Latest / The Starting Block / Wall Street Surrenders To Crypto & AI, feat. David Pakman

Transcript

Wall Street Surrenders To Crypto & AI, feat. David Pakman · The Starting Block

• 0:09
  Welcome back to The Starting Block.
• 0:11
  Happy Tuesday to you.
• 0:13
  We were supposed to broadcast
• 0:13
  yesterday, but we had a late pull-up
• 0:15
  with Pierre Richard. So we're hoping
• 0:16
  to get him back on the show a little
• 0:18
  bit later this week to chat to us
• 0:19
  about BIP 110, but we will
• 0:21
  continue with regular programming
• 0:23
  today. If you're joining us for the
• 0:24
  first time, welcome to the show.
• 0:25
  This is The Starting block.
• 0:27
  My name is Gareth Jenkinson.
• 0:28
  I'm the head of multimedia here at
• 0:29
  The Block. And of course, your host.
• 0:32
  It's been a fantastic three weeks so
• 0:33
  far of the new show.
• 0:35
  And if you missed it, we had big
• 0:36
  grand launch with an exclusive
• 0:38
  interview. With Binance founder CZ
• 0:40
  in Abu Dhabi.
• 0:42
  If you wanna go and catch up on that
• 0:43
  interview, it's about 17 minutes
• 0:45
  worth of riveting conversation.
• 0:46
  You can find that on the Block's
• 0:48
  YouTube channel.
• 0:50
  And if you are new here and this is
• 0:51
  the first time you're watching the
• 0:52
  show, I'll give you a brief rundown.
• 0:54
  As you can see above my head, we've
• 0:55
  got four segments of the show.
• 0:57
  Three of which are currently live.
• 0:59
  We start with the Block briefing,
• 1:00
  which is a rundown of the top news
• 1:02
  in the last 24 hours.
• 1:04
  We move on to a market poll segment
• 1:05
  where we give you.
• 1:06
  An update in the markets, and then
• 1:08
  we move on to the hot seat, which is
• 1:10
  an interview with a big newsmaker or
• 1:11
  guest of the day.
• 1:13
  Today, we've got David Pakman
• 1:14
  joining us a little bit later on the
• 1:15
  show. We'll be giving him a proper
• 1:17
  introduction and getting into an
• 1:19
  in-depth conversation with him, if
• 1:21
  you stay with us, but we'll be
• 1:22
  moving straight into the block
• 1:24
  briefing right now.
• 1:26
  So if we bring up our first story of
• 1:27
  the Day, Washington is back
• 1:29
  from recess and crypto is
• 1:32
  at the top of the agenda.
• 1:33
  President Trump and senior White
• 1:35
  House officials are pressing the
• 1:36
  Senate to advance the Clarity Act,
• 1:38
  the sweeping market structure bill
• 1:39
  in a window of just a couple of
• 1:41
  weeks.
• 1:42
  On Monday, Trump urged passage in
• 1:44
  tribute to Senator Lindsey Graham,
• 1:46
  who died unexpectedly over the
• 1:48
  weekend, warning that China and
• 1:49
  other countries would love to seize
• 1:51
  control of what he called a
• 1:53
  major financial happening.
• 1:54
  But it's not a clear runway.
• 1:57
  An ethics fight is looming over the
• 1:58
  bill and the clock is ticking on
• 2:00
  the congressional calendar.
• 2:02
  Staying in Washington, the banks are
• 2:04
  pushing back.
• 2:05
  The American Bankers Association,
• 2:07
  the Independent Community Bankers of
• 2:08
  America, and 76 state
• 2:11
  banking associations have written to
• 2:12
  Senate leaders, Tune and Schumer,
• 2:15
  demanding tighter stablecoin
• 2:16
  language in the Clarity Act.
• 2:18
  Their target is section 404.
• 2:21
  It bans crypto firms from paying
• 2:23
  direct or indirect yield on
• 2:25
  payment stablecoins, but
• 2:27
  still allows activity or
• 2:28
  transaction-based rewards.
• 2:30
  The banking groups say that
• 2:32
  ambiguity could let stable coins
• 2:33
  function as deposit substitutes
• 2:36
  and warn of deposit flight from
• 2:38
  community banks.
• 2:39
  Expect this fight to shape the
• 2:41
  final text in the Clarity Act.
• 2:43
  Then across the Atlantic, the UK
• 2:45
  wants in on tokenization.
• 2:47
  The government has published its
• 2:49
  first roadmap for tokenizing
• 2:50
  wholesale financial markets,
• 2:52
  projecting up to £33
• 2:54
  billion in annual
• 2:56
  economic output and £14 billion.
• 2:58
  In yearly tax revenue by 2035.
• 3:02
  That's if the infrastructure gets
• 3:03
  built fast enough.
• 3:05
  The report from wholesale digital
• 3:06
  markets champion Christopher Willard
• 3:09
  centers on tokenized government
• 3:10
  bonds, stable coins and
• 3:12
  on chain settlement.
• 3:14
  And the warning's pretty blunt here.
• 3:16
  Tokenization has moved beyond
• 3:18
  experimentation and without a
• 3:20
  solid plan, the UK risks losing
• 3:22
  liquidity influence and
• 3:24
  market share.
• 3:25
  A landmark migration out of Japan,
• 3:28
  Progmat, the country's dominant
• 3:29
  digital asset issuance platform
• 3:31
  has completed the move of
• 3:33
  its entire security token business.
• 3:36
  Every active tokenized asset
• 3:37
  worth around 452
• 3:40
  billion yen or roughly $2.7
• 3:42
  billion onto a dedicated
• 3:44
  public avalanche
• 3:46
  layer one.
• 3:47
  It's coming off a private,
• 3:48
  permissioned quarter ledger.
• 3:50
  And ProgMat said, this wasn't a
• 3:52
  straight swap. The architecture was
• 3:53
  redesigned to be chain-agnostic.
• 3:55
  With all existing smart contracts
• 3:57
  ported to EVM without changing
• 3:59
  the behavior of any live
• 4:01
  project.
• 4:02
  A big public chain, a vote of
• 4:04
  confidence from Japanese
• 4:05
  institutional finance for a
• 4:07
  big player in the space, Avalanche.
• 4:09
  Then the Ether accumulation machine
• 4:11
  rolls on.
• 4:11
  Bitmine has added another 27,801
• 4:15
  ETH to its treasury last
• 4:17
  week, taking that up to 5.77
• 4:20
  million tokens.
• 4:21
  That's 4.8% of
• 4:23
  all the Ether.
• 4:24
  In circulation worth roughly 10
• 4:26
  and a half billion dollars.
• 4:28
  It takes Bitmine, the largest
• 4:29
  corporate ether holder on the planet
• 4:31
  and the second biggest corporate
• 4:33
  crypto treasury overall behind
• 4:35
  Only Strategies Bitcoin stack.
• 4:37
  The company is now 96%
• 4:39
  of its way to its alchemy of 5%
• 4:42
  target.
• 4:42
  And Chairman Tom Lee says users are
• 4:44
  starting to see Ethereum as money,
• 4:46
  pointing to the launch of Robinhood
• 4:48
  Chain. As evidence of that
• 4:50
  and it's been really interesting to
• 4:51
  see a lot of talk around protocol
• 4:53
  fees generated by Robinhood in the
• 4:55
  last 10 days and
• 4:57
  the likes of Joe Lubin coming out
• 4:58
  saying that this is not necessarily
• 5:00
  a bad thing for Ethereum and
• 5:02
  definitely a theme that we're going
• 5:03
  to be talking about in the coming
• 5:04
  weeks right here on the show.
