Latest / The Learning Corner by Precursor / Episode #76: My Biggest Founder Regret, Fundamental Truths in VC, SpaceX Cursor Partnership

Transcript

Episode #76: My Biggest Founder Regret, Fundamental Truths in VC, SpaceX Cursor Partnership · The Learning Corner by Precursor

• 0:00
  Strategy always looks easier in a business school case study than it does in real life
• 0:05
  because in real life, you don't actually know what happens at the end.
• 0:07
  I've had portfolio companies of ours sort of had somebody sneak up on them and eat a part of
• 0:12
  their lunch and vice versa.
• 0:13
  So I think this is one of the challenges in dynamic tech markets is you don't always know
• 0:17
  what new feature or what new thing is really gonna matter at the end.
• 0:20
  It's the week of April 20.
• 0:22
  Welcome to the Learning Corner, a Precursor Ventures podcast.
• 0:28
  Welcome to the Learning Corner by Precursor.
• 0:30
  I'm Mia Farnam, and I'm joined by RGP and managing partner Charles Hudson, where we will
• 0:34
  be talking through the articles and Goodreads that caught our team's attention over the past
• 0:38
  week.
• 0:39
  Our first Goodread this week is a LinkedIn post from Manny Medina, cofounder of Outreach, and
• 0:44
  his reflection on a lesson he wishes he had fully embraced from the start that the job of a
• 0:48
  VC backed founder is to win, and winning means actively trying to kill your competition.
• 0:53
  Early on at Outreach, Manny had no choice but to outmaneuver rivals who had raised
• 0:58
  dramatically more capital, and it worked.
• 1:00
  But over time, he says he was talked into going soft by conventional wisdom like focus on your
• 1:05
  own race, and there will be many winners.
• 1:07
  Advice he now calls out as fundamentally misaligned with what founders owe their
• 1:12
  investors.
• 1:13
  He points to real examples of companies that won by acquiring or suffocating competitors,
• 1:18
  DiscoverOrg, OneTrust, and Salesforce versus his own company that held back and watched
• 1:23
  billion dollar businesses like Apollo and Gong fill the gaps they left open.
• 1:27
  The post mentions of observation at the end.
• 1:30
  Uber's market cap is 30 x that of Lyft.
• 1:32
  And while everyone can talk about Uber's leader, most people can't even name who ran
• 1:37
  Lyft.
• 1:38
  Manny's framing here is really clear that killing competition is your job as a VC backed
• 1:45
  founder.
• 1:45
  We've, the podcast and then internally on our team Yeah.
• 1:48
  Also talked about the risks of having your head too much on a swivel.
• 1:51
  So it's difficult to actually focus on your customers when you're so locked in on people
• 1:57
  building around you.
• 1:58
  Wanted to pause for your reflections here, Charles.
• 2:01
  Yeah.
• 2:01
  The first time I read it, I was like, oh, there's something in here that resonates, but
• 2:05
  I'm not sure I would describe it in exactly the same way that Manny did.
• 2:09
  I started thinking about the companies he mentioned, Uber and Salesforce in particular.
• 2:14
  And I think it wasn't even as much about killing the competition as it was having this
• 2:18
  really big holistic view of what you're trying to do.
• 2:21
  And whenever I talked to my friends who worked at Uber in this ascendant days, they had this
• 2:25
  like world domination idea that we're gonna be in a lot of businesses, and we're gonna compete
• 2:30
  in a lot of geographies.
• 2:32
  And that's what it's gonna take to become the global leader in transportation, logistics, and
• 2:36
  delivery.
• 2:37
  And I think it was more about the breadth of the vision required them to win in all of the
• 2:42
  and remember, like, didn't win everywhere.
• 2:44
  They spent a lot of money.
• 2:45
  But I will say Uber competed in every region, in every geo, in every category that they went
• 2:49
  into.
• 2:49
  And if I think about Salesforce, I think my sense has always been that's a company that
• 2:53
  wanted to be one of the, if not the most important software vendor to their customers.
• 2:58
  And to do that, you have to provide more of the things that your customers need.
• 3:02
  And again, not every acquisition that Salesforce has made has worked out.
• 3:06
  But Salesforce has always been, I think, aggressive in trying to make sure that if one
• 3:10
  of our customers has a business problem, whether it's customer service, whether it's
• 3:14
  CRM, we should have a solution for them.
• 3:16
  And I think it to me, it's more about two things.
• 3:19
  What does it mean to kind of dominate your category?
