Latest / Connecting the Dots / Spatial AI Meets Wall Street Access

Transcript

Spatial AI Meets Wall Street Access · Connecting the Dots

• 0:00
  Today's tech news moves fast, but understanding
• 0:03
  it shouldn't be hard. That's why Snarfle Solutions
• 0:06
  Group brings you Connecting the Dots, where we
• 0:09
  break down what's happening in tech and why it
• 0:11
  matters to you. Now, let's dive in with your
• 0:14
  hosts. Good morning, everyone. You are listening
• 0:17
  to Connecting the Dots. I'm Alex. And I'm Morgan.
• 0:20
  We are coming to you live from the Snarfle Solutions
• 0:23
  Group studios right here in... very windy Sacramento,
• 0:28
  California. You are not kidding about the wind.
• 0:31
  It is Wednesday, February 18th, 2026. And if
• 0:35
  you're local, you know exactly what we're talking
• 0:37
  about. Oh, yeah. The windows have been rattling
• 0:39
  since about 4 0 a .m. Yeah, we have a pretty
• 0:42
  serious wind advisory in effect until what? Ten
• 0:45
  point zero zero p .m. on Thursday. So if you
• 0:48
  haven't tied down the patio furniture or that
• 0:50
  one recycling bin that always tips over that
• 0:52
  one bin, consider this your official warning
• 0:54
  and looking at the national map. is not much
• 0:56
  calmer elsewhere. We are seeing critical fire
• 0:59
  weather developing in the high plains and parts
• 1:01
  of the Midwest. Right. It's that dangerous combination
• 1:04
  of high winds and really low humidity. At the
• 1:06
  same time, the West is just getting hammered
• 1:09
  by two different Pacific storms bringing heavy
• 1:11
  snow. And then you've got a wintry mix stretching
• 1:13
  all the way from the northern plains to New England.
• 1:16
  So it is a very active map today. Stay safe out
• 1:19
  there. Yeah. And while the weather is volatile,
• 1:22
  The market seemed to be taking a little nap this
• 1:26
  morning. A well -deserved nap, maybe. The Dow
• 1:29
  is up just a hair, about 37 points, closing at
• 1:32
  49 ,609. And the S &P 500 is basically flat,
• 1:37
  sitting at 6 ,857. Yeah, not a lot of movement.
• 1:40
  And for the crypto crowd, Bitcoin is down a little
• 1:43
  bit, just under 1%, so it's trading around $67
• 1:45
  ,059. So, quite on the financial front, but...
• 1:49
  Loud on the atmospheric front I will take that
• 1:52
  trade off any day me too absolutely now before
• 1:55
  we get into the morning star we have to talk
• 1:57
  about what is feeling the show you see us with
• 2:00
  these eggs every single morning we can't function
• 2:02
  without it we really can't we are talking about
• 2:04
  our sponsor strong coffee. Look, I love caffeine,
• 2:08
  but I hate that jagged nervous energy you get
• 2:10
  from cheap coffee or energy drinks. Oh, the worst.
• 2:13
  Where you feel like you're going to vibrate through
• 2:15
  your chair. Yes. And the crash that comes after
• 2:17
  is just brutal. But strong coffee is different
• 2:20
  because it's instant, which I know that usually
• 2:22
  sounds terrible. But this stuff is premium. It
• 2:25
  tastes like a traditional brew, but it's functional.
• 2:27
  It's the ingredients. It's packed with 15 grams
• 2:30
  of protein. And MCTs, plus adaptogens like Ashwagandha.
• 2:34
  It's all designed to give you clean, sustained
• 2:36
  energy. No jitters, just focus. Exactly. You
• 2:40
  can head over to strongcoffeecompany .com and
• 2:43
  use our promo code SNARFL. That's S -N -A -R
• 2:46
  -F -U -L for 20 % off your entire purchase. It
• 2:49
  is a game changer for the morning routine. It
• 2:52
  really is. Alright, let's open to the morning
• 2:53
  star. That's the Snarfletec article report for
• 2:56
  anyone new. We have three stories today that
• 2:59
  really paint a picture of how barriers are breaking
• 3:01
  down. Yeah, both in the digital world and in
• 3:04
  the financial world. And we're starting with
• 3:07
  a massive check being written in the world of
• 3:09
  artificial intelligence. Let's jump right in.
• 3:11
  The headline here is really the dollar amount.
• 3:13
  Autodesk. giant behind, I mean basically every
• 3:17
  building and engine designed in the last 30 years.
• 3:19
  Yeah for sure. They've invested 200 million dollars.
• 3:22
  into a company called World Labs. And just to
• 3:24
  set the context, World Labs is the spatial AI
• 3:28
  startup founded by Fei -Fei Li. She's often called
• 3:31
  the godmother of AI. Right. They came out of
• 3:33
  stealth mode in 2024 with a, what, a one billion
• 3:37
  dollar valuation. And now reports are saying
• 3:39
  they're in talks for a five billion dollar valuation.
• 3:42
  OK, so obviously the market thinks they're valuable.
• 3:44
  But I want to push back on this a little bit.
• 3:46
  We hear about big AI investments every single
• 3:49
  day. Autodesk is a serious engineering company.
• 3:53
  They deal in physics, load -bearing walls, precise
• 3:55
  measurements. Why are they throwing money at
• 3:58
  a generative AI company? Isn't generative AI
• 4:01
  just about making cool pictures? That is the
• 4:04
  exact problem they're trying to solve. You fit
• 4:06
  the nail in the head. Right now, there's a massive
