The Paragraph version of That Was The Week

Subscribe to That Was The Week
<100 subscribers
<100 subscribers
A reminder for new readers. That Was The Week collects the best writing on key issues in tech, startups, and venture capital. I select the articles because they are interesting. The selections often include things I disagree with. The articles are only snippets. Click on the headline to go to the full original. My editorial and the weekly video are where I express my point of view.

AGI will not happen in your lifetime. Or will it?
Google vs. ChatGPT
Unicorn Valuations Are On The Chopping Block
Thinning The Herd
Netflix’s New Chapter
Why Now is a Great Time to Raise Seed Funding. Even If It’s Awful for Series A-E Rounds.
Japan to remove limit on overseas investment by startup funds
Cowboy Ventures goes bigger with $260M across two new funds, including an opportunity fund
Lightspeed - Fintech Trends for 2023 and Beyond
What Microsoft gets from betting billions on the maker of ChatGPT
NEA Announces Two New Funds Totaling $6.2B
Substack
Dave Rubin on Twitter
There are three possible leads this week. ChatGPT’s ongoing success and Microsoft’s acknowledgment of a very large investment; The ongoing issue of social media, in particular, the return of “the former guy” to Facebook; or the continuing impact of the public market decline on private company valuations and on venture capital.
I vote for the latter. Social media is tiring and over-discussed. The former guy is increasingly irrelevant. So private company valuations and venture capital it is.
I make no excuses for stealing this week’s headline from an interpretation of articles written by Gené Tear at Crunchbase News, and Kyle Harrison from . Unicorn valuations on the Block and Thinning the Herd are two of my essays of the week.
First Kyle:
For many of these companies, the biggest reality that they will have to face is that for most of these companies, any valuation mark from 2020 - 2021 aren't just one of many data points. They're irrelevant data points.
In the words of Bill Gurley, "forget those prices happened."

And Gené:
Of the 1,433 companies on the board, more than 1,100 — or 78% — had their valuations set in the past two years. Those companies with new valuations represent almost three quarters of the entire board’s total value.
Keep in mind, our Unicorn Board reflects disclosed valuations tied to a priced funding event — not other valuations, such as those set in an internal event via a 409a valuation, or when investors revalue their portfolios via writedowns. (That’s why, for example, the board still lists Stripe’s valuation at $95 billion, although the payments company has reportedly reset its internal valuation to $63 billion — the third time in fewer than 12 months it has trimmed its valuation.)
The implication is clear. Many of the companies described as unicorns are not. And many venture funds that invested in them will mark down their investments if they haven’t already.
It also seems clear that new unicorn production will slow significantly. That will take two forms. First, it will take longer for companies to travel from early-stage rounds to later stages. Second, as they move from Series A to Series B, C, and D, the valuations will reflect the new reality where a 10x valuation measured from annual revenue will be rare.
Most likely, as the IPO market freezes over, later-stage investors will not show up as often or with as much capital. Exits will be rare.
In this market environment, a premium will be placed on a venture fund being able to find the best growth companies early. Access to the companies will become super competitive. Returns will be highly sought after. And liquidity will be valued more highly than paper multiples.
I met an LP this week who had decided not to re-up in many very prized venture funds. The complaint was that he can’t bank MOIC or TVPI. He only values DPI. In other words cash, liquidity, not paper gains.
The well-documented “dry powder” in the venture ecosystem may be a short-lived reality.
of course, where there is pain, there also is gain. Last week we covered Blackstone raising a $25bn secondary fund focused on buying distressed assets with good prospects. One person’s distress is another person’s opportunity. And this week we cover SaaStr’s article stating that seed investing will flourish, even though A-E rounds will struggle.
One thing is for sure - nothing remains the same.
The Video and Podcast with myself and Keen On publisher Andrew Keen that accompanies That Was The Week is recorded separately and delivered to paying subscribers via email on Friday or Saturday each week.

Grady Booch, a Fellow at IBM, and one of the deepest minds on software development in the history of the field, is one of my favorite people that I have met through Twitter.
We rarely disagree about anything. On Twitter we sometimes tag team Yann LeCun, and on more than a few occasions we have retweeted each other’s cynical tweets. Neither one of us believes for a minute that large language models are remotely close to artificial general intelligence, and what’s more, we mostly agree about what’s missing: semantics, reasoning, common sense, theory of mind and so on.
But I found myself disagreeing with the Tweet above. My own skepticism has limits. I do think we will eventually reach AGI (artificial general intelligence), and quite possibly before the end of this century.
On the theory that our differences might represent some kind of teachable moment, I invited him to talk about it on the record; he immediately agreed. Here’s a very lightly edited transcript of our conversation (approved by both of us).
Gary Marcus: Grady, I get that AGI is hard, and that we aren’t there yet. I think we are wasting funding and bright young minds on an approach that probably isn’t on the right path. I don’t think we have great answers yet for a bunch of core issues, like how to get machines to represent abstract knowledge, and how to get them to induce complex models from the events that they see or read about it. When I think about how little deep progress has been made in the last 75 years, I get depressed.
But I am cautiously optimistic that we’ll do better in the next 75, that once the hype cools off, people will finally dive deeper into neurosymbolic AI, and start to take some important steps. Our data problems are solved, our compute problems are mostly solved; it’s now mostly a matter of software, and of rethinking how we build AI. Why be so sure we can’t do that in the next 75 years?
Hi! After a long hiatus, Can here again. Sorry, been a while!
I don’t really cook often, but I do love making pasta aglio e olio. It’s a simple recipe that I discovered many years ago on YouTube. Yet, being the extremely irregular cook I am, I have to look up that simple recipe every once in a while. And every time I have to search for it, which is every six months or go, I feel like I have a worse experience on the internet than the last time. And by internet, I mean the Google monopoly.
I'm old enough to remember when Google came out, which makes me old enough to remember at least 20 different companies that touted as called Google-killers. I applaud every single one of them! A single American company operating as a bottleneck behind the world's information is a dangerous, and inefficient proposition and a big theme of the Margins is that monopolies are bad so it’s also on brand.
But for one reason and another, none of the Google competitors have seemed to capture the world's imagination.
Until now, that is. Yes, sorry, I’m talking about that AI bot.

