I don't understand why Paul Graham is quoted here so much, given that I have never once seen him say something above the level of GPT3. I get that he's a well-connected schelling person it might be useful to cultivate in VC circles, but I don't get seeing him as a source of information instead of an NPC to be played.
Imagine reading this essay in a world where the conventional wisdom was that you should get an MBA to learn the relevant skills to start a company. https://paulgraham.com/start.html
Yeah, this is the best understanding I have of why people a generation older than me seem to hold him in unreasonably high regard. If he had interesting takes that were unusual for an authority figure when they were young and making life plans, he'd get tenured into a kind of "favourite teacher" status that doesn't really get updated away from just because the things he's saying are now boring cliches and he doesn't think of anything new.
Criticizing Paul Graham's advice is like criticizing Tolkien because it sounds just like every other fantasy novel. So much of his correct advice has simply become conventional wisdom, taken for granted, repeated daily by hundreds of people connected to the startup community.
Listening to Paul Graham is like hiring a personal trainer. If you're in the gym, working hard, your natural inclination is to slack off and relax. The personal trainer helps you stay focused and push yourself to your limits. If you aren't in the gym, and you're simply reading a transcript of the advice given by the personal trainer, it might not seem all that brilliant. Of course you should push yourself to do a few more reps! Of course you should talk to users and make something people want! And having someone hanging around short selling your workout isn't really going to help you.
I say "Paul Graham", but the same goes for most YC partners. The standard "Paul Graham thought" is good for startups, YC can provide it, and if you want to start a startup it's a great deal.
Not really? Tolkien is a really good writer who subsequent writers mostly fail to emulate (notably, if you train GPT3 on Lord of the rings, you will absolutely not get a Tolkien level piece of writing).
(The fact that Graham doesn't have anything original or interesting to say alsoakes me suspect that he wasn't novel twenty years ago either, rather than just being a schelling point for what was then a much smaller group's conventional wisdom.)
I think the personal trainer analogy holds. But it'd be weird to take a famous fitness influencer and block quote him saying "squats are good" if he didn't provide any nontrivial insight. It's not a quotable statement.
I have known Ayahuasca success stories - for treating PTSD. The most extreme is a friend who was heavily traumatized, could barely leave the house, and considering suicide. He credits aya with saving his life.
I don't get the use case for successful tech CEOs though. It's not something to use as a general psychedelic trip. I just don't want to see it written off entirely either when it can have significant benefits for trauma.
Everything I've read about psychedelics makes me think they're like rolling a randomiser on your brain - if it's badly broken when you start, the upside potential is high; if your brain is working fine before you start, why risk that it might not be afterwards?
"We raised $20m at a $100m valuation, and realistically it should have been less than half of that. We made damn sure the terms were airtight, so we can safety do a down round if needed, and we are setting aside the extra money so we don’t go too big or spend too much, and we’re setting your options deals accordingly."
In practice, founders are limited in their ability to offer these kind of deals, because tax law essentially requires that the strike price for employee options match an independent 409A valuation of the common stock. That valuation will be heavily anchored on the valuation from the most recent round.
There are two things that founders can do to to right-size employee options in this situation:
- Include valuable "structure" in the preferred stock they sell to VC funds: liquidation preference, anti-dilution mechanisms, etc. This allows them to make the case for a 409A valuation that might be significantly lower than the naive math implied by dollars / percentage ownership, because the common stock should be discounted vs. the preferred. But obviously this involves making real economic transfers to the VC funds.
- Offer employees options for a higher-than-typical percent of ownership to compensate for the fact that the strike price is in some real sense "out of the money". But this gives your employees a payoff curve that neither you or the employee probably prefer. Also, increasing the size of the employee option pool would typically dilute founders and other common stock holders, but not the most recent investors, so you bear the full cost of this.