• 3:22
  Well, sometimes dominating your category does mean you have to snuff out competition early to
• 3:26
  not give them room.
• 3:26
  And sometimes it means figuring out, like, well, how much space can we really take up in
• 3:30
  our customers' lives?
• 3:31
  And how do I make sure that we're executing on a strategy that maximizes that?
• 3:34
  And to be fair to many, I don't think a lot of investors thought that companies like Gong and
• 3:39
  Apollo I met the Apollo team early.
• 3:41
  I was like, this is a really neat feature or a really neat tool.
• 3:45
  And so I think the the danger is if you try to kill everything, like, you don't build anything
• 3:50
  on your own.
• 3:50
  You're just being very reactive.
• 3:52
  And it's hard sometimes.
• 3:53
  You know, we have to put this in our team meeting.
• 3:54
  Strategy always looks easier in a business school case study than it does in real life
• 3:59
  because in real life, you don't actually know what happens at the end.
• 4:02
  And so we've been on both ends.
• 4:04
  I've had portfolio companies of ours sort of had somebody sneak up on them and eat a part of
• 4:08
  their lunch and and vice versa.
• 4:09
  So I think this is one of the challenges in dynamic tech markets is you don't always know
• 4:13
  what new feature or what new thing is really gonna matter in the end.
• 4:16
  I wanna talk about the piece that Manny mentioned earlier about competitors who raise
• 4:20
  far more capital than you did, which seems to be happening more and more, especially if
• 4:25
  you're building with an AI.
• 4:26
  If that is the space that you're in as a founder and you see a quasi competitor or a
• 4:31
  direct competitor come out of stealth having raised three to four x the amount of capital
• 4:35
  that you did, how do you make the decision as a founder of how much to take that into
• 4:40
  consideration when you're running after the same customers?
• 4:44
  Or do you really try to just stay focused on the path that you have already outlined for
• 4:49
  yourself?
• 4:50
  I find that that's Yeah.
• 4:51
  Just something that comes up in meetings with founders oftentimes when they it's like the
• 4:55
  elephant in the room almost.
• 4:57
  As investors, I've never felt like it's a helpful thing to bring up of like, oh, did you
• 5:00
  see the 30,000,000 like, it's just obviously, they've seen it.
• 5:04
  I don't unless I have something really valuable to share about how they can use that to their
• 5:09
  advantage, I just don't think it's a good use of time.
• 5:11
  But I don't know sometimes if founders know where to prioritize after seeing stuff like
• 5:15
  that.
• 5:15
  I just talked to a founder in our portfolio who's selling her company.
• 5:20
  And she said, look, a bunch of my competitors have raised a ton of money.
• 5:24
  None of their businesses are really working, but a, they've sort of sucked the fundraising
• 5:29
  oxygen out of the room.
• 5:30
  And b, like, they're engaging in business practices that, like, don't make any sense from
• 5:36
  a pricing go to market standpoint.
• 5:37
  And she's just like, I don't know if any of them will win, and I'm not gonna pick on
• 5:41
  specific companies.
• 5:42
  But I can think of a few categories where VCs piled in, funded a bunch of other companies,
• 5:46
  and then the VCs moved on, but those founders were stuck with those businesses.
• 5:50
  They raised $50.60, $70,000,000 to dominate a category and turned out the category either was
• 5:55
  harder to dominate or was less interesting than originally thought, and they're still running
• 5:58
  those businesses.
• 5:59
  And I'm like, well, to what end?
• 6:01
  My sense is when people raise lots of money, it tends to create some fairly predictable
• 6:06
  distortions in your category.
• 6:08
  They tend to spend a lot more money on GTM, so they will flood the zone from a marketing
• 6:13
  standpoint.
• 6:14
  The smart companies will spend that money in ways that actually create value.
• 6:18
  A lot of it's wasted.
• 6:18
  They will look around at like, hey.
• 6:20
  We have product and technology gaps.
• 6:22
  We should maybe go acquire companies that can help us solve those things, and cash makes that
• 6:27
  easier.
• 6:27
  I think the third thing is I tend to find the people who are able to raise those routes are
• 6:31
  naturally more aggressive Mhmm.
• 6:32
  Or can convince VCs that their vision for the category or the product is more expansive.
• 6:38
  So it's kind of like a what's first?
• 6:40
  Is it the expansive vision that comes first, or is it the money that enables expansive vision?
• 6:45
  I'm pretty sure it's the expansive vision that comes first.