Google is done. Compare the quality of these responses (ChatGPT)


So, is ChatGPT a real threat to Google search business?
Yes, but not for the reasons you might think……

January 20, 2023
The ranks of The Crunchbase Unicorn Board have swelled in the past two years, adding almost 1,000 new companies and trillions of dollars in reported value. But now, as private company valuations go through a hard reset in a tough market, the board faces a reckoning.
Since the beginning of 2021, 925 companies have been added to the Unicorn Board and more than $400 billion was invested in those unicorns. But many of those same companies that raised in 2021 and the first half of 2022 will find their valuations — even just 18 months after their most recent funding — are too high in the current economy.
Of the 1,433 companies on the board, more than 1,100 — or 78% — had their valuations set in the past two years. Those companies with new valuations represent almost three quarters of the entire board’s total value.
Keep in mind, our Unicorn Board reflects disclosed valuations tied to a priced funding event — not other valuations, such as those set in an internal event via a 409a valuation, or when investors revalue their portfolios via writedowns. (That’s why, for example, the board still lists Stripe’s valuation at $95 billion, although the payments company has reportedly reset its internal valuation to $63 billion — the third time in fewer than 12 months it has trimmed its valuation.)
According to Forge, a private securities marketplace, investors are purchasing private company stock on the secondary market at a median discount of 50.5% compared to their most recent funding. If that discount were applied across the Unicorn Board, 821 companies — or 57% of the companies on the board — would drop off, and the value of the board would decline to $2.4 trillion from its current total valuation of $4.9 trillion.
The funding environment has changed swiftly after a heady couple of years.
Valuations at 10x forward ARR in 2021 became “100x, 200x, and 400x multiples,” Tomasz Tunguz, previously a partner at Redpoint and now raising his own fund, wrote on his blog in April 2022.
Now, we’re seeing internal valuations revised downward — often drastically — for highly valued unicorns. Many companies are quietly raising extension rounds from existing investors, while those unable to do so sometimes crash and shut down. (That was the case with self-driving car technology company Argo AI, which shut down in the fourth quarter of 2022 after being unable to raise fresh capital).
In July 2022, Klarna lost $39 billion in value, plummeting from a $45 billion valuation to $6.7 billion in a new funding round.
Boston-based OutSystems, a low-code engineering platform, was valued at $9.5 billion in February 2021. It quietly raised funding eight months later, led by existing investor KKR — more than halving its valuation to $4.3 billion, according to Forbes.
And Oda, a door-to-door grocery delivery service based in Norway, was valued below a billion dollars in a December fundraising, dropping from the unicorn ranks entirely.
Crypto companies are also facing a harsh new reality. Crypto exchange FTX, valued at $32 billion in a January 2022 funding, is valued at nothing. Other crypto unicorns in bankruptcy proceedings include crypto wealth management platform BlockFi and blockchain lending platform Celsius Network……

In the last ~3 years, we've seen hundreds of companies reach unicorn status. Many of them are now armed with their $1B+ valuations, years of runway (maybe), and in almost every case, less than $100M in revenue. I can't tell you the number of companies I saw raise at $1B+ valuations on <$5M of revenue.
A lot of VCs are predicting that many of them will fail, in some instances as high as 50%+. I also think that a lot of companies who are overvalued are going to face a significant punch to the face. Many of them are reacting dramatically to the changing market. But I've been surprised how many are sort of going along business as usual. So I started to put down thoughts on how that might play out, for better or worse (mostly worse).
For many of these companies, the biggest reality that they will have to face is that for most of these companies, any valuation mark from 2020 - 2021 aren't just one of many data points. They're irrelevant data points.
Posted on Monday, January 23, 2023
Netflix’s moment of greatest peril is, in retrospect, barely visible in the company’s stock chart:

I’m referring to 2004-2007 and the company’s battle with Blockbuster:

The simplified story of Netflix’s founding starts with Reed Hastings grumbling over a $40 late charge from Blockbuster, and ends with the brick-and-mortar giant going bankrupt as customers came to prefer online rentals from Netflix, with streaming providing the final coup de grâce.
Neither are quite right……
by Jason Lemkin | Blog Posts, Early, Fundraising