Section 409A is an underrated source of friction in startup hiring. In practice, it gives VC funds a "most favored nation" clause for equity offerings: Startups can't offer employees stock options with strike prices that are significantly lower than the valuations from their funding rounds. If you as a founder negotiate a sharp deal during fundraising, you have to apply the same sharp deal to your own employees. As a potential employee, I've seen multiple cases where option strike prices were so high that it essentially killed the financial upside of working at a promising company.
This is also a major part of what makes down rounds so painful - employee options had strike prices determined by the previous round, so a down round almost certainly marks to market in a way that makes clear that those options are underwater.
Founders also can't make these types of deals because this is not the anonymous public market. The VC is someone you're going to be working with for years, and you need a good working relationship with them. They will be unhappy if you do a deal with them and then immediately announce "I just ripped this VC off, they gave us a $100m valuation but we aren't worth that much money at all!"
If you get someone on your board of directors who thinks "This is a good investment at $100m because we will take lots of risks and have a 50-50 shot at being a $1B company, and the CEO agrees with this plan" and then you announce that your strategy is to avoid these risks and accept that you're only worth $50m, you're just going to have an unhappy board member who's constantly trying to undermine you.
> If Silicon Valley is to take its rightful place as a true part of Nate Silver’s The River, or simply it wants to make good decisions, this requires a desire to understand the world and be accurate.
Re bad VC deals: herding behavior rules everything around me.
Propsective employees (and to a lesser extent customers) know they can't trust founders telling them that this product is totally going to (a) still exist in 2 years and (b) take over the world in 5. (Startups are less exposed to reputational risks from dissatisfied counterparties than established companies; extreme overpromising and outright deception are rampant.) So they look, among other things, for endorsements from a16z Sequoia etc. to know whether a pitch is legit. So do other VCs, including in subsequent funding rounds.
Startups are a classic anti-inductive regime; their work is novel enough that there's no great way to know it's legit short of doing it yourself. You mostly can't trust inside information because startups aren't repeat players in this signaling game. (Successful repeat founders are the exception that proves the rule-- startups very much market themselves on that!) And getting actionable outside information about a startup's propsects from sources other than the Paul Graham school of talk-to-the-founders can be prohibitively hard. Top VCs are the proverbial one-eyed man here and as such have disproportionate pricing power.
In my past start-up we've done AI-assisted psychedelics design and well, ayahuasca often shows you there're things you're neglecting, or well... that working 14/7 is simply mad? Anyway - I'm starting a new AI safety R&D lab now so it's not like psychedelics kill startup mentality or anything.
Hm, I remember reading an interesting Vox piece by Sean Illing about his ayahuasca excursion some years ago...it sounds pretty intense, one of the more risky "play to your outs" choices for when life's stuck in a really bad rut. Wouldn't touch it for mere recreational purposes with an eleven-foot pole. I myself have done shrooms a few times, which is fun but relatively benign long-term. Also had a...near-fatal acid trip a few weeks back, more in line with what I'd expect from ayahuasca (or Unsong's version of peyote). Obviously would not lightly recommend going up to Death and telling him Not Today, but also that was a big part of what made it a valuable experience? Literally almost dying makes so many quotidian inconveniences seem cosmically unimportant by contrast. I can totally see how this might cash out to "don't wanna be a hard-charging CEO anymore", or maybe even getting fed up with a more usual job/life...the overwhelming throughline of the trip was a pervasive sense of tiredness, of exhaustion from fighting so hard all the time. Kept hearing the siren call to Just Let Go (which would have been fatal). Cf. Scott getting out of the car in Universal Love, Said The Cactus Person. Overall I think the changes were positive; people worry a lot about not rising to meet acute stressors, but doing a brain reformat made me realize the steady erosion of excessively sweating the small stuff was quite bad already.
Unfortunately a lot of Grand Achievements do require some level of self-martyrdom such as obsessing over petty details, so it's still plausibly negative ROI at the societal level for anyone who might've been A Contender. So it makes sense that psychedelics and VC would be a poor mix.
Thanks! Great review.