• 5:06
  Moving on, speaking of strategy the
• 5:07
  world's biggest corporate Bitcoin
• 5:08
  holder didn't buy any Bitcoin last
• 5:10
  week. If you missed the story
• 5:11
  yesterday it sold nearly 467
• 5:14
  million dollars of MSTR common
• 5:16
  stock. And grew its dollar reserve
• 5:18
  to $3 billion, leaving holdings
• 5:20
  unchanged at $843,775.
• 5:25
  Bitcoin analysts at Benchmark
• 5:27
  and TD Cowan say
• 5:29
  that's not weakness, it's discipline
• 5:30
  and early proof the company is
• 5:32
  following through on its new digital
• 5:34
  credit capital framework.
• 5:36
  TD Cowon reiterated its buy rating
• 5:38
  and a $2.60 price
• 5:40
  target.
• 5:40
  The takeaway here is strategy
• 5:41
  strengthening its balance sheet
• 5:43
  ahead of its next move rather than
• 5:45
  just buying on autopilot.
• 5:47
  Then on-chain stock trading is
• 5:48
  quietly eating hyperliquid.
• 5:51
  HIP3, the exchange's permissionless
• 5:53
  framework that lets builders deploy
• 5:54
  their own perpetual markets has
• 5:56
  exploded from roughly 2% of
• 5:58
  total per volume at the start of the
• 6:00
  year to nearly 50%
• 6:02
  now. The category is dominated by
• 6:04
  trade XYZ, which runs
• 6:06
  markets like XYZ 100,
• 6:08
  tracking the NASDAQ 100, plus single
• 6:10
  stock contracts on names like Nvidia
• 6:12
  and Tesla all settled in
• 6:14
  stable coins with no shares
• 6:15
  involved.
• 6:16
  The draw is leveraged without
• 6:17
  expiry, and like options, perps
• 6:19
  don't decay.
• 6:20
  You just pay the funding and retail
• 6:22
  appetite for trading equities on
• 6:24
  chain seems to be growing pretty
• 6:26
  fast. Our final story, gotta catch
• 6:27
  them all on chain.
• 6:29
  Jupiter, the major Solana decks has
• 6:30
  launched Jupiter Gatcha, a
• 6:32
  venue where collectors buy random
• 6:33
  packs containing real graded Pokemon
• 6:35
  and one piece cards.
• 6:37
  Every card is authenticated, sealed
• 6:39
  in a slab and paired with an on
• 6:41
  chain token so it can be traded
• 6:42
  directly on the exchange.
• 6:44
  As Jupiter puts it, these are the
• 6:45
  same cards you chased as a kid, but
• 6:47
  they're now tradeable on chain.
• 6:49
  That product is still in beta, and
• 6:51
  tokenized Pokemon cards aren't
• 6:53
  new, but a platform of Jupiter's
• 6:55
  size entering the space could
• 6:56
  seriously boost tokenized
• 6:58
  collectibles. It'll be interesting
• 6:59
  to see how this one kind of
• 7:00
  plays out, and I know my colleague
• 7:02
  Calvin Sparks has been
• 7:04
  watching the tokenized Pokemon
• 7:06
  cards space quite closely, so if
• 7:08
  you're not already, go and give him
• 7:09
  a follow. I'm sure we'll be talking
• 7:10
  about this one for the rest of the
• 7:12
  week. That's it for the block
• 7:14
  briefing. We're gonna move
• 7:16
  on to our hot seat segment today and
• 7:18
  we're gonna amalgamate market polls
• 7:20
  in hot seat because we've got David
• 7:21
  Packman in the house.
• 7:23
  So Jordan, I don't know if you can
• 7:24
  bring Packman up on stage with me or
• 7:25
  if he's already here.
• 7:29
  All right, perfect.
• 7:31
  Pacman is going to be joining us
• 7:32
  very, very soon.
• 7:33
  If you don't, okay, there we go.
• 7:35
  David's on screen with us now.
• 7:36
  David, thank you so much for joining
• 7:38
  us. First things first, it's been a
• 7:39
  little while since we last caught
• 7:40
  up.
• 7:41
  How are you doing? How's 2026
• 7:42
  treating you? I know it's already
• 7:44
  more than halfway through the year,
• 7:46
  but it's being that long since we
• 7:47
  were in contact.
• 7:48
  Great to see you, Garrett.
• 7:49
  Thanks for having me on.
• 7:50
  It's awesome to be part of your
• 7:52
  newly launched program and you guys
• 7:54
  are doing great. It's an honor to
• 7:56
  follow CZ also,
• 7:58
  fantastic guest getting on your
• 8:00
  part.
• 8:01
  26 has been exciting from a metric
• 8:03
  standpoint. You know, we watch
• 8:05
  crypto activity adoption by
• 8:07
  major financial institutions,
• 8:09
  TradFi, Web3, and
• 8:11
  boy, everything's up and to the
• 8:12
  right. I wouldn't say that near term
• 8:14
  prices have been following that
• 8:16
  though. And so it's easy to kind of,
• 8:18
  conflate those two things and
• 8:20
  get sad about token prices
• 8:22
  just at the same moment when
• 8:24
  crypto adoption skyrocketing.
• 8:26
  So I'm sure we'll get into that.
• 8:28
  Yeah, sure. I mean, just as a
• 8:30
  bit of a scene setter here, can you
• 8:31
  talk a little bit about your role at
• 8:32
  CoinFund? I mean you've been there
• 8:33
  for a few years.
• 8:35
  You've got a big following online
• 8:36
  and the insights that you put
• 8:38
  out into the market are very
• 8:40
  well respected.
• 8:43
  2026 as a whole, a lot of people
• 8:45
  have been talking about this
• 8:45
  obviously being kind of the depths
• 8:47
  of a bear market.
• 8:49
  A lot of the analysts we've spoken
• 8:50
  to have spoken about things
• 8:52
  perhaps bottoming out by
• 8:54
  October.
• 8:56
  But before we get into that, maybe
• 8:57
  you can just talk a little bit about
• 8:58
  your role and CoinFund's general
• 9:00
  thesis on the markets where we stand
• 9:02
  in Midway through 2026.
• 9:04
  Sure. Well, I'm a venture
• 9:06
  capitalist. I've been a VC for
• 9:08
  18 or 19 years now.
• 9:10
  I did about 13 years at Venrock
• 9:12
  as a partner there, focused on early
• 9:14
  stage web to tech,
• 9:16
  consumer and enterprise.
• 9:17
  And then the last 60 years or
• 9:19
  so focused entirely on crypto.
• 9:22
  That's been at Coin Fund.
• 9:23
  Coin Fund is one of the oldest
• 9:25
  crypto native investing firms in
• 9:27
  the world.
• 9:28
  So we're started in 2015
• 9:30
  and we operate two strategies.
• 9:31
  We have a seed stage fund.
• 9:33
  And that focuses on early two
• 9:36
  to $5 million checks at
• 9:38
  seed stage crypto companies.
• 9:39
  And we have a more traditional
• 9:40
  series A venture fund which
• 9:42
  writes like eight to $15 million
• 9:44
  checks into series A stage
• 9:46
  venture, early
• 9:48
  venture bets in crypto companies,
• 9:50
  we invest all along the crypto
• 9:52
  waterfront from a
• 9:54
  technology perspective, from an
• 9:56
  application perspective, up and down
• 9:57
  the stack.
• 9:58
  Um, we've been fortunate enough
• 10:00
  to back more than a hundred grade,
• 10:02
  uh, founders over the
• 10:04
  decade or 11 years or so.
• 10:06
  And, uh we're active in the market.
• 10:08
  We're, we're about 18 or 20 people.
• 10:10
  And, you know, we love to work with
• 10:12
  builders who are trying to change
• 10:14
  the financial system and
• 10:15
  build decentralized networks.