So now is simply a terrible time to be raising growth stage venture capital. With the BVP Cloud Index down 40%+ from a year ago, and many of the top Cloud leaders trading for as low as 5x ARR, it’s just hard to justify a later stage $300m-$400m valuation round these days, let alone a true unicorn round.
We just did our latest Workshop Wednesday with Omers Growth, and they only did one growth investment last year. One. A great deep dive on later-stage SaaS investing here:
And look venture is a mess today. Crypto investments are going bankrupt. Unicorns that looked like sure things a year ago are now running out of money. Board meetings are endless drama sessions. And there are too many founders looking for bridge financing, and not enough bridge financing to go around,.
And yet, these are still very good times to raise a seed round. If … you are capital efficient.
Why?
First, there are tons of new and newer seed managers that are ready to go. They don’t have 20 broken unicorns to deal with, nor lots of mark down to manage.
Second, top funds (including Sequoia this week) have brand-new Seed funds. They are ready to go, too. Again, without the direct baggage of prior funds.
Third, VCs believe they can still make money in seed. Yes, if you invested in Figma at a $10m pre and it sold for $2B instead of $20B to Adobe, you would have made a lot less money. But still plenty of money.
Fourth, many growth and mid-stage VCs are fleeing to seed. This might not be as smart as it looks, as risk of scaling goes up. But many later stage VCs are doing more seed because growth is mostly frozen.

This isn’t to say it’s quite as good of times for Seed as it was during the crazy times of late 2020 to very early 2022. Not even close. But — it’s still pretty darn good because there are so many more players that want to invest.
It’s just, there’s one huge change. You have to be capital efficient. The odds of another check coming anytime soon have plummeted. Series A rounds have become very tough, and Series B-E rounds have become downright rare.
So no one can stomach a high burn rate anymore.
But if you have a smidge of traction, early strong growth, and are capital efficient — there’s far more capital out there than founders to invest in. In fact, every great Seed VC fund I know of has more capital to deploy right now than deals. They’re ready to go.
Even at SaaStr Fund, I’m looking to do 2-4 deals this year from $750k-$4m. And I’m already behind.
A deep dive on this and much more in our latest conversation with Harry Stebbings on 20VC here:
Portfolio cap seen hindering ability to attract foreign investors

Japan aims to increase annual investment in startups by more than tenfold by fiscal 2027. (Photo by Hirofumi Yamamoto)KOSUKE TAKEUCHI, Nikkei staff writerJanuary 23, 2023 07:05 JST
TOKYO -- Japan intends to scrap a key restriction on overseas investments by domestic venture funds dedicated to backing startups, a move intended to attract more foreign investors into the sector.
Domestic startup funds currently must limit investment in foreign companies to less than 50% of their portfolio holdings. Ending this restriction would allow more flexibility in choosing investment targets, encouraging overseas money to flow into such Japanese funds.
The planned deregulation applies to startup funds that take the form of limited partnerships, a common method tapped by venture capital to curb risk exposure. Investors participating as limited partners are not required to take on risk beyond their investment. The arrangement not only provides an easy way for investors to supply money, but also is suited for collecting small amounts from a wide range of investors.
The requirement for limited partnerships in Japan to invest at least 50% domestically hampers foreign investors that see Asia as a whole as a target. Removing the restriction would help Japanese venture capital funds expand their investment targets overseas, and an increase in money flowing in the domestic venture capital market ultimately will benefit Japanese startups.
Japan's government will submit to parliament a draft amendment to the law governing limited partnerships in 2024. The cap on investing in foreign companies was introduced as part of the law on limited partnerships that took effect in 1998.
Japan has set a goal of increasing annual investment in startups to the scale of 10 trillion yen ($77.2 billion) by fiscal year 2027, more than a tenfold increase from current levels.

Cowboy Ventures, the now-10-year-old, Bay Area-based seed-stage focused fund founded by renowned investor Aileen Lee, has closed on two new funds totaling $260 million in capital commitments. The outfit garnered $140 million in commitments for its fourth flagship fund and another $120 million for its first opportunity-type fund (its “Mustang Fund”).
The amount is more than all the capital that the outfit has raised across its previous funds, which were sized at $40 million, $60 million and $95 million, respectively. Then again, the team has grown over the years from being a one-person firm to an outfit with an investor team, including fintech specialist Jill Williams, who Lee recruited from Anthemis, and Amanda Robson, who was pulled out of Norwest Venture Partners, where she worked with numerous enterprise software companies, including some focused on AI and robotics. (Longtime Silicon Valley attorney Ted Wang is also closely associated with the fund as a “board partner” and advises more than a dozen of its portfolio companies.)
It’s easy to appreciate why LPs committed more capital to Cowboy, even in a market that seems to be actively shrinking given broader market turmoil.
First and foremost are its numbers, which look good, particular given the size of its earlier funds. Cowboy was among the first investors in Guild Education, for example, an online education company that’s focused on upskilling frontline employees, and was valued at $4.4 billion when it closed its most recent round of funding in June of last year. Cowboy is also a seed investor in the security and compliance automation platform Drata, assigned a $2 billion valuation in December when it raised $200 million in Series C funding.