• 6:47
  And then once you've got the $30,000,000, you probably told the investors, hey, we're gonna
• 6:50
  flood the zone with marketing.
• 6:51
  We're gonna go acquire some smaller companies, and we're gonna patch this thing together, and
• 6:55
  we're gonna build the brand.
• 6:56
  And yes, is it a Frankenstein company?
• 6:58
  Sure.
• 6:58
  All those all those things apply.
• 7:00
  But I think for the people who are wired that way, that playbook's never been easier to
• 7:03
  execute than it is right now.
• 7:05
  The one thing that I do want founders to prioritize, or they just have to if they're
• 7:10
  going out to raise money and they have competitors who have raised much more capital
• 7:13
  than they have in the market, is like you have to be able to speak to the story as to what
• 7:16
  your competitors have done that you're not going to be doing that's going to be more
• 7:20
  effective.
• 7:20
  Because to your point, I find that's actually the likely scenario.
• 7:24
  There's been way more money that's flooded into a specific industry that leaves scar tissue
• 7:28
  with investors.
• 7:29
  So it's not like, oh, these wonderful experiences make me more excited to jump in.
• 7:33
  I've seen that a lot with health care.
• 7:35
  If we've done this back in 2021, we're all set.
• 7:38
  So I would say for founders who see competitors building, if you were choosing to wanna stay
• 7:42
  your path, just know that you do have to speak to the folks who are in your own space and how
• 7:46
  you're gonna be differentiated and maybe what they've left open as gaps that you can take
• 7:50
  over.
• 7:51
  You have to be educated about that space for sure.
• 7:53
  And and I think you have to be realistic.
• 7:55
  I sometimes Yeah.
• 7:56
  Hear founders tell me a story and I'm like, I understand why you want that to true.
• 8:00
  That, oh, you know, the product sucks and this and that.
• 8:03
  But is it true?
• 8:04
  And I think it takes a lot of intellectual honesty to be able to say that, like, yeah,
• 8:08
  it's frustrating when your when your competitor who you think has an inferior product raises a
• 8:12
  lot of money.
• 8:12
  The dismissal can't be, well, their product sucks.
• 8:15
  Well, they have a lot of money.
• 8:15
  They're probably doing something.
• 8:16
  So let's at least be honest about what are the things that they're doing that we cannot do?
• 8:20
  Yeah.
• 8:20
  Also, the product could suck, but does that even matter?
• 8:23
  Like, if I I would say for most b to b SaaS features and products that are sold at Fortune
• 8:28
  500 companies, they're not good products.
• 8:29
  But they have multi million dollar deals under the belt, which means that it stinks.
• 8:33
  But people are happy to use this, like, half baked product because it's what they've used
• 8:37
  before.
• 8:38
  So being realistic is such a throughput here.
• 8:40
  Our next Goodread is also a LinkedIn post from Samir Kaji, CEO of Allocate and host of Venture
• 8:46
  Unlocked, sharing what he calls his fundamental truths in VC right now.
• 8:49
  It's a really tight list of observations, but a few stood out.
• 8:53
  The barbell between large and small VC firms is wider than it ever has been, and LPs should
• 8:59
  treat them as entirely distinct asset classes.
• 9:02
  Zmir also flags that no portfolio construction model, whether concentrated or diversified,
• 9:07
  rescues a bad portfolio.
• 9:10
  And that the real variable is whether you are in great companies to begin with.
• 9:13
  He raises a quiet but growing concern around ARR reporting, noting that founders are
• 9:18
  increasingly coached to show growth to maintain valuations, and that is getting dangerous.
• 9:23
  He closes on a note of honest uncertainty about AI, arguing that the people who sound most
• 9:28
  certain about how it all plays out tend to be the most financially invested in a particular
• 9:32
  outcome being true.
• 9:34
  Samir says that the barbell between large and small firms is as wide as it's ever been, and
• 9:39
  they're just genuinely two different asset classes now.
• 9:41
  As a seed investor, how do you think about where we sit or maybe, you know, folks who've
• 9:47
  started funds back ten years ago are and haven't grown their fund sizes out too
• 9:53
  tremendously just to keep that same strategy?
• 9:55
  And has the pressure from both ends changed how you see the firm operating at all?
• 10:00
  I didn't realize this when you picked out.
• 10:02
  There's like a through line with the last article too, which is, you know, we talked
• 10:05
  about what do you do when you're a portfolio company.
• 10:08
  I think the same question applies to venture capitalists.