• 10:17
  Um, you may be more broadly kind
• 10:20
  of our view about what's happening.
• 10:22
  It's been a, it's been an
• 10:25
  amazing 10 or 11 years if
• 10:27
  you go back early.
• 10:28
  To what blockchains
• 10:30
  and Bitcoin were originally promised
• 10:32
  to do. We're building these
• 10:33
  decentralized networks that process
• 10:35
  transactions. And that is a really
• 10:37
  threatening idea to
• 10:39
  centralized systems.
• 10:41
  Centralized systems require trust
• 10:42
  and middlemen and somebody to
• 10:45
  operate a network.
• 10:46
  And the most obvious places where
• 10:48
  those systems remain closed
• 10:50
  are in payments and financial
• 10:52
  markets.
• 10:53
  And I think, no surprise,
• 10:55
  crypto has always represented
• 10:57
  at the very least, a really awesome
• 10:59
  innovation to go and attack
• 11:01
  payments and financial markets
• 11:04
  with open decentralized
• 11:06
  systems. And so here we are, we're
• 11:07
  at 2026 and you have the
• 11:09
  largest banks in the world, the
• 11:11
  largest financial companies in the
• 11:12
  word who have been, you
• 11:14
  know. Really just super
• 11:16
  negative on crypto and blockchains
• 11:18
  for a decade,
• 11:20
  stripping over themselves to
• 11:22
  announce adoption of
• 11:24
  them, the movement of financial
• 11:25
  assets and products on chain as your
• 11:27
  opening segment just discussed.
• 11:29
  We've called this a phenomenon
• 11:31
  convergence and we've predicted it
• 11:33
  for a decades now and it's
• 11:34
  happening. Uh, you're, you, we
• 11:36
  fully.
• 11:37
  Believe there's enough data for you
• 11:39
  to conclude that the majority of
• 11:41
  payments will happen on chain and
• 11:42
  the majority asset trading will
• 11:44
  happen on chain. And what is an
• 11:45
  asset is sort of anything that you
• 11:47
  can make digital from stocks
• 11:49
  and bonds to real estate
• 11:51
  and contracts.
• 11:53
  So the activities up and to the
• 11:55
  right, we'll talk about why that's
• 11:57
  important.
• 11:59
  But it's sort of never been a better
• 12:00
  time to argue that
• 12:03
  crypto represents the future of
• 12:04
  payments in financial markets.
• 12:07
  I took a little kind
• 12:09
  of scroll down your X
• 12:10
  timeline over the last few months.
• 12:12
  And I think one of the big meters
• 12:13
  that you were talking about is like
• 12:16
  stable coins and wake up your money.
• 12:17
  And this is something I've spoken to
• 12:19
  many different people at length.
• 12:20
  I think, one, one of, the
• 12:22
  conversations that stands out to the
• 12:23
  most to me is like off the record
• 12:25
  chats with Stanley Kulikov from Aave
• 12:27
  and just talking about, you
• 12:29
  know, DeFi and, and stable coin
• 12:31
  yields and putting people
• 12:33
  sort of back in control of their
• 12:34
  money. It's interesting to see
• 12:36
  all the banks pushing back on the
• 12:38
  Clarity Act as it
• 12:40
  currently stands and the ability to
• 12:42
  offer yields on stablecoins
• 12:44
  directly to users because
• 12:46
  it really does disintermediate
• 12:49
  the banks.
• 12:50
  What I want to know from you is how
• 12:51
  do you really see this playing out
• 12:53
  and do you think that this does
• 12:55
  become some sort of an existential
• 12:56
  threat for the traditional banking
• 12:58
  system in its current form.
• 13:01
  It's absolutely a threat to deposit
• 13:03
  banking, but let's zoom out
• 13:05
  and discuss why.
• 13:07
  Money can be productive.
• 13:09
  It's always been true that if you
• 13:10
  have some money, you can do
• 13:11
  something with it to make more
• 13:13
  money. You can invest it, sure, but
• 13:15
  you also can loan it to
• 13:16
  other people and those two
• 13:18
  basic mechanisms produce
• 13:20
  return, some with higher risk
• 13:22
  and some with lower risk.
• 13:24
  But most of
• 13:26
  retail banking users,
• 13:28
  when they deposit money in a
• 13:29
  checking or savings account.
• 13:31
  Are deprived of really any
• 13:33
  productivity on that money.
• 13:34
  In Western markets, most
• 13:36
  Western markets.
• 13:38
  Not Brazil, but most Western
• 13:40
  Markets, you earn
• 13:42
  almost nothing by depositing
• 13:43
  your money into an account
• 13:45
  where you feel that it's safe.
• 13:48
  And the people who do
• 13:50
  take that money from you
• 13:52
  and offer it to you as
• 13:54
  a place to make it safe, charge you
• 13:55
  fees for the purpose of doing so,
• 13:57
  and then go and loan that money out
• 13:58
  and make money off you.
• 14:00
  You could be smart and
• 14:02
  put some money in a money market
• 14:03
  fund or maybe try
• 14:05
  to buy some treasury bills, but the
• 14:07
  problem is that there's a smaller
• 14:09
  number of people who are permitted
• 14:10
  to do that. You have to open up
• 14:12
  some type of brokerage or investment
• 14:14
  account in order to do it.
• 14:16
  So banks have been in the business
• 14:18
  of pocketing the yield difference
• 14:19
  there and not really giving you very
• 14:21
  much. I think .04 or .07
• 14:23
  percent is the average rate
• 14:25
  in savings accounts across the
• 14:26
  United States today.
• 14:27
  But That is
• 14:29
  an unfortunate by-product of the
• 14:31
  centralized banking system we have
• 14:33
  today. Crypto offers an
• 14:35
  alternative to that.
• 14:36
  We have a product called a
• 14:37
  stablecoin, which is the same as the
• 14:39
  dollar, but it's digital.
• 14:41
  But that stablecoin can
• 14:43
  offer you the
• 14:45
  treasury yield that it's
• 14:47
  using to establish its reserves and
• 14:49
  at least pay you some or most of
• 14:51
  that yield.
• 14:52
  Now today, that yield's pretty
• 14:53
  meaningful, 3.5%, 4%
• 14:55
  over the last few years.
• 14:56
  Even if you're getting 3%
• 14:58
  of that, that's...
• 14:59
  Pretty much 300 basis points more
• 15:01
  than you're getting today from a
• 15:03
  savings account or a checking
• 15:04
  account.
• 15:05
  That's amazing.
• 15:06
  That 3% difference at
• 15:08
  a restaurant or a retailer is
• 15:10
  often the difference between being
• 15:12
  in the red and in the black.
• 15:14
  Many low margin businesses would
• 15:16
  appreciate having 3% on their
• 15:18
  idle cash flow and every consumer
• 15:20
  should earn 3% or more
• 15:22
  on their Idle balances when
• 15:24
  they're just not spending it.
• 15:25
  But we're just used to a system that
• 15:27
  doesn't offer that. So stable coins
• 15:29
  and crypto do and of course
• 15:31
  that is threatening to the deposit
• 15:33
  banking system.
• 15:33
  It's important to differentiate.
• 15:35
  There are banks that actually are
• 15:36
  not in the deposit world.
• 15:37
  They're more investment banks or
• 15:38
  wealth management banks and they're
• 15:40
  not nearly impacted in this
• 15:42
  way. But the large deposit banks are
• 15:44
  fearful that if you can just earn
• 15:46
  three or four percent just by
• 15:47
  holding a dollar in a wallet,
• 15:49
  well then why would you ever use a
• 15:51
  savings account?