Image Credits: Cowboy Ventures
In conversation with Lee, Williams and Robson late last week, Lee noted that Cowboy thinks of itself as a generalist firm, but that 70% of its most recent fund was funneled into enterprise startups and 30% into consumer startups, given Cowboy has also enjoyed success with the latter. (Most notably, one of its first checks went to Dollar Shave Club, the men’s grooming company acquired by Unilever in 2016 for a reported $1 billion.)
Others of the firm’s bets include Vic.ai, a startup that’s automating accounting processes and just closed a $52 million Series C round in December; Homebase, a platform for small to mid-size businesses that helps with scheduling, payroll, cash advances and HR stuff and has raised roughly $100 million from investors to date; and SVT Robotics, whose software organizes robots in warehouses and factories (it closed $25 million in Series A funding in late 2021).
Lightspeed’s investment team shares what’s fascinating them and what to watch for in the year ahead

Our global fintech team covers all sectors — from infrastructure, to enterprise application layer, to consumer across India, China, South East Asia, Europe, Israel, and the Americas.
As we collected our thoughts on what’s ahead, the world is experiencing rising interest rates, falling equities, and rampant inflation. It’s clear the global financial system is more interconnected now than ever. Our team’s belief in a world where finances are a seamless part of everyone’s lives, not a frustration, is stronger than ever.
We see fintech leading the way toward a financial ecosystem that is accessible, affordable, transparent, data driven, frictionless, cross-border, and convenient. Below are key areas we’re watching in 2023 that advance this vision:
……
The reported $10 billion investment in OpenAI will keep the hottest AI company on Microsoft’s Azure cloud platform.
By Sara Morrisonsara.morrison@recode.net Jan 23, 2023, 3:10pm EST

John Smith/VIEWpress/Corbis via Getty Images
Sara Morrison is a senior Vox reporter who covers data privacy, antitrust, and Big Tech’s power over us all.
Microsoft revealed last week that it will lay off 10,000 people throughout 2023. But don’t think that means the company is having money problems. On Monday, the company announced that it’s investing billions of dollars into the hot artificial intelligence platform OpenAI.
This is Microsoft’s third investment in the company, and cements Microsoft’s partnership with one of the most exciting companies making one the most exciting technologies today: generative AI. It also shows that Microsoft is committed to making the initiative a key part of its business, as it looks to the future of technology and its place in it. And you can likely expect to see OpenAI’s services in your everyday life as companies you use integrate it into their own offerings.
Microsoft told Recode it was not disclosing the deal’s specifics, but Semafor reported two weeks ago that the two companies were talking about $10 billion, with Microsoft getting 75 percent of OpenAI’s profits until it recoups its investment, after which it would have a 49 percent stake in the company. The New York Times has since confirmed the $10 billion amount.
With the arrangement, OpenAI runs and powers its technology through Microsoft’s Azure cloud computing platform, which allows it to scale and make it available to developers and companies looking to use AI in their own services (rather than have to build their own). Think of it as AIaaS — AI as a service. Microsoft recently made its OpenAI services widely available, allowing more businesses to integrate some of the hottest AI technologies, including word generator ChatGPT and image generator DALL-E 2, into their own companies’ offerings.
Meanwhile, OpenAI also gets a needed cash infusion — key for a company with a lot of potential but not much to show in terms of monetization. And Microsoft can offer something to its cloud customers that rivals Google and Amazon can’t yet: one of the most advanced AI technologies out there, as well as one of the buzziest. They do have their own AI initiatives, like Google’s DeepMind, which is reportedly rolling out a ChatGPT rival at some point. But it’s not here yet. ChatGPT is, and it’s gone mainstream.
January 26, 2023
Venture giant New Enterprise Associates announced Thursday it has closed on approximately $6.2 billion across two funds.
The firm, known for investments in transformational startups like Salesforce
, Workday and Robinhood, said one fund will be dedicated to early-stage investing and the other to venture growth-stage opportunities.
NEA will invest in a range of different technology and health care sectors — including both enterprise and consumer technology, digital health, and life sciences.
The firm had more than $25 billion of assets under management as of the end of 2022.
“We are deeply grateful to our limited partners for the trust they have placed in our team, and excited to have raised the largest pool of capital in NEA’s history at a time of great uncertainty, but also tremendous opportunity,” said Scott Sandell, managing general partner at NEA.
Founded in 1977, NEA is one of the oldest VC firms in the country. The firm has had more than 270 portfolio company IPOs and over 450 mergers and acquisition events.
Just in the last several months, NEA has participated in rounds worth $100 million or more for companies such as Escient Pharmaceuticals, MBX Biosciences and Swift Navigation.
The announcement comes at a time when many in the industry believe fund managers may have a hard time raising new funds as limited partners and institutions try to avoid risk and rebalance their portfolios after the beating many public tech stocks have taken.
However, such an environment is likely to favor more established firms such as NEA and those who have a proven track record.
Just last week, another seasoned firm, Sequoia, announced a new $195 million fund for seed-stage startups, according to Forbes.

Image Credits: Substack
Newsletter platform Substack announced today that it’s introducing several new features, including private Substacks. A private Substack is a publication that you can host alone or readers can request to subscribe to read your posts. Writers can choose to approve or decline each subscription request.
In a blog post, Substack explained that private Substack accounts work similarly to private Instagram profiles. A private Substack can be used to keep in touch with friends, build communities of interest and test the waters for a new publication, the company says.
Users can change their Substack from public to private at any time by navigating to their settings and selecting “Private” in the “Import” section. Once a user makes their publication private, readers won’t be able to see any posts. When a reader requests to subscribe to a Substack, the writer will receive an email notification with their details. You can see your requests on your Subscribers page. If you approve a request, the reader will be automatically subscribed and sent a Welcome email.
The launch of the new feature comes as Substack has been hoping to capitalize on Twitter’s upheaval following Elon Musk’s takeover. The company openly targeted Twitter’s user base in the past few months and recently threw its hat into the ring as a more direct competitor with the launch of Substack Chat, which allows writers to communicate directly with their loyal readers right in the Substack mobile app.