• 10:10
  You know, we've kept our funds small for out of a mix of strategy, personal preference, a lot
• 10:15
  of things.
• 10:16
  And I tell a lot of I told a of people last year, it'd be very comforting to think that
• 10:20
  you're allowed to execute whatever strategy you want, but the market has the vote too.
• 10:24
  And so I think part of what's happening in Venture, we've this has been, I think, the
• 10:28
  theme of the last year, is this continued pulling apart of small and large firms.
• 10:33
  And, you know, as a small fund, you gotta remember, there's a lot of short term
• 10:37
  incentives for you to to mimic the investment decisions of big firms.
• 10:41
  I just saw someone else posted a one of these charts.
• 10:43
  These are the seed funds that have the biggest follow on uptake rate by the big tier one firm.
• 10:47
  Like, great.
• 10:48
  I don't I don't know what I don't know what to do with that information other than to say,
• 10:50
  like, congratulations.
• 10:51
  Like, doesn't necessarily mean that those companies will be more successful or better.
• 10:54
  It is just a data point.
• 10:56
  So I think about I think about this a lot.
• 10:58
  Yeah.
• 10:59
  I'd like to say I spent most of last year re underwriting the whole use case of small firms
• 11:04
  in venture.
• 11:05
  And I I came away with one big conclusion, which is on average, to make a broad
• 11:09
  generalization, most of the larger players in our space, this whole legibility.
• 11:13
  This is one of those words that Michael, but I get it.
• 11:15
  They're looking for legibility as in, like, founders with backgrounds and working in
• 11:19
  categories that are easy to understand and underwrite for them.
• 11:23
  So people who went to a narrow set of schools, people with a really specific set of
• 11:26
  backgrounds, maybe they graduated from an accelerator, know.
• 11:29
  That is not going to be the totality of returns in venture.
• 11:31
  If it ever becomes 100%, then everybody else should just go home.
• 11:35
  But I think we're far from that world.
• 11:37
  And so I do think that no matter what your strategy is in venture, you should constantly
• 11:42
  re underwrite it every twelve to twenty four months and say, like, is the thing I'm doing
• 11:46
  does it still make and I'm I'm I've talked to a lot of friends who are simultaneously
• 11:49
  increasing their fund size and becoming even more concentrated because they're like, look,
• 11:54
  the only way to make the returns math work is we need more money to compete for the best
• 11:58
  deals.
• 11:59
  And when we win them, we have to own more of them.
• 12:01
  And I'm like, that is a strategy that's based on a set of assumptions about where venture is
• 12:06
  going.
• 12:06
  And if you're right, then that's the right strategy.
• 12:08
  If not, you're gonna end up with a gigantic hole.
• 12:10
  So I think strategy making right now for VCs is no easier than it is for founders.
• 12:14
  And the other thing I will say is that I read everything Sameer writes.
• 12:17
  I would say a subtle tone shift from him expressing caution about AI.
• 12:21
  He's been one of the most bullish, positive, I wouldn't say unabashed, but really positive AI
• 12:27
  voices I know.
• 12:28
  And I think I too have had the same questions and concerns around ARR reporting and, you
• 12:32
  know, forcing people to grow.
• 12:34
  It's dangerous.
• 12:35
  So Yeah.
• 12:36
  It's really a tough time.
• 12:37
  Do you think that last piece that you mentioned of making sure you're being cautious around
• 12:42
  reporting, do you think that that's something that is a call for maybe LPs to carry out a
• 12:47
  little bit more heavily?
• 12:48
  Because I think in our job, we've known this to be true for the last year and a half.
• 12:51
  I think part of the challenge is the reward to caution is you maybe don't raise your route.
• 12:55
  And the question is, like, are companies fooling VCs or are VCs in on it too?
• 13:00
  I kinda think it's more of the latter.
• 13:02
  You don't have to usually scratch too hard to figure out, is this ARR number, is it all
• 13:06
  deployed or is some of it contracted?
• 13:08
  As someone once said, they have to pay you twice before you can call it recurring.
• 13:11
  There there's a lot of things that but the question is, like, how hard do you wanna look?
• 13:15
  And I I don't think that many VCs I know are actively being misled.
• 13:19
  I think they're choosing to say, look, there's a chance that this isn't all recurring.
• 13:23
  There's a chance that this is a little bit funny.
• 13:25
  But the growth is good, like, everybody's got a little bit of this funny on the books.
• 13:30
  And so as long as we're not investing in fraud, it's probably okay.
• 13:34
  Yeah.