• 15:52
  And and the answer
• 15:54
  is you wouldn't like their
• 15:56
  right to be concerned because the
• 15:58
  product they offer you is worse
• 16:00
  than having a wallet that just
• 16:02
  pays you money for
• 16:04
  keeping your balances there and
• 16:05
  charges you a whole lot less to
• 16:07
  send that money to someone else, a
• 16:09
  consumer or a business.
• 16:10
  Transaction fees are much lower on
• 16:12
  chain. We have built a
• 16:14
  way more
• 16:16
  compelling modern financial system
• 16:18
  and I saw this movie
• 16:20
  in the internet 1.0 days
• 16:22
  where incumbent institutions go
• 16:24
  cry to Congress for laws.
• 16:26
  That prevent them from offering
• 16:28
  a better product to their customers.
• 16:30
  Banks could choose to compete with
• 16:33
  stablecoins by paying yield, but
• 16:35
  that impacts their margins.
• 16:37
  So they're not willing to do it.
• 16:38
  Instead, they want to fight
• 16:40
  stablecoins.
• 16:40
  So I think they're right to be
• 16:42
  concerned.
• 16:43
  But everyone should appreciate that
• 16:44
  what's happening is the banks are
• 16:46
  trying to have a law passed to
• 16:48
  prevent them for building a better
• 16:49
  product to give to your customers.
• 16:52
  I mean, that's the crazy thing to
• 16:53
  me. Do you think it's wrong that
• 16:55
  they can they can lobby this problem
• 16:57
  away, right?
• 16:57
  Like they're actually just using the
• 17:00
  political system to to
• 17:02
  have their cake and eat it as
• 17:03
  opposed to, you know, actually
• 17:05
  playing by the rules of capitalism,
• 17:07
  where the product and service
• 17:09
  dictates who wins and loses.
• 17:11
  Well, unfortunately or fortunately,
• 17:14
  the laws of capitalism include
• 17:15
  regulatory capture and
• 17:17
  what some people in business schools
• 17:19
  call non-market strategies, using
• 17:21
  the courts, using political threats.
• 17:23
  This is all part of capitalism.
• 17:25
  And so they're doing exactly what
• 17:26
  they should be doing, like use
• 17:28
  levers to try to win.
• 17:31
  But in tech, most of
• 17:33
  us have believed for decades that
• 17:34
  the real way to win is by building
• 17:36
  the best product first and not
• 17:38
  trying to use the non- market
• 17:39
  strategies as ways to.
• 17:41
  Promulgate your power or to keep
• 17:43
  bad products in the
• 17:45
  market.
• 17:46
  But I'm pretty optimistic that
• 17:48
  the better product will win here.
• 17:50
  Yes, we're going to see some
• 17:52
  friction imposed by Congress
• 17:54
  on consumers because
• 17:56
  of the bank's concerns, but I think
• 17:58
  that friction will be easily
• 18:00
  evaded or moved past with a few
• 18:02
  clicks. And so I still think
• 18:04
  consumers and businesses are going
• 18:05
  to earn yield on their money.
• 18:06
  In fact, I'm super bullish on yield.
• 18:08
  I really think that the future
• 18:10
  of all of global economics
• 18:13
  will be to deliver yield to
• 18:15
  idle money holders.
• 18:16
  And that is a fundamental change to
• 18:18
  the entire payments and
• 18:20
  market systems around the world.
• 18:21
  Money and assets will become
• 18:23
  productive because of the
• 18:24
  blockchain.
• 18:25
  Yeah, for me, the most interesting
• 18:27
  thing is I held this view very
• 18:29
  strongly for about two and a half
• 18:30
  years and just a personal anecdote
• 18:32
  about a few months ago,
• 18:34
  my wife and I sold some property
• 18:36
  in South Africa and obviously you
• 18:37
  have cash in the bank and if you
• 18:39
  don't put it into a deposit account
• 18:40
  or a savings account, you're not
• 18:41
  earning any interest on your money,
• 18:44
  which anyone would tell you is
• 18:45
  a terrible thing to do.
• 18:47
  You should be earning interest on
• 18:48
  your own money and I was kind of
• 18:49
  thinking I should pile all of this
• 18:51
  money into Aave or some
• 18:53
  DeFi protocol.
• 18:54
  And then, you know, we saw the whole
• 18:56
  CalpDal exploit
• 18:58
  and the amount of money that was
• 19:00
  extracted from that system, just,
• 19:02
  you now, unfortunately that
• 19:04
  was the timing.
• 19:05
  At least I didn't put my money
• 19:07
  there, but I really did think that
• 19:09
  this is tried and tested now and I
• 19:10
  could just go and put my into DeFi.
• 19:13
  Do you think that
• 19:15
  this latest kind of
• 19:18
  trend in 2026 of DeFi
• 19:20
  exploits has put a bit of
• 19:22
  a question mark around?
• 19:24
  The safety of those protocols right
• 19:25
  now or do you think that it's
• 19:27
  not that much to worry about and
• 19:29
  this will very soon be the reality
• 19:31
  in the next year or two where we're
• 19:33
  plugged into systems where you are
• 19:34
  just earning yield on your money,
• 19:36
  whether you know it's happening or
• 19:38
  not, the complexity is abstracted
• 19:40
  away.
• 19:41
  Well, you're talking about risk, and
• 19:43
  I think all of us want to see
• 19:45
  products that offer yield to be
• 19:47
  as low risk as possible.
• 19:49
  I would break the risk up into two
• 19:51
  buckets. One is sort of the exploit
• 19:53
  or security related risks, and
• 19:54
  we'd love those to be zero or near
• 19:56
  zero.
• 19:57
  The way that the banks accomplish
• 19:59
  that zero or near zero sort
• 20:01
  of systemic risk is through federal
• 20:03
  insurance.
• 20:04
  The US government ensures if a bank
• 20:06
  has a failure, that
• 20:08
  your money is still safe.
• 20:09
  And so one question would be, what's
• 20:11
  going to be the?
• 20:11
  The failure prevention mechanism
• 20:13
  for on-chain yield
• 20:15
  products, and is it appropriate
• 20:17
  to have some governmental or
• 20:20
  institutional or insurance type
• 20:22
  products that make you
• 20:24
  feel much better that in the event
• 20:25
  that there's a security incident or
• 20:27
  an exploit, you're safe.
• 20:28
  The second question you're also may
• 20:30
  be asking is, is there some inherent
• 20:32
  risk in the
• 20:34
  economic decision you're making by
• 20:36
  buying the yield-bearing
• 20:38
  product? And that I'm less.
• 20:41
  Focused on you know trying to
• 20:43
  prevent that's an individual
• 20:44
  investment decision and we can talk
• 20:46
  all about today the yield curve and
• 20:48
  how there are people who are moving
• 20:50
  out way farther on the yield curve
• 20:51
  to really face
• 20:54
  a way to improve
• 20:55
  their economic standing in life
• 20:57
  so I would
• 20:59
  separate the two but I agree with
• 21:01
  your observation that there still is
• 21:02
  a bunch of risk in the first bucket
• 21:04
  we can engineer some of that away
• 21:06
  through mature products I
• 21:08
  think the longer these products are
• 21:09
  in market and they're faced with all
• 21:11
  the potential weaknesses being
• 21:13
  exploited and penetrated, the more
• 21:15
  hardened they get.
• 21:17
  But I do think we're probably going
• 21:18
  to need to see some type of
• 21:20
  insurance or federal backstop
• 21:22
  product for those products just like
• 21:23
  we see for traditional banks.