A reminder for new readers. That Was The Week collects the best writing on key issues in tech, startups, and venture capital. I select the articles because they are interesting. The selections often include things I disagree with. The articles are only snippets. Click on the headline to go to the full original. My editorial and the weekly video are where I express my point of view.

AGI will not happen in your lifetime. Or will it?
Google vs. ChatGPT
Unicorn Valuations Are On The Chopping Block
Thinning The Herd
Netflix’s New Chapter
Why Now is a Great Time to Raise Seed Funding. Even If It’s Awful for Series A-E Rounds.
Japan to remove limit on overseas investment by startup funds
Cowboy Ventures goes bigger with $260M across two new funds, including an opportunity fund
Lightspeed - Fintech Trends for 2023 and Beyond
What Microsoft gets from betting billions on the maker of ChatGPT
NEA Announces Two New Funds Totaling $6.2B
Substack
Dave Rubin on Twitter
There are three possible leads this week. ChatGPT’s ongoing success and Microsoft’s acknowledgment of a very large investment; The ongoing issue of social media, in particular, the return of “the former guy” to Facebook; or the continuing impact of the public market decline on private company valuations and on venture capital.
I vote for the latter. Social media is tiring and over-discussed. The former guy is increasingly irrelevant. So private company valuations and venture capital it is.
I make no excuses for stealing this week’s headline from an interpretation of articles written by Gené Tear at Crunchbase News, and Kyle Harrison from . Unicorn valuations on the Block and Thinning the Herd are two of my essays of the week.
First Kyle:
For many of these companies, the biggest reality that they will have to face is that for most of these companies, any valuation mark from 2020 - 2021 aren't just one of many data points. They're irrelevant data points.
In the words of Bill Gurley, "forget those prices happened."

And Gené:
Of the 1,433 companies on the board, more than 1,100 — or 78% — had their valuations set in the past two years. Those companies with new valuations represent almost three quarters of the entire board’s total value.
Keep in mind, our Unicorn Board reflects disclosed valuations tied to a priced funding event — not other valuations, such as those set in an internal event via a 409a valuation, or when investors revalue their portfolios via writedowns. (That’s why, for example, the board still lists Stripe’s valuation at $95 billion, although the payments company has reportedly reset its internal valuation to $63 billion — the third time in fewer than 12 months it has trimmed its valuation.)
The implication is clear. Many of the companies described as unicorns are not. And many venture funds that invested in them will mark down their investments if they haven’t already.
It also seems clear that new unicorn production will slow significantly. That will take two forms. First, it will take longer for companies to travel from early-stage rounds to later stages. Second, as they move from Series A to Series B, C, and D, the valuations will reflect the new reality where a 10x valuation measured from annual revenue will be rare.
Most likely, as the IPO market freezes over, later-stage investors will not show up as often or with as much capital. Exits will be rare.
In this market environment, a premium will be placed on a venture fund being able to find the best growth companies early. Access to the companies will become super competitive. Returns will be highly sought after. And liquidity will be valued more highly than paper multiples.
I met an LP this week who had decided not to re-up in many very prized venture funds. The complaint was that he can’t bank MOIC or TVPI. He only values DPI. In other words cash, liquidity, not paper gains.
The well-documented “dry powder” in the venture ecosystem may be a short-lived reality.
of course, where there is pain, there also is gain. Last week we covered Blackstone raising a $25bn secondary fund focused on buying distressed assets with good prospects. One person’s distress is another person’s opportunity. And this week we cover SaaStr’s article stating that seed investing will flourish, even though A-E rounds will struggle.
One thing is for sure - nothing remains the same.
The Video and Podcast with myself and Keen On publisher Andrew Keen that accompanies That Was The Week is recorded separately and delivered to paying subscribers via email on Friday or Saturday each week.

Grady Booch, a Fellow at IBM, and one of the deepest minds on software development in the history of the field, is one of my favorite people that I have met through Twitter.
We rarely disagree about anything. On Twitter we sometimes tag team Yann LeCun, and on more than a few occasions we have retweeted each other’s cynical tweets. Neither one of us believes for a minute that large language models are remotely close to artificial general intelligence, and what’s more, we mostly agree about what’s missing: semantics, reasoning, common sense, theory of mind and so on.
But I found myself disagreeing with the Tweet above. My own skepticism has limits. I do think we will eventually reach AGI (artificial general intelligence), and quite possibly before the end of this century.
On the theory that our differences might represent some kind of teachable moment, I invited him to talk about it on the record; he immediately agreed. Here’s a very lightly edited transcript of our conversation (approved by both of us).
Gary Marcus: Grady, I get that AGI is hard, and that we aren’t there yet. I think we are wasting funding and bright young minds on an approach that probably isn’t on the right path. I don’t think we have great answers yet for a bunch of core issues, like how to get machines to represent abstract knowledge, and how to get them to induce complex models from the events that they see or read about it. When I think about how little deep progress has been made in the last 75 years, I get depressed.
But I am cautiously optimistic that we’ll do better in the next 75, that once the hype cools off, people will finally dive deeper into neurosymbolic AI, and start to take some important steps. Our data problems are solved, our compute problems are mostly solved; it’s now mostly a matter of software, and of rethinking how we build AI. Why be so sure we can’t do that in the next 75 years?
Hi! After a long hiatus, Can here again. Sorry, been a while!
I don’t really cook often, but I do love making pasta aglio e olio. It’s a simple recipe that I discovered many years ago on YouTube. Yet, being the extremely irregular cook I am, I have to look up that simple recipe every once in a while. And every time I have to search for it, which is every six months or go, I feel like I have a worse experience on the internet than the last time. And by internet, I mean the Google monopoly.
I'm old enough to remember when Google came out, which makes me old enough to remember at least 20 different companies that touted as called Google-killers. I applaud every single one of them! A single American company operating as a bottleneck behind the world's information is a dangerous, and inefficient proposition and a big theme of the Margins is that monopolies are bad so it’s also on brand.
But for one reason and another, none of the Google competitors have seemed to capture the world's imagination.
Until now, that is. Yes, sorry, I’m talking about that AI bot.