• 13:34
  That that is a position to take.
• 13:36
  I was at a breakfast yesterday, and it was a mix of, like, LPs and GPs.
• 13:40
  And one person was at a CBC of a large Fortune 500 company.
• 13:44
  I'm not gonna name which one.
• 13:45
  And he mentioned that in the past year alone, their company signed off for 02/1963 pilots for
• 13:53
  new AI tools.
• 13:54
  Two of them went through the contract.
• 13:57
  It was Adobe and another company that was not any of the fancy AI companies because I would
• 14:02
  have remembered it.
• 14:02
  It was not that.
• 14:03
  And so that really does also reinforce young companies that you see that state a lot of
• 14:08
  really exciting ARR numbers.
• 14:09
  You have to really hammer in of, like, how much of this is pilot?
• 14:12
  And I asked him too.
• 14:13
  I said, like, how long are the time lengths for the pilots?
• 14:16
  He said, all range.
• 14:17
  It could be a month.
• 14:18
  It could be a year.
• 14:18
  They're all different types.
• 14:19
  We had them simultaneously running, but only two of them went through.
• 14:22
  Do you think 2026 and maybe even the beginning of 2027 will illuminate any reckonings between
• 14:28
  what's real and what's elaborated, or do you think we've got still a little bit more to go
• 14:32
  here?
• 14:32
  I do for a very different reason.
• 14:34
  It's a small thing, but a lot of people I don't notice.
• 14:36
  You know, Anthropic took Claude Code out of the pro plan.
• 14:40
  I use Mac, so it's fine.
• 14:41
  I do think one thing that's gonna happen is we're definitely in the $5 get across town in
• 14:46
  San Francisco Uber era of tokens.
• 14:49
  And I do think as that gap between true cost and cost of the consumer begins to close, some
• 14:55
  things right now that are not worth paying attention to because they're little
• 14:57
  experiments.
• 14:58
  People are gonna say, wait a minute.
• 14:59
  How much are we spending on that thing?
• 15:00
  Are we really getting anything out of it?
• 15:02
  And I think prices will go up to consumers, to business consumers, and scrutiny will go up.
• 15:07
  I think the other thing is many of these things have been running long enough that people ought
• 15:10
  to be seeing some ROI from them.
• 15:12
  And so I think there's gonna be a combination of factors that are gonna make it harder to
• 15:16
  sell the AI dream without some reality in the back half of this year and all of next year.
• 15:21
  Yeah.
• 15:21
  I assume that founders will have to be able to show the ROI much quicker than you would be
• 15:26
  able to kind of like show a little AI dance before and by like six or seven months to just
• 15:31
  kind
• 15:31
  of Yeah.
• 15:32
  Win people over.
• 15:33
  Our last piece this week is a post on X from Tuesday night from SpaceX announcing a new
• 15:38
  partnership with Cursor to build what they're calling the world's most useful models for
• 15:42
  coding and knowledge work.
• 15:44
  Cursor brings leading product and distribution to expert software engineers and SpaceX brings
• 15:49
  Colossus, their million h 100 equivalent training supercomputer to the table.
• 15:53
  What makes this one really worth discussing is the deal structure attached to it that we
• 15:57
  really only know service level on of of what's been shared.
• 16:00
  SpaceX has been granted the right to acquire Cursor later this year at a $60,000,000,000
• 16:04
  valuation or alternatively pay $10,000,000,000 for the work produced through the
• 16:08
  collaboration.
• 16:09
  For context, Cursor only recently crossed 500,000,000 in ARR, making a $60,000,000,000
• 16:15
  price a jaw dropping multiple and a signal of just how aggressively the market is pricing AI
• 16:20
  infrastructure and distribution right now.
• 16:22
  That deal structure obviously is is what caught my eye here.
• 16:25
  SpaceX essentially has an option to buy it at 60,000,000,000 or pay 10,000,000,000 for the
• 16:29
  output of the partnership.
• 16:31
  What does that structure tell you about how companies are thinking about locking in AI
• 16:35
  capability before the window, quote unquote, closes?
• 16:38
  It reminds me of a very convoluted Microsoft OpenAI deal back in the day where they're gonna
• 16:43
  buy some compute.
• 16:44
  The person who sent me this right before it got announced, I was like, oh, wow.
• 16:47
  Are you sure?
• 16:48
  I but immediately, I was like, are you sure this is from a reputable source, or were you
• 16:52
  spending too much time?
• 16:52
  He's like, no.
• 16:53
  No.