• 21:25
  In exchange for that will come
• 21:27
  regulation and maybe
• 21:29
  capital requirements, just like you
• 21:30
  see with stablecoins and
• 21:33
  the Genius Act. So,
• 21:35
  I'm pretty bullish that you will
• 21:36
  have. Yield yield bearing
• 21:38
  products that feel mostly safe
• 21:40
  other than the other economic risk
• 21:42
  you're taking the black swan risk of
• 21:44
  macro economic effects or things
• 21:46
  like that
• 21:47
  I wanted to get your thoughts on
• 21:49
  tokenization and TradFi,
• 21:52
  the likes of New York Stock
• 21:53
  Exchange, big cryptocurrency
• 21:56
  exchanges like Robinhood
• 21:58
  and a handful of others
• 22:00
  offering tokenized
• 22:02
  securities.
• 22:03
  It seems like that's the inevitable
• 22:05
  future where we can
• 22:07
  just trade these markets 24-7.
• 22:09
  It seems by far superior
• 22:11
  to the existing financial
• 22:13
  system or how conventional
• 22:15
  exchanges work.
• 22:17
  Maybe you've been in this game a lot
• 22:18
  longer than I have and can probably
• 22:20
  offer some interesting insights
• 22:21
  here, but do you think that is a
• 22:23
  better model, actually just having
• 22:24
  things run 24-7,
• 22:26
  or was there an overarching
• 22:28
  reason why these
• 22:31
  trading houses only traded Mondays
• 22:32
  to Fridays and very set trading
• 22:34
  hours? Do you think it's important
• 22:35
  to have those guardrails in place,
• 22:37
  or should it really be 24-
• 22:40
  7 open markets?
• 22:41
  Couple of questions in your question
• 22:43
  there.
• 22:44
  Just going back, I think there's a
• 22:46
  good parallel to help us think
• 22:47
  through your questions.
• 22:50
  When the commercialization
• 22:52
  of the internet began in sort of the
• 22:54
  mid-90s, what we were building
• 22:56
  by building the internet was a
• 22:58
  globally connected, always on,
• 23:00
  low-cost network that
• 23:02
  was permissionless.
• 23:04
  Anyone could join by taking a
• 23:05
  computer and plugging it into the
• 23:07
  internet. You had to pay somebody
• 23:08
  for some bandwidth, but anyone could
• 23:10
  offer that bandwidth.
• 23:11
  And so it was an open system, and
• 23:13
  it was a network that was always on
• 23:15
  24-7.
• 23:16
  And so data could move at much lower
• 23:18
  costs.
• 23:19
  And because it was based on open
• 23:21
  source software, most of its
• 23:23
  weaknesses got
• 23:25
  exploited early on and cleaned up.
• 23:28
  And everyone who was a developer
• 23:30
  could work on it.
• 23:31
  And so you had the collective minds
• 23:32
  of thousands and millions of
• 23:34
  developers working to build a better
• 23:35
  network. And that's what we have.
• 23:36
  The internet is a much better
• 23:37
  network then token ring and all
• 23:39
  the other closed network standards
• 23:41
  that exist before it.
• 23:42
  And now everyone moves the data on
• 23:44
  the open internet.
• 23:45
  People said they wouldn't do it.
• 23:46
  Banks said they would never send
• 23:48
  transactions on the internet.
• 23:50
  Corporations said we're never going
• 23:51
  to let our data be stored in the
• 23:52
  cloud. All of that fell away because
• 23:55
  we've built better networks.
• 23:57
  Well, what are blockchains?
• 23:58
  They're just better networks,
• 24:00
  they're always on, they are
• 24:01
  decentralized, they are
• 24:02
  permissionless, they have a very,
• 24:03
  very low fee, and they're global by
• 24:05
  nature. So what do they go
• 24:07
  if the thing that's different about
• 24:08
  them than the internet is that the
• 24:09
  really great for economic
• 24:11
  transactions You know, they have a
• 24:13
  a great cryptographic protection
• 24:15
  mechanism to validate transactions
• 24:17
  So no trust is needed for those
• 24:19
  transactions where it matters So now
• 24:21
  we have these better networks and
• 24:22
  the question you're asking is is it
• 24:23
  time to move assets onto them and
• 24:25
  trade them? The answer is of course
• 24:26
  it is and that's exactly what's
• 24:28
  happening. And and yes, uh
• 24:30
  All assets are going to trade 24
• 24:32
  seven. They're going to treat it
• 24:33
  much lower fees.
• 24:34
  Those fees are going be transparent
• 24:36
  It'll be much harder for
• 24:38
  middlemen to obfuscate Transaction
• 24:41
  fees which is what happens today
• 24:42
  when you buy a stock and someone's
• 24:43
  selling your order flow to someone
• 24:45
  else who's front-running you That
• 24:46
  can still happen on blockchains and
• 24:48
  it is But
• 24:50
  but you can watch it.
• 24:51
  You can watch. It happen because the
• 24:52
  transactions are Transparent.
• 24:54
  So yes, I think all I think you're
• 24:56
  seeing this now with equities moving
• 24:58
  on chain and we're
• 25:00
  going to get lower fees and always
• 25:01
  on and Yes, so
• 25:03
  is it good for asset prices
• 25:05
  that they trade 24-7?
• 25:07
  I think it's inconvenient for some
• 25:08
  traders who need to sleep
• 25:10
  But that's why we have a lot of
• 25:12
  automated trading But
• 25:14
  yeah, I have no
• 25:17
  Fondness for you know market hours
• 25:19
  and the same The same
• 25:21
  can be said about sending a bank
• 25:23
  wire, which can only happen during
• 25:24
  banking hours, or sending an
• 25:26
  overseas payment, which takes
• 25:28
  multiple days across correspondent
• 25:29
  banks, because people work at
• 25:31
  banks need to sleep.
• 25:32
  So all these closed systems are
• 25:34
  going to go away. It's not
• 25:35
  overnight. We're going to see many
• 25:37
  of these systems stay for a long
• 25:39
  time, but volumes are transferring
• 25:41
  over, and we will hit a
• 25:43
  tipping point where all financial
• 25:45
  volumes will move on from
• 25:46
  blockchains.
• 25:48
  You recently tweeted about
• 25:50
  prediction markets and young
• 25:51
  investors, and you are arguing
• 25:53
  there, and I stand to be corrected
• 25:55
  if I've taken this out of context,
• 25:57
  but younger people
• 25:59
  are moving on to prediction markets
• 26:01
  in crypto because they're doing
• 26:03
  the math and they can just make more
• 26:05
  money doing this than they would in
• 26:07
  TradFi or on gambling
• 26:09
  platforms. What's your general
• 26:10
  thesis on prediction markets?
• 26:12
  Do you think it's a great
• 26:13
  innovation?
• 26:15
  And yeah, just the younger
• 26:16
  generation of people being a bit
• 26:17
  more sort of financially savvy
• 26:19
  and diving headfirst into
• 26:21
  the wonderful world of crypto.
• 26:23
  Thanks for asking.
• 26:24
  I think this thesis is really based
• 26:26
  on observation, so it's not like
• 26:28
  some point of view we've concocted
• 26:30
  in a dark room.
• 26:31
  The data shows that younger
• 26:33
  generations, Millennials and Gen
• 26:35
  Z, are taking more economic
• 26:37
  risk with their idle
• 26:39
  balances.
• 26:40
  And they're doing it largely
• 26:42
  because, and you ask them these
• 26:43
  questions and the research shows,
• 26:46
  because the traditional forms of
• 26:48
  wealth building are really closed
• 26:49
  off to them. The number one form
• 26:51
  of wealth building too.