Google is done. Compare the quality of these responses (ChatGPT)


So, is ChatGPT a real threat to Google search business?
Yes, but not for the reasons you might think……

January 20, 2023
The ranks of The Crunchbase Unicorn Board have swelled in the past two years, adding almost 1,000 new companies and trillions of dollars in reported value. But now, as private company valuations go through a hard reset in a tough market, the board faces a reckoning.
Since the beginning of 2021, 925 companies have been added to the Unicorn Board and more than $400 billion was invested in those unicorns. But many of those same companies that raised in 2021 and the first half of 2022 will find their valuations — even just 18 months after their most recent funding — are too high in the current economy.
Of the 1,433 companies on the board, more than 1,100 — or 78% — had their valuations set in the past two years. Those companies with new valuations represent almost three quarters of the entire board’s total value.
Keep in mind, our Unicorn Board reflects disclosed valuations tied to a priced funding event — not other valuations, such as those set in an internal event via a 409a valuation, or when investors revalue their portfolios via writedowns. (That’s why, for example, the board still lists Stripe’s valuation at $95 billion, although the payments company has reportedly reset its internal valuation to $63 billion — the third time in fewer than 12 months it has trimmed its valuation.)
According to Forge, a private securities marketplace, investors are purchasing private company stock on the secondary market at a median discount of 50.5% compared to their most recent funding. If that discount were applied across the Unicorn Board, 821 companies — or 57% of the companies on the board — would drop off, and the value of the board would decline to $2.4 trillion from its current total valuation of $4.9 trillion.
The funding environment has changed swiftly after a heady couple of years.
Valuations at 10x forward ARR in 2021 became “100x, 200x, and 400x multiples,” Tomasz Tunguz, previously a partner at Redpoint and now raising his own fund, wrote on his blog in April 2022.
Now, we’re seeing internal valuations revised downward — often drastically — for highly valued unicorns. Many companies are quietly raising extension rounds from existing investors, while those unable to do so sometimes crash and shut down. (That was the case with self-driving car technology company Argo AI, which shut down in the fourth quarter of 2022 after being unable to raise fresh capital).
In July 2022, Klarna lost $39 billion in value, plummeting from a $45 billion valuation to $6.7 billion in a new funding round.
Boston-based OutSystems, a low-code engineering platform, was valued at $9.5 billion in February 2021. It quietly raised funding eight months later, led by existing investor KKR — more than halving its valuation to $4.3 billion, according to Forbes.
And Oda, a door-to-door grocery delivery service based in Norway, was valued below a billion dollars in a December fundraising, dropping from the unicorn ranks entirely.
Crypto companies are also facing a harsh new reality. Crypto exchange FTX, valued at $32 billion in a January 2022 funding, is valued at nothing. Other crypto unicorns in bankruptcy proceedings include crypto wealth management platform BlockFi and blockchain lending platform Celsius Network……

In the last ~3 years, we've seen hundreds of companies reach unicorn status. Many of them are now armed with their $1B+ valuations, years of runway (maybe), and in almost every case, less than $100M in revenue. I can't tell you the number of companies I saw raise at $1B+ valuations on <$5M of revenue.
A lot of VCs are predicting that many of them will fail, in some instances as high as 50%+. I also think that a lot of companies who are overvalued are going to face a significant punch to the face. Many of them are reacting dramatically to the changing market. But I've been surprised how many are sort of going along business as usual. So I started to put down thoughts on how that might play out, for better or worse (mostly worse).
For many of these companies, the biggest reality that they will have to face is that for most of these companies, any valuation mark from 2020 - 2021 aren't just one of many data points. They're irrelevant data points.
Posted on Monday, January 23, 2023
Netflix’s moment of greatest peril is, in retrospect, barely visible in the company’s stock chart:

I’m referring to 2004-2007 and the company’s battle with Blockbuster:

The simplified story of Netflix’s founding starts with Reed Hastings grumbling over a $40 late charge from Blockbuster, and ends with the brick-and-mortar giant going bankrupt as customers came to prefer online rentals from Netflix, with streaming providing the final coup de grâce.
Neither are quite right……
by Jason Lemkin | Blog Posts, Early, Fundraising

