his exchange and trading company, were matryoshka dolls of crime. Today, the defense started its case, which should theoretically present Bankman-Fried in a better light. But if what I saw of him on the stand is any indication, he may be more damning for himself than any of the prosecution’s witnesses.

Whatever Bankman-Fried can’t pin on Alameda CEO Caroline Ellison, he is essentially trying to off-load onto FTX lawyer Dan Friedberg. But blaming your lawyers for your decisions often implicates stuff — conversations, communications, documents — that are sometimes covered by attorney-client privilege. (If you blow up your own attorney-client privilege, it’s much worse for you than it is for anyone else, which is why most adults blame their own lawyers only under extreme circumstances.) The defense appears to be trying to thread the needle by saying that Bankman-Fried believed everything at FTX was fine because lawyers had been involved.

So today the jury got to go home early while the judge conducted an odd evidentiary hearing to figure out exactly what Bankman-Fried wants to tell the jury — and how much of it going to be admissible. 

Bankman-Fried took the stand as part of this hearing. This meant that prosecutor Danielle Sassoon got a crack at him, and boy howdy, she beat him like a piñata. 

We’ve heard a lot of testimony in this trial about disappearing messages on Signal, which the prosecution has strongly implied are evidence of wrongdoing. I don’t believe this to be true! Plenty of businesses destroy documents as a matter of course, for a wide variety of reasons, many of them harmless. 

The defense is seeking to testify that the disappearing text messages were part of a document retention policy which had been approved by FTX general counsel Friedberg. Though Bankman-Fried has testified that important business records were retained, the defense has been unable to produce the actual document retention policy, though they say it exists. 

News Of the Week

Threads Is Closing in on 100M Users Says Zuckerberg

Published Oct. 25, 2023

By Andrew Hutchinson Content and Social Media Manager

Well, this is something.

In today’s Q3 earnings call, Meta CEO Mark Zuckerberg shared a quick note about the growth of Threads thus far, and what he sees as the potential for the Twitter-like app.

Here’s Zuckerberg’s full summary:

“I want to give a quick update on Threads. We’re three months in now, and I’m very happy with the trajectory. There are just under 100 million monthly actives at this point, and we’re now getting to the point where we’re going to be focusing on growing the community further. From what we can tell, people love it so far. I’ve thought for a long time that there should be a billion-person public conversations app that’s a bit more positive, and I think that if we keep at this for a few more years then I think we have a good chance of achieving our vision there.”

A billion users. For a Twitter-like app.

For context, Twitter, now known as “X”, currently has around 253 million daily actives, and 550 million monthly active users, which is by far its highest ever usage rate.

Apple TV+ Show ‘The Problem With Jon Stewart’ Reportedly Canceled

Friday October 20, 2023 3:53 am PDT by Tim Hardwick

Apple TV+ show “The Problem With Jon Stewart” has been canceled because of “creative differences” between Stewart and Apple executives, reports The New York Times.

Members of the show’s staff were informed about its end on Thursday, according to several people with knowledge of the situation, just a couple of weeks before episodes for the third season were scheduled to begin.

The editorial disagreements reportedly revolved around some of the planned guests on “The Problem,” as well as potential show topics related to China and artificial intelligence that were causing concern among Apple executives. Apple did not respond to NYT‘s request for comment.

Apple in 2020 secured a multi-year deal with the writer, producer, and former host of “The Daily Show”, which saw Stewart return to television following his 2015 retirement.

The name of the show, “The Problem with Jon Stewart,” is a nod to the issues that Stewart explored on the current affairs new series, with topics in the 20 episodes ranging from what’s in the national conversation to issues that are part of Stewart’s advocacy work.

Note: Due to the political or social nature of the discussion regarding this topic, the discussion thread is located in our Political News forum. All forum members and site visitors are welcome to read and follow the thread, but posting is limited to forum members with at least 100 posts.

BRICS: Chinese Yuan Surpasses Euro, Becomes Second Main SWIFT Currency

Vinod Dsouza, October 19, 2023

BRICS member China is aggressively pushing its native currency, the Chinese Yuan, to compete against the US dollar, Euro, Pound, and other local currencies. In a first, the Chinese Yuan surpassed the Euro becoming the second main currency in SWIFT transactions. Euro’s use in international SWIFT payments is at its 3-year low while the Chinese Yuan is at its 5-month high.

Startup of the Week

Hands-On With the New Journal App in the iOS 17.2 Beta

Thursday October 26, 2023 1:45 pm PDT by Juli Clover

Apple today released the first iOS 17.2 beta, and it has the long-awaited Journal app that Apple initially announced as part of iOS 17. We’ve installed iOS 17.2 and thought we’d go hands-on with the Journal app to give MacRumors readers a first look at how it works.

Subscribe to the MacRumors YouTube channel for more videos.

The Journal app has a simple, straightforward UI. When you open it up, you’re presented with a “+” button, and tapping on it lets you start a new entry. When you first open the app, you can decide what information the Journal app can pull from your phone for writing prompts. Options include workout activity, media like podcasts and music, photos, and significant locations.

There is a setting to prefer suggestions with others when you’re around your contacts or large groups, and an option to be discoverable by others to help prioritize the suggestions of other people. Journal options can be toggled on and off in the Settings app under Privacy & Security > Journaling Suggestions.

You can opt to write whatever you like, or choose from one of Apple’s prompts that are based on your device usage. Prompts include invitations to reflect on your day, recall a new idea, think of what made you smile, think about what you could do to improve your life, and more.

Journal entries can be customized with images, location information, and voice recordings. Entries are organized by date, and can also be bookmarked so you can save your favorite entries. To keep your Journal secure, you can set it up so you need a Face ID scan to unlock it.

X of the Week