• 26:53
  To the boomers was was
• 26:55
  home buying a house and
• 26:57
  home prices Relative
• 26:58
  to average incomes are way
• 27:01
  up It's about a seven and a half
• 27:02
  times your average annual salary
• 27:04
  today to buy a house
• 27:06
  When that was about three to three
• 27:07
  and a halve two generations ago
• 27:09
  So, you know just it feels to
• 27:11
  many younger people that
• 27:12
  homeownership is impossible or out
• 27:14
  of reach And so that as a mechanism
• 27:16
  for wealth creation is you know
• 27:17
  closed off to them and you know Just
• 27:19
  buying some equities and holding
• 27:21
  them also doesn't feel great,
• 27:23
  especially in a high inflation
• 27:24
  environment that we've been in for a
• 27:26
  while. So what do you do?
• 27:27
  Well, it's natural to want to
• 27:29
  take some more asymmetric risk
• 27:31
  or some high variance bets.
• 27:33
  And it just so happens that that's
• 27:35
  coinciding with all these mobile
• 27:37
  products that are always on that
• 27:39
  let you trade fractions of stocks,
• 27:41
  buy options,
• 27:43
  gamble, play
• 27:45
  real money gaming,
• 27:49
  play on hyper liquid with leverage.
• 27:52
  Or participate in prediction markets
• 27:54
  or even zero data expiry
• 27:56
  options, these are all
• 27:58
  high risk. In some ways, they're
• 28:00
  entertaining.
• 28:00
  They're shorter-term investments.
• 28:03
  Some of them give you an answer in
• 28:04
  five minutes like a crash game.
• 28:07
  Some give you an answer in a day
• 28:08
  or a week, and those are
• 28:10
  more interesting to many young
• 28:12
  people than buying the
• 28:14
  S&P 500 and holding it for 30 years.
• 28:17
  So we're seeing significant increase
• 28:19
  in what most people would call
• 28:21
  bedding. I think I can
• 28:23
  argue that stock picking,
• 28:24
  especially when the average duration
• 28:27
  of an equity held today is five
• 28:29
  and a half months, is also a form
• 28:31
  of wagering.
• 28:32
  Yes, you can point to cash flows and
• 28:34
  other more sensible fundamentals
• 28:37
  behind stock prices
• 28:39
  than maybe behind crypto
• 28:41
  asset prices.
• 28:42
  But younger people are taking more
• 28:43
  economic risk.
• 28:44
  And there's many more of these
• 28:46
  products.
• 28:46
  So, um, We need to
• 28:49
  face that fact.
• 28:50
  My argument is that that's probably
• 28:51
  irreversible and that it would be
• 28:53
  better to focus on, rather than sort
• 28:55
  of poo-pooing the whole thing and
• 28:56
  saying these kids don't know what
• 28:57
  they're doing, they know exactly
• 28:59
  what they are doing, but we should
• 29:00
  make sure we're building transparent
• 29:02
  products, hopefully ones on chain,
• 29:04
  where the payout ability of
• 29:06
  the site you're using is
• 29:08
  clear and the
• 29:10
  transactions are clear and the fees
• 29:11
  are clear. And so we can just have
• 29:13
  like kind of better products than
• 29:14
  sort off-shore dark bedding parlors.
• 29:17
  Yeah, for sure.
• 29:18
  I completely agree with you on
• 29:20
  the status quo of
• 29:22
  what we have right now and the
• 29:23
  alternative that is prediction
• 29:25
  marks. It's been fascinating to just
• 29:26
  see the evolution.
• 29:29
  I wanted to chat to you a little bit
• 29:30
  about probably your thesis on
• 29:32
  AI and I guess just investing in
• 29:34
  general. Obviously, you're sort of
• 29:35
  head of venture investing and I
• 29:37
  would love to know how
• 29:40
  your approach to this has changed in
• 29:41
  the last.
• 29:43
  Year and a half and I would just
• 29:44
  even say in a year because if I
• 29:46
  think back to what I was able to do
• 29:47
  a year ago versus what I can
• 29:49
  do now with the tools I have
• 29:51
  at my disposal, it's
• 29:53
  actually unreal the
• 29:56
  pace of innovation that's happened
• 29:57
  in 12 months.
• 30:00
  What are you looking at investing
• 30:01
  in? How are you judging this?
• 30:03
  How much have your theses
• 30:05
  changed as these tools have
• 30:07
  developed?
• 30:08
  Two main observations, maybe.
• 30:10
  First, it's unbelievably exciting.
• 30:12
  This is just an extraordinary
• 30:14
  set of innovations.
• 30:16
  And I've been in tech 30 years and
• 30:18
  I've seen a lot of different waves
• 30:20
  go by us.
• 30:22
  The internet in the mid-90s and
• 30:24
  mobile devices in the early 2000s,
• 30:26
  social media, cloud computing,
• 30:29
  crypto, and certainly now
• 30:31
  AI is the wave.
• 30:33
  And it's incredibly fun also,
• 30:36
  like, I mean, that's one of the
• 30:37
  things that makes technology so
• 30:39
  adopting is that it's fun.
• 30:41
  Mobile phones are fun, social media
• 30:42
  is fun, crypto is fun and
• 30:44
  AI is really fun.
• 30:46
  So like, this trend has not
• 30:48
  gone away and it's accelerating.
• 30:50
  I would agree with your
• 30:51
  characterization that for me,
• 30:53
  the real big tipping point was sort
• 30:55
  of last November.
• 30:57
  We've had machine learning and,
• 30:58
  you know, early LLMs for some time,
• 31:00
  but once...
• 31:01
  Cloud Code really shipped
• 31:03
  in Q4 and
• 31:06
  the quality level of the
• 31:08
  products you could build with
• 31:09
  automated AI-driven code
• 31:11
  creation reached
• 31:13
  what it did.
• 31:14
  Then it was really clear, that was
• 31:16
  the light bulb moment to so many
• 31:17
  people that you now can just build.
• 31:20
  It doesn't matter your skill set,
• 31:22
  obviously it takes skills to be
• 31:23
  really good at automated product
• 31:25
  development, but But still,
• 31:27
  it's a lot less than going to
• 31:29
  school for four years to learn
• 31:30
  computer science like I did and then
• 31:32
  learning how to write code for
• 31:33
  another 10 years.
• 31:34
  So it's incredible.
• 31:35
  What does this mean for investing?
• 31:37
  Well, first, the most obvious thing
• 31:39
  is just anything that's automatable
• 31:40
  will now be automated.
• 31:42
  And so processes, low-level tasks,
• 31:45
  research tasks, things
• 31:47
  that the LLM is particularly good
• 31:49
  at, information gathering,
• 31:51
  processing of large amounts of data,
• 31:53
  finding insights and large amounts
• 31:55
  of data.
• 31:56
  This has all happened already
• 31:59
  and if you're not taking advantage
• 32:00
  of it, you're far behind.
• 32:02
  So that's the first obvious place is
• 32:04
  like let's just move faster as an
• 32:06
  organization because we can automate
• 32:07
  the heck out of anything that's part
• 32:09
  of a business process or an insight
• 32:10
  process.
• 32:12
  One question that we ask ourselves
• 32:14
  is when is the AI going to
• 32:16
  be as good or better then
• 32:18
  we, venture capitalists, are
• 32:19
  making an investment decision.
• 32:21
  And so much of an early-stage
• 32:22
  investment decision is tied up in
• 32:24
  who is the person or people who
• 32:26
  are building the product and doing a
• 32:27
  team evaluation.
• 32:29
  We've spent a bunch of time trying
• 32:30
  to train the LLMs on our
• 32:32
  own observations of people, and
• 32:34
  I think it's really helpful in
• 32:35
  pointing things out, but I don't
• 32:37
  know that I fully trust its judgment
• 32:39
  versus a
• 32:41
  great venture capitalist, but
• 32:43
  clearly investment decisions are
• 32:45
  going to be severely impacted by
• 32:48
  LLM's. And so I think VCs and
• 32:50
  investors in general have something
• 32:51
  to gain.