So now is simply a terrible time to be raising growth stage venture capital. With the BVP Cloud Index down 40%+ from a year ago, and many of the top Cloud leaders trading for as low as 5x ARR, it’s just hard to justify a later stage $300m-$400m valuation round these days, let alone a true unicorn round.
We just did our latest Workshop Wednesday with Omers Growth, and they only did one growth investment last year. One. A great deep dive on later-stage SaaS investing here:
And look venture is a mess today. Crypto investments are going bankrupt. Unicorns that looked like sure things a year ago are now running out of money. Board meetings are endless drama sessions. And there are too many founders looking for bridge financing, and not enough bridge financing to go around,.
And yet, these are still very good times to raise a seed round. If … you are capital efficient.
Why?
First, there are tons of new and newer seed managers that are ready to go. They don’t have 20 broken unicorns to deal with, nor lots of mark down to manage.
Second, top funds (including Sequoia this week) have brand-new Seed funds. They are ready to go, too. Again, without the direct baggage of prior funds.
Third, VCs believe they can still make money in seed. Yes, if you invested in Figma at a $10m pre and it sold for $2B instead of $20B to Adobe, you would have made a lot less money. But still plenty of money.
Fourth, many growth and mid-stage VCs are fleeing to seed. This might not be as smart as it looks, as risk of scaling goes up. But many later stage VCs are doing more seed because growth is mostly frozen.

This isn’t to say it’s quite as good of times for Seed as it was during the crazy times of late 2020 to very early 2022. Not even close. But — it’s still pretty darn good because there are so many more players that want to invest.
It’s just, there’s one huge change. You have to be capital efficient. The odds of another check coming anytime soon have plummeted. Series A rounds have become very tough, and Series B-E rounds have become downright rare.
So no one can stomach a high burn rate anymore.
But if you have a smidge of traction, early strong growth, and are capital efficient — there’s far more capital out there than founders to invest in. In fact, every great Seed VC fund I know of has more capital to deploy right now than deals. They’re ready to go.
Even at SaaStr Fund, I’m looking to do 2-4 deals this year from $750k-$4m. And I’m already behind.
A deep dive on this and much more in our latest conversation with Harry Stebbings on 20VC here:
Portfolio cap seen hindering ability to attract foreign investors

Japan aims to increase annual investment in startups by more than tenfold by fiscal 2027. (Photo by Hirofumi Yamamoto)KOSUKE TAKEUCHI, Nikkei staff writerJanuary 23, 2023 07:05 JST
TOKYO -- Japan intends to scrap a key restriction on overseas investments by domestic venture funds dedicated to backing startups, a move intended to attract more foreign investors into the sector.
Domestic startup funds currently must limit investment in foreign companies to less than 50% of their portfolio holdings. Ending this restriction would allow more flexibility in choosing investment targets, encouraging overseas money to flow into such Japanese funds.
The planned deregulation applies to startup funds that take the form of limited partnerships, a common method tapped by venture capital to curb risk exposure. Investors participating as limited partners are not required to take on risk beyond their investment. The arrangement not only provides an easy way for investors to supply money, but also is suited for collecting small amounts from a wide range of investors.
The requirement for limited partnerships in Japan to invest at least 50% domestically hampers foreign investors that see Asia as a whole as a target. Removing the restriction would help Japanese venture capital funds expand their investment targets overseas, and an increase in money flowing in the domestic venture capital market ultimately will benefit Japanese startups.
Japan's government will submit to parliament a draft amendment to the law governing limited partnerships in 2024. The cap on investing in foreign companies was introduced as part of the law on limited partnerships that took effect in 1998.
Japan has set a goal of increasing annual investment in startups to the scale of 10 trillion yen ($77.2 billion) by fiscal year 2027, more than a tenfold increase from current levels.

Cowboy Ventures, the now-10-year-old, Bay Area-based seed-stage focused fund founded by renowned investor Aileen Lee, has closed on two new funds totaling $260 million in capital commitments. The outfit garnered $140 million in commitments for its fourth flagship fund and another $120 million for its first opportunity-type fund (its “Mustang Fund”).
The amount is more than all the capital that the outfit has raised across its previous funds, which were sized at $40 million, $60 million and $95 million, respectively. Then again, the team has grown over the years from being a one-person firm to an outfit with an investor team, including fintech specialist Jill Williams, who Lee recruited from Anthemis, and Amanda Robson, who was pulled out of Norwest Venture Partners, where she worked with numerous enterprise software companies, including some focused on AI and robotics. (Longtime Silicon Valley attorney Ted Wang is also closely associated with the fund as a “board partner” and advises more than a dozen of its portfolio companies.)
It’s easy to appreciate why LPs committed more capital to Cowboy, even in a market that seems to be actively shrinking given broader market turmoil.
First and foremost are its numbers, which look good, particular given the size of its earlier funds. Cowboy was among the first investors in Guild Education, for example, an online education company that’s focused on upskilling frontline employees, and was valued at $4.4 billion when it closed its most recent round of funding in June of last year. Cowboy is also a seed investor in the security and compliance automation platform Drata, assigned a $2 billion valuation in December when it raised $200 million in Series C funding.