Podcast episode for this post:
https://open.substack.com/pub/dwatvpodcast/p/startup-roundup-3?r=67y1h&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true
I don't understand why Paul Graham is quoted here so much, given that I have never once seen him say something above the level of GPT3. I get that he's a well-connected schelling person it might be useful to cultivate in VC circles, but I don't get seeing him as a source of information instead of an NPC to be played.
Imagine reading this essay in a world where the conventional wisdom was that you should get an MBA to learn the relevant skills to start a company. https://paulgraham.com/start.html
Yeah, this is the best understanding I have of why people a generation older than me seem to hold him in unreasonably high regard. If he had interesting takes that were unusual for an authority figure when they were young and making life plans, he'd get tenured into a kind of "favourite teacher" status that doesn't really get updated away from just because the things he's saying are now boring cliches and he doesn't think of anything new.
Criticizing Paul Graham's advice is like criticizing Tolkien because it sounds just like every other fantasy novel. So much of his correct advice has simply become conventional wisdom, taken for granted, repeated daily by hundreds of people connected to the startup community.
Listening to Paul Graham is like hiring a personal trainer. If you're in the gym, working hard, your natural inclination is to slack off and relax. The personal trainer helps you stay focused and push yourself to your limits. If you aren't in the gym, and you're simply reading a transcript of the advice given by the personal trainer, it might not seem all that brilliant. Of course you should push yourself to do a few more reps! Of course you should talk to users and make something people want! And having someone hanging around short selling your workout isn't really going to help you.
I say "Paul Graham", but the same goes for most YC partners. The standard "Paul Graham thought" is good for startups, YC can provide it, and if you want to start a startup it's a great deal.
Not really? Tolkien is a really good writer who subsequent writers mostly fail to emulate (notably, if you train GPT3 on Lord of the rings, you will absolutely not get a Tolkien level piece of writing).
(The fact that Graham doesn't have anything original or interesting to say alsoakes me suspect that he wasn't novel twenty years ago either, rather than just being a schelling point for what was then a much smaller group's conventional wisdom.)
I think the personal trainer analogy holds. But it'd be weird to take a famous fitness influencer and block quote him saying "squats are good" if he didn't provide any nontrivial insight. It's not a quotable statement.
I have known Ayahuasca success stories - for treating PTSD. The most extreme is a friend who was heavily traumatized, could barely leave the house, and considering suicide. He credits aya with saving his life.
I don't get the use case for successful tech CEOs though. It's not something to use as a general psychedelic trip. I just don't want to see it written off entirely either when it can have significant benefits for trauma.
Everything I've read about psychedelics makes me think they're like rolling a randomiser on your brain - if it's badly broken when you start, the upside potential is high; if your brain is working fine before you start, why risk that it might not be afterwards?
"We raised $20m at a $100m valuation, and realistically it should have been less than half of that. We made damn sure the terms were airtight, so we can safety do a down round if needed, and we are setting aside the extra money so we don’t go too big or spend too much, and we’re setting your options deals accordingly."
In practice, founders are limited in their ability to offer these kind of deals, because tax law essentially requires that the strike price for employee options match an independent 409A valuation of the common stock. That valuation will be heavily anchored on the valuation from the most recent round.
There are two things that founders can do to to right-size employee options in this situation:
- Include valuable "structure" in the preferred stock they sell to VC funds: liquidation preference, anti-dilution mechanisms, etc. This allows them to make the case for a 409A valuation that might be significantly lower than the naive math implied by dollars / percentage ownership, because the common stock should be discounted vs. the preferred. But obviously this involves making real economic transfers to the VC funds.
- Offer employees options for a higher-than-typical percent of ownership to compensate for the fact that the strike price is in some real sense "out of the money". But this gives your employees a payoff curve that neither you or the employee probably prefer. Also, increasing the size of the employee option pool would typically dilute founders and other common stock holders, but not the most recent investors, so you bear the full cost of this.