• 32:52
  But maybe what we're most excited
• 32:53
  about is just the fact that you can
• 32:55
  now have agents do lots of
• 32:57
  transactions for you that are
• 32:58
  financial in nature.
• 32:59
  And whether that's stock trading or
• 33:02
  moving of money,
• 33:04
  that implicates blockchains in a
• 33:06
  really big power fight.
• 33:09
  Blockchains are much better for
• 33:11
  programming software-based
• 33:13
  transactions on than traditional
• 33:15
  networks like credit cards.
• 33:17
  Credit card networks are really bad
• 33:19
  for agentic transactions because
• 33:21
  they cost $0.25 plus 3.25%
• 33:24
  in order to do a single transaction.
• 33:26
  As you know, most API calls
• 33:28
  and agentic transaction are
• 33:30
  pennies or fractions of a penny
• 33:32
  in nature, and so you need a
• 33:34
  small denominated network that
• 33:36
  can be really low-fee.
• 33:38
  You need something that's always on.
• 33:39
  It doesn't like reconcile in T plus
• 33:41
  one or T plus two days.
• 33:42
  You need to reconcile immediately,
• 33:44
  and of course you need them
• 33:44
  programmable.
• 33:45
  So there's a pretty strong argument
• 33:46
  that agentic-based transactions
• 33:49
  can and should happen using
• 33:50
  stablecoins on top of blockchains,
• 33:52
  and we've made a lot of investments
• 33:53
  around that thesis.
• 33:54
  I'm not declaring that it's true
• 33:56
  yet, and the data is really early,
• 33:58
  and to be honest, you know,
• 34:00
  stablecoin-based agentic transaction
• 34:02
  volume is still really small
• 34:03
  compared to global transaction
• 34:05
  volume, but it's a thesis and it's
• 34:06
  interesting.
• 34:07
  The other place where AI intersects
• 34:09
  crypto that we've been pursuing, I
• 34:11
  think for at least four years now,
• 34:13
  is the notion that it
• 34:16
  seems kind of
• 34:18
  sucky to live in a world where
• 34:20
  three or four companies control
• 34:22
  all of the
• 34:24
  future of human thought because
• 34:26
  they're the foundational model
• 34:28
  companies. It seems
• 34:30
  like some type of decentralized
• 34:32
  alternative for at least trading
• 34:34
  your own.
• 34:35
  LLMs or AI models should
• 34:37
  exist.
• 34:38
  But there's been just this, you
• 34:40
  know, bland grab of
• 34:42
  GPUs that kind of prevents anyone
• 34:44
  else from getting in on the training
• 34:45
  side of things at scale, which means
• 34:47
  you're sort of forced to pay the
• 34:49
  piper, those three or four
• 34:50
  companies, in order to use
• 34:52
  LLM.
• 34:53
  So there is this question of can we
• 34:55
  build decentralized training and
• 34:57
  maybe even decentralized inference
• 34:58
  networks to bring the cost down and
• 35:00
  the accessibility of model building
• 35:02
  and model utilization.
• 35:04
  And again, I would say that that's
• 35:05
  not definitively established, but
• 35:07
  it's an investment thesis that we've
• 35:09
  pursued for some time.
• 35:10
  Yeah, I mean, I've looked, you're
• 35:12
  sort of intimating that, you know,
• 35:14
  the deep end space where, you know
• 35:15
  people are building these networks
• 35:17
  that can connect an
• 35:19
  individual user to a
• 35:21
  network of people
• 35:23
  making available their GPU
• 35:25
  processing power, whether that's in
• 35:27
  data centers or just, you know, a
• 35:28
  few, whatever it is, 50
• 35:30
  ATs, it's someone's house,
• 35:32
  you know, which I think, which I
• 35:34
  think is very interesting.
• 35:35
  And you know the few projects
• 35:37
  that I looked into over the last
• 35:38
  year or two for reporting on—
• 35:41
  that were running on Solana,
• 35:43
  et cetera.
• 35:44
  They needed a high throughput
• 35:46
  blockchain like
• 35:48
  that in order to manage
• 35:50
  the inferences and the actual
• 35:51
  microtransactions that are needed to
• 35:53
  pay for who's
• 35:56
  made what available and then paying
• 35:57
  them for that.
• 35:59
  And my question on that is,
• 36:00
  is there actually enough, are there
• 36:02
  enough people and enough small
• 36:03
  devices that could compete with,
• 36:05
  like you say, the Anthropics?
• 36:07
  Of the world that just have huge
• 36:09
  amounts of capital and massive data
• 36:10
  centers full of institutional
• 36:14
  grade graphics cards.
• 36:16
  I don't know, I don't think so,
• 36:19
  but we definitely want a
• 36:20
  decentralized alternative to
• 36:22
  be explored and can we
• 36:24
  get there?
• 36:25
  Yeah, well, so we have
• 36:27
  we have done that math and by
• 36:29
  far there is
• 36:31
  Super sufficient amount
• 36:33
  of gpu capacity sitting
• 36:35
  on people's desks On
• 36:37
  their laps of their laptops and in
• 36:39
  their pocket in their phones That if
• 36:41
  you sum together a bunch
• 36:43
  of that idle capacity you could
• 36:44
  train really large
• 36:46
  ai models The issue is
• 36:48
  that the latency across those
• 36:50
  devices is way different
• 36:52
  and suboptimal compared to the way
• 36:54
  training systems work today when
• 36:56
  all of the GPUs are centralized and
• 36:58
  separated by amazingly high
• 37:00
  throughput connections.
• 37:02
  And so we have companies investments
• 37:04
  we've made in the software
• 37:06
  and technology necessary to
• 37:08
  operate training and inference
• 37:10
  across highly distributed
• 37:12
  networks where latency is
• 37:14
  high. So if that problem is
• 37:16
  solved, yes, there is enough,
• 37:18
  by far, enough GPU
• 37:19
  capacity.
• 37:20
  The question is, can you convince
• 37:21
  people to put their idle cards
• 37:23
  or computers or phones
• 37:25
  in a network?
• 37:26
  And if there's economic reason to do
• 37:28
  so, well, maybe that can happen.
• 37:30
  But you know, Filecoin and the early
• 37:32
  storage models are interesting
• 37:33
  examples where
• 37:35
  I think many of us, when we
• 37:37
  first heard about decentralized
• 37:38
  storage, assumed that it was going
• 37:39
  to be like our idle hard drives that
• 37:42
  were going to power decentralized
• 37:43
  storage networks.
• 37:44
  But it quickly turned out that, no,
• 37:45
  no. It's an industrial product.
• 37:46
  Like, it's an enterprise product
• 37:48
  that data centers and
• 37:50
  even hard drive manufacturers
• 37:52
  will put supply
• 37:54
  into the network. So we'll see what
• 37:55
  happens with decentralized training
• 37:57
  and inference.
• 37:58
  But I think we have plenty of GPU
• 38:00
  power in the world, also making
• 38:02
  sure that training can happen across
• 38:03
  heterogeneous networks where you
• 38:05
  cannot assume that every GPU is the
• 38:07
  same.
• 38:08
  So early days here, but there's
• 38:10
  a lot of demand for much lower cost
• 38:12
  training.
• 38:13
  And if you could bring down not just
• 38:15
  the cost of inference, which
• 38:16
  everyone talks about because no one
• 38:17
  wants to pay for tokens, but also
• 38:19
  bring down cost of training, well,
• 38:20
  then you can have a lot of other
• 38:21
  people training specialty models.