Image Credits: Cowboy Ventures
In conversation with Lee, Williams and Robson late last week, Lee noted that Cowboy thinks of itself as a generalist firm, but that 70% of its most recent fund was funneled into enterprise startups and 30% into consumer startups, given Cowboy has also enjoyed success with the latter. (Most notably, one of its first checks went to Dollar Shave Club, the men’s grooming company acquired by Unilever in 2016 for a reported $1 billion.)
Others of the firm’s bets include Vic.ai, a startup that’s automating accounting processes and just closed a $52 million Series C round in December; Homebase, a platform for small to mid-size businesses that helps with scheduling, payroll, cash advances and HR stuff and has raised roughly $100 million from investors to date; and SVT Robotics, whose software organizes robots in warehouses and factories (it closed $25 million in Series A funding in late 2021).
Lightspeed’s investment team shares what’s fascinating them and what to watch for in the year ahead

Our global fintech team covers all sectors — from infrastructure, to enterprise application layer, to consumer across India, China, South East Asia, Europe, Israel, and the Americas.
As we collected our thoughts on what’s ahead, the world is experiencing rising interest rates, falling equities, and rampant inflation. It’s clear the global financial system is more interconnected now than ever. Our team’s belief in a world where finances are a seamless part of everyone’s lives, not a frustration, is stronger than ever.
We see fintech leading the way toward a financial ecosystem that is accessible, affordable, transparent, data driven, frictionless, cross-border, and convenient. Below are key areas we’re watching in 2023 that advance this vision:
……
The reported $10 billion investment in OpenAI will keep the hottest AI company on Microsoft’s Azure cloud platform.
By Sara Morrisonsara.morrison@recode.net Jan 23, 2023, 3:10pm EST

John Smith/VIEWpress/Corbis via Getty Images
Sara Morrison is a senior Vox reporter who covers data privacy, antitrust, and Big Tech’s power over us all.
Microsoft revealed last week that it will lay off 10,000 people throughout 2023. But don’t think that means the company is having money problems. On Monday, the company announced that it’s investing billions of dollars into the hot artificial intelligence platform OpenAI.
This is Microsoft’s third investment in the company, and cements Microsoft’s partnership with one of the most exciting companies making one the most exciting technologies today: generative AI. It also shows that Microsoft is committed to making the initiative a key part of its business, as it looks to the future of technology and its place in it. And you can likely expect to see OpenAI’s services in your everyday life as companies you use integrate it into their own offerings.
Microsoft told Recode it was not disclosing the deal’s specifics, but Semafor reported two weeks ago that the two companies were talking about $10 billion, with Microsoft getting 75 percent of OpenAI’s profits until it recoups its investment, after which it would have a 49 percent stake in the company. The New York Times has since confirmed the $10 billion amount.
With the arrangement, OpenAI runs and powers its technology through Microsoft’s Azure cloud computing platform, which allows it to scale and make it available to developers and companies looking to use AI in their own services (rather than have to build their own). Think of it as AIaaS — AI as a service. Microsoft recently made its OpenAI services widely available, allowing more businesses to integrate some of the hottest AI technologies, including word generator ChatGPT and image generator DALL-E 2, into their own companies’ offerings.
Meanwhile, OpenAI also gets a needed cash infusion — key for a company with a lot of potential but not much to show in terms of monetization. And Microsoft can offer something to its cloud customers that rivals Google and Amazon can’t yet: one of the most advanced AI technologies out there, as well as one of the buzziest. They do have their own AI initiatives, like Google’s DeepMind, which is reportedly rolling out a ChatGPT rival at some point. But it’s not here yet. ChatGPT is, and it’s gone mainstream.
January 26, 2023
Venture giant New Enterprise Associates announced Thursday it has closed on approximately $6.2 billion across two funds.
The firm, known for investments in transformational startups like Salesforce
, Workday and Robinhood, said one fund will be dedicated to early-stage investing and the other to venture growth-stage opportunities.
NEA will invest in a range of different technology and health care sectors — including both enterprise and consumer technology, digital health, and life sciences.
The firm had more than $25 billion of assets under management as of the end of 2022.
“We are deeply grateful to our limited partners for the trust they have placed in our team, and excited to have raised the largest pool of capital in NEA’s history at a time of great uncertainty, but also tremendous opportunity,” said Scott Sandell, managing general partner at NEA.
Founded in 1977, NEA is one of the oldest VC firms in the country. The firm has had more than 270 portfolio company IPOs and over 450 mergers and acquisition events.
Just in the last several months, NEA has participated in rounds worth $100 million or more for companies such as Escient Pharmaceuticals, MBX Biosciences and Swift Navigation.
The announcement comes at a time when many in the industry believe fund managers may have a hard time raising new funds as limited partners and institutions try to avoid risk and rebalance their portfolios after the beating many public tech stocks have taken.
However, such an environment is likely to favor more established firms such as NEA and those who have a proven track record.
Just last week, another seasoned firm, Sequoia, announced a new $195 million fund for seed-stage startups, according to Forbes.

Image Credits: Substack
Newsletter platform Substack announced today that it’s introducing several new features, including private Substacks. A private Substack is a publication that you can host alone or readers can request to subscribe to read your posts. Writers can choose to approve or decline each subscription request.
In a blog post, Substack explained that private Substack accounts work similarly to private Instagram profiles. A private Substack can be used to keep in touch with friends, build communities of interest and test the waters for a new publication, the company says.
Users can change their Substack from public to private at any time by navigating to their settings and selecting “Private” in the “Import” section. Once a user makes their publication private, readers won’t be able to see any posts. When a reader requests to subscribe to a Substack, the writer will receive an email notification with their details. You can see your requests on your Subscribers page. If you approve a request, the reader will be automatically subscribed and sent a Welcome email.
The launch of the new feature comes as Substack has been hoping to capitalize on Twitter’s upheaval following Elon Musk’s takeover. The company openly targeted Twitter’s user base in the past few months and recently threw its hat into the ring as a more direct competitor with the launch of Substack Chat, which allows writers to communicate directly with their loyal readers right in the Substack mobile app.

Share Dialog
Share Dialog
No activity yet