Section 409A is an underrated source of friction in startup hiring. In practice, it gives VC funds a "most favored nation" clause for equity offerings: Startups can't offer employees stock options with strike prices that are significantly lower than the valuations from their funding rounds. If you as a founder negotiate a sharp deal during fundraising, you have to apply the same sharp deal to your own employees. As a potential employee, I've seen multiple cases where option strike prices were so high that it essentially killed the financial upside of working at a promising company.
This is also a major part of what makes down rounds so painful - employee options had strike prices determined by the previous round, so a down round almost certainly marks to market in a way that makes clear that those options are underwater.
Founders also can't make these types of deals because this is not the anonymous public market. The VC is someone you're going to be working with for years, and you need a good working relationship with them. They will be unhappy if you do a deal with them and then immediately announce "I just ripped this VC off, they gave us a $100m valuation but we aren't worth that much money at all!"
If you get someone on your board of directors who thinks "This is a good investment at $100m because we will take lots of risks and have a 50-50 shot at being a $1B company, and the CEO agrees with this plan" and then you announce that your strategy is to avoid these risks and accept that you're only worth $50m, you're just going to have an unhappy board member who's constantly trying to undermine you.
> If Silicon Valley is to take its rightful place as a true part of Nate Silver’s The River, or simply it wants to make good decisions, this requires a desire to understand the world and be accurate.
Strongly agree
Boston startup week is Sept8, you are reading zvi’s blog come through :
https://www.startupbos.org/sbw2025
Re bad VC deals: herding behavior rules everything around me.
Propsective employees (and to a lesser extent customers) know they can't trust founders telling them that this product is totally going to (a) still exist in 2 years and (b) take over the world in 5. (Startups are less exposed to reputational risks from dissatisfied counterparties than established companies; extreme overpromising and outright deception are rampant.) So they look, among other things, for endorsements from a16z Sequoia etc. to know whether a pitch is legit. So do other VCs, including in subsequent funding rounds.
Startups are a classic anti-inductive regime; their work is novel enough that there's no great way to know it's legit short of doing it yourself. You mostly can't trust inside information because startups aren't repeat players in this signaling game. (Successful repeat founders are the exception that proves the rule-- startups very much market themselves on that!) And getting actionable outside information about a startup's propsects from sources other than the Paul Graham school of talk-to-the-founders can be prohibitively hard. Top VCs are the proverbial one-eyed man here and as such have disproportionate pricing power.
In my past start-up we've done AI-assisted psychedelics design and well, ayahuasca often shows you there're things you're neglecting, or well... that working 14/7 is simply mad? Anyway - I'm starting a new AI safety R&D lab now so it's not like psychedelics kill startup mentality or anything.
Hm, I remember reading an interesting Vox piece by Sean Illing about his ayahuasca excursion some years ago...it sounds pretty intense, one of the more risky "play to your outs" choices for when life's stuck in a really bad rut. Wouldn't touch it for mere recreational purposes with an eleven-foot pole. I myself have done shrooms a few times, which is fun but relatively benign long-term. Also had a...near-fatal acid trip a few weeks back, more in line with what I'd expect from ayahuasca (or Unsong's version of peyote). Obviously would not lightly recommend going up to Death and telling him Not Today, but also that was a big part of what made it a valuable experience? Literally almost dying makes so many quotidian inconveniences seem cosmically unimportant by contrast. I can totally see how this might cash out to "don't wanna be a hard-charging CEO anymore", or maybe even getting fed up with a more usual job/life...the overwhelming throughline of the trip was a pervasive sense of tiredness, of exhaustion from fighting so hard all the time. Kept hearing the siren call to Just Let Go (which would have been fatal). Cf. Scott getting out of the car in Universal Love, Said The Cactus Person. Overall I think the changes were positive; people worry a lot about not rising to meet acute stressors, but doing a brain reformat made me realize the steady erosion of excessively sweating the small stuff was quite bad already.
Unfortunately a lot of Grand Achievements do require some level of self-martyrdom such as obsessing over petty details, so it's still plausibly negative ROI at the societal level for anyone who might've been A Contender. So it makes sense that psychedelics and VC would be a poor mix.