• 38:23
  I wonder also just if
• 38:25
  the approach to
• 38:27
  the tokenomics of all these projects
• 38:29
  has been perhaps one of the failings
• 38:31
  because at the end of the day,
• 38:32
  people are getting paid in a native
• 38:34
  token of a protocol that's building
• 38:35
  this kind of infrastructure and then
• 38:37
  you're encouraged to stake the
• 38:39
  tokens and that kind of things and
• 38:41
  maybe we just need to pay people in
• 38:42
  a stablecoin for availing
• 38:44
  their hardware to make this happen.
• 38:46
  And once you tick
• 38:48
  all the boxes and the tech side of
• 38:49
  it is sorted.
• 38:50
  The incentive mechanisms just need
• 38:52
  to be very straightforward, and you
• 38:54
  think of networks like Bitcoin where
• 38:56
  the incentive mechanism is very
• 38:58
  clear and very straightforward.
• 39:00
  But the adoption of Bitcoin,
• 39:02
  the actual token, the money,
• 39:04
  as it were, has also been really,
• 39:06
  really high and the value
• 39:07
  proposition is understood.
• 39:09
  And probably a lot of these other
• 39:10
  new deepened networks just don't
• 39:12
  have that. Their tokenomics don't
• 39:13
  make sense.
• 39:14
  People don't trust that the token
• 39:15
  will have long-term value.
• 39:18
  It's hard to compete with what...
• 39:19
  What Bitcoin is and maybe that's a
• 39:21
  missing piece of the puzzle as well.
• 39:23
  You're bringing up kind of the most
• 39:25
  fundamental question about
• 39:27
  tokens in general, not just
• 39:29
  for deep-in networks, but it's
• 39:31
  this battle between economic
• 39:34
  incentive in
• 39:36
  the task you're doing to help build
• 39:38
  out a network and short-termism
• 39:40
  versus long-term-ism view on
• 39:42
  when you want that return paid to
• 39:43
  you. So if you're taking a
• 39:45
  longer-term perspective
• 39:47
  and you're helping build a network,
• 39:49
  and you think the future value of
• 39:51
  the network will be very high
• 39:52
  relative to today's current price,
• 39:54
  well then you'd love to be paid in
• 39:55
  some native token.
• 39:57
  But if you really are more
• 39:59
  transactional in nature, and we just
• 40:00
  talked earlier about how we're in
• 40:02
  this world where younger people
• 40:03
  really want their bets to resolve
• 40:05
  much more quickly and take their
• 40:07
  winnings or losings and then move
• 40:09
  on, well then maybe you should just
• 40:10
  get paid in a stable coin.
• 40:11
  But when most of these deep in
• 40:13
  networks, and even, look, I was an
• 40:14
  Ethereum miner for many years, I was
• 40:16
  paid in ETH. I'm glad I was pay in
• 40:17
  Eth. I made great.
• 40:20
  Incremental returns because they
• 40:21
  held my ETH, but it
• 40:23
  might not have gone that way.
• 40:24
  There were plenty of other tokens
• 40:25
  that I was mining also and those
• 40:27
  we don't even remember their names.
• 40:29
  Right. So I'm with you that
• 40:31
  like this is a fundamental question,
• 40:33
  but we didn't have stable coins back
• 40:35
  when I was finding ETH.
• 40:36
  And so I didn't a choice, but why
• 40:38
  couldn't we offer a choice?
• 40:39
  Or I could see your.
• 40:41
  You know, your wallet for your
• 40:43
  GPU payments, having an
• 40:45
  auto convert to USDC in it,
• 40:47
  right? As my payments come in.
• 40:48
  So I think this is now resolvable.
• 40:50
  You are bringing up the
• 40:51
  uncomfortable question though
• 40:54
  around crypto tokens in general,
• 40:56
  just so many are so wrapped for
• 40:58
  so long that, you
• 41:00
  know, are they, do they
• 41:02
  have some updraft that's gonna power
• 41:04
  their long-term returns?
• 41:06
  Or are they just these sort of
• 41:07
  short-term market perturbations
• 41:09
  that are based on narrative
• 41:11
  ever.
• 41:12
  Twitter activity, and boy, if it's
• 41:14
  that, then they're not interesting
• 41:15
  as products.
• 41:16
  But if they really can resort
• 41:18
  back to some tie-in to
• 41:20
  fundamental value of the network
• 41:22
  or whatever product they're
• 41:23
  building, well, then I think it
• 41:24
  could be interesting.
• 41:25
  And I think we've been waiting for
• 41:27
  some regulatory clarity to be able
• 41:29
  to do that.
• 41:30
  We're investors in some really great
• 41:32
  decentralized networks.
• 41:33
  You know, EtherFi is an
• 41:35
  amazing protocol and product.
• 41:37
  That consumers love, they have
• 41:38
  really high AOM, they pay great
• 41:40
  yield to their users, they offer
• 41:41
  awesome payment mechanisms through
• 41:43
  card integration, great products,
• 41:45
  people love them.
• 41:46
  There's a native token that governs
• 41:48
  EtherFi. If you follow this company,
• 41:50
  you would probably like to hold
• 41:52
  some piece of their future value.
• 41:54
  But you got to believe that there's
• 41:55
  a relationship between that token
• 41:57
  and their business.
• 41:59
  And we've been prevented by
• 42:01
  SEC and Aggressive Government Action
• 42:03
  from linking those two things.
• 42:05
  But that's what hopefully this
• 42:06
  forthcoming bill sells.
• 42:09
  Yeah, beautiful.
• 42:10
  I love that you tied that in on the
• 42:11
  end there, because it's
• 42:13
  way more complicated than just does
• 42:15
  this network have long-term
• 42:17
  value.
• 42:18
  David, thank you so much for taking
• 42:19
  the time to chat to us today.
• 42:20
  I'm sure we could probably speak for
• 42:21
  about an hour, but we try to keep
• 42:23
  the show as formatted as possible
• 42:24
  and not run too long.
• 42:26
  We'll definitely have to have you
• 42:28
  back for a
• 42:30
  second conversation in the next few
• 42:31
  weeks and months.
• 42:33
  But again, thanks so much for taking
• 42:34
  your time.
• 42:34
  Thank you for having me, Gareth.
• 42:36
  Awesome to be on your show, and I
• 42:37
  hope it continues to do great.
• 42:39
  Fantastic, thank you very much.
• 42:40
  David Pakman, everyone, if you're
• 42:42
  not already, you can go and follow
• 42:43
  him on X at Pakman.
• 42:45
  That's P-A-K-M-A N.
• 42:47
  If you're new here to the show,
• 42:48
  welcome to The Starting Block.
• 42:49
  This is The Block's new daily show.
• 42:51
  We broadcast 830 a.m.
• 42:53
  Eastern Time if you are in the U.S.,
• 42:56
  230 Central European Time if your in
• 42:58
  Europe. Please set a reminder on
• 42:59
  your calendar every day to come and
• 43:00
  join us. I promise you'll be
• 43:02
  entertained. Tomorrow we've got a
• 43:03
  great show lined up and we've show
• 43:05
  the rest of the week.
• 43:06
  And if you'd like to catch up on
• 43:08
  some of the content that we've been
• 43:09
  having here on the show, you can go
• 43:11
  and check out the BlockX account or
• 43:12
  you can come to our YouTube channel.
• 43:14
  Please subscribe to the Block on
• 43:15
  YouTube. And of course, I encourage
• 43:17
  you to go back and watch our launch
• 43:19
  episode, our interview with
• 43:21
  Finance Founder CZ.
• 43:23
  Promise you'll be entertained.
• 43:25
  Hope to see you back here tomorrow.
• 43:26
  Wherever you are, have a fantastic
• 43:27
  morning, evening, and we
• 43:29
  will catch you next time.