• 16:53
  It's gonna come from SpaceX.
• 16:54
  And it it did a couple minutes later.
• 16:56
  And I asked him, I said, who's who's saving who?
• 16:59
  And he's like, what do you mean?
• 17:00
  I go, is this Cursor helping Grok remain relevant and get more value out of Colossus?
• 17:06
  Or is this Grok helping Cursor remain relevant by giving them a bunch of compute for their new
• 17:12
  for the Kimi based, their their new set of models?
• 17:15
  And he said, could it be both?
• 17:16
  I'm like, yeah.
• 17:17
  Actually, it could be both.
• 17:19
  And so I think there's like a whole another layer, which is like, if you believe SpaceX
• 17:23
  will go public and that the company's valuation will hold, even in remotely the one to two
• 17:29
  trillion dollar range, $10,000,000,000 means nothing.
• 17:31
  I thought about this a was like, it's actually, for me, fairly interesting that in the end, the
• 17:36
  smartest thing to do might have been to sell Cursor at 30 or 40 or 50 if you had the chance.
• 17:41
  In the grand scheme of like how this all plays out, if it if it trades out at 10, and, you
• 17:46
  know, I don't know all of the conditions that have to be met for that $10,000,000,000 payout
• 17:49
  to be made.
• 17:50
  These things are it's never that simple.
• 17:52
  But it might turn out that a company that we once thought was worth nothing and then thought
• 17:56
  would be the dominant coding player might end up being a very valuable company, but not
• 18:01
  nowhere near as valuable as we had once thought.
• 18:03
  Or it could turn out to be an incredibly valuable company.
• 18:06
  We were all wrong to be skeptical or doubtful at all.
• 18:09
  But a lot of this is contingent on SpaceX's IPO, and the question I keep coming with, where
• 18:14
  do you find that much demand?
• 18:17
  Even in Robinhood, like, the irony is a lot of the big institutions that you would go to for
• 18:21
  the IPO, they're already on the books on the private side.
• 18:24
  So this is a this is an enormous amount of money to raise in the context of an IPO.
• 18:28
  But I just immediately thought, like, well, what's $10,000,000,000 or even $60,000,000,000
• 18:33
  to a $2,000,000,000,000 company?
• 18:34
  It's actually not very much money in the grand scheme of things.
• 18:37
  And and it gives them the ability to have and again, of course, our really deterrence, look,
• 18:42
  I assume that Graca's probably building their own coding agent too or wants to have some
• 18:45
  ability to do that.
• 18:46
  So I think this is a very interesting tie up.
• 18:49
  I'll be curious to see whether the the end price ends up being closer to 10,000,000,000 or
• 18:53
  60,000,000,000.
• 18:54
  It is interesting too because the deal terms that we'll never see likely are what will be
• 18:59
  interesting to parse out because right now, Cursor's trained on Anthropic, OpenAI's models.
• 19:03
  And so I wonder, like, is this they have to kind of, like, co locate all of their training
• 19:07
  on what SpaceX, XAI, Grok is doing?
• 19:10
  Or they're really just like cruising and and sharing distribution?
• 19:13
  I don't I don't know.
• 19:14
  My take is, like, you know, the the last one they released, they used the Kymi model.
• 19:18
  They're they're musing, I think, wisely, trying to find good open source models so that they're
• 19:22
  not dependent on their competitors for infrared models.
• 19:26
  And I don't use cursor anymore.
• 19:27
  I've sort of gone all in on Cloud Code.
• 19:29
  But I have friends who use cursor who've been very happy with the new models.
• 19:32
  And obviously, they're open source.
• 19:33
  They're much cheaper to serve.
• 19:35
  And if you had, you know to me, the logic of the deal actually makes sense.
• 19:39
  If you're a cursor, like, I gotta find a lot of compute somewhere.
• 19:42
  Yeah.
• 19:42
  But, like, our cost to compute on these models is much lower than it would have been if we
• 19:45
  were using one of our former partners.
• 19:47
  And if you're Grok, you're like, I got all these I got all this compute.
• 19:49
  I gotta find somebody to use it.
• 19:51
  Sure.
• 19:51
  So it kinda makes sense.
• 19:52
  We are all out of time.
• 19:54
  Thank you, Charles, for jumping in.
• 19:56
  Thank you to our listeners for listening in as well.
• 19:58
  Please like and subscribe, and we will talk with you all next week.
• 20:01
  Thank you.
• 20:02
  Bye.
• 20:08
  Wow.