• 4:08
  disconnect between generative AI and engineering.
• 4:13
  How do you mean? Well, think about something
• 4:15
  like mid -journey or, you know, the early image
• 4:18
  generators. You could type in, show me a futuristic
• 4:21
  car, and it would give you this beautiful image.
• 4:24
  It looks like a car. But if you zoomed in, the
• 4:26
  wheels might be ovals. There's no engine inside.
• 4:28
  If you sent that image to a factory, they would
• 4:30
  just laugh at you. Because it's just pixels.
• 4:33
  It's a hallucination of a car. Yeah, exactly.
• 4:35
  It has no structural integrity. You can't build
• 4:37
  from it. Autodesk deals in CAD computer aided
• 4:40
  design. That's all math. It's geometry. Pure
• 4:43
  logic. And Darren Green, the chief scientist
• 4:45
  at Autodesk, is talking about merging these two
• 4:48
  worlds into something he calls neural CAD. Neural
• 4:51
  CAD. Okay, that sounds like a buzzword, but break
• 4:53
  it down. What does that actually mean for, say,
• 4:56
  an architect? It means bridging the gap between
• 4:58
  the dreaming phase and the building phase. So
• 5:02
  imagine you're an architect. Right now you might
• 5:04
  sketch an idea or use an AI to generate a vibe
• 5:08
  for a building, but then you have to go into
• 5:10
  AutoCAD and manually draw Every single beam,
• 5:14
  every window, every vent from scratch. It's two
• 5:17
  completely different workflows. Totally. With
• 5:19
  NeuralCAD, the idea is you use WorldLab's model
• 5:22
  to generate the 3D environment, the office layout,
• 5:25
  the aesthetic. But because of this integration,
• 5:27
  that model isn't just like dummy pixels. It's
• 5:30
  actual 3D geometry. It's geometry that Autodesk
• 5:33
  software can actually understand. So I could
• 5:35
  generate a world and then click on a wall, and
• 5:38
  Autodesk would know it's a hall with specific
• 5:40
  dimensions. That is the vision. You know, Fei
• 5:43
  -Fei Li said something really profound about
• 5:45
  this. She said, if AI is to be truly useful,
• 5:49
  it must understand worlds, not just words. Worlds,
• 5:52
  not just words. I like that. It's about reconciling
• 5:54
  semantics, the meaning of things with geometry
• 5:57
  and, you know, physics. That distinction is really
• 6:00
  sticking with me because up until now, AI has
• 6:03
  been a wordsmith. It's great at writing essays
• 6:06
  and it's great at making images based on words.
• 6:09
  But it doesn't know gravity. It doesn't know
• 6:11
  gravity or collision or depth. And that's where
• 6:13
  the other side of this deal comes in. It's not
• 6:16
  just about designing buildings, it's about simulation.
• 6:19
  Okay, give me an example of that. Alright, say
• 6:21
  you design a new drone in Autodesk Inventor.
• 6:23
  You have the perfect mechanical engineering for
• 6:26
  the drone. Now you want to test how it flies
• 6:28
  in a dense forest. Right. Which usually means
• 6:31
  you'd have to manually build a 3D forest, which
• 6:34
  takes forever. Correct. But with world labs,
• 6:36
  you could just instantly generate a photorealistic,
• 6:40
  physics -compliant forest. And because it understands
• 6:43
  geometry, the drone can actually crash into a
• 6:45
  tree branch. So it bridges that gap between the,
• 6:48
  like, generative imagination and engineering
• 6:51
  precision. You got it. We are moving from AI
• 6:55
  as a kind of parlor trick to AI as a true industrial
• 6:59
  tool, it's a huge shift. It really is. And to
• 7:01
  understand how they're doing this, we have to
• 7:03
  look at our second story. Because Autodesk isn't
• 7:06
  just buying a brand, they're buying into a specific
• 7:08
  product. A product called Marble. Let's get into
• 7:11
  the weeds on Marble. This is World Labs flagship
• 7:14
  product. They're calling it the first commercial
• 7:16
  world model. Right. And we need to be clear,
• 7:18
  this is different from the video generators people
• 7:20
  have seen, like Sora or Runway. Yeah, because
• 7:23
  those make videos that look 3D. If I watch a
• 7:26
  drone shot generated by Sora, it looks like a
• 7:29
  world. How is Marble different? The key word
• 7:32
  is persistence. When a video generator makes
• 7:34
  a clip, it's just predicting the next frame based
• 7:37
  on the last one. It's dreaming forward. But it
• 7:40
  has no memory of what's behind the camera. If
• 7:42
  you turn the camera around in a generated video,
• 7:45
  the room might be totally different. It gets
• 7:47
  all Morphe. Like a dream where the geography
• 7:49
  keeps shifting. Exactly. Marble builds a consistent
• 7:53
  3D environment. It's not a video, it's a place.
• 7:56
  You can walk around it, leave, come back, and
• 7:59
  the chair is still in the same spot. And you
• 8:01
  can export it. You mentioned in the pre -show
• 8:03
  notes that you can export these worlds as Gaussian
• 8:06
  splats. I have to stop you there. I knew you
• 8:09
  would. That sounds like a messy accident, not
• 8:11
  a file format. What is a Gaussian splat? It does
• 8:13
  sound weird, doesn't it? But it's actually a
• 8:16
  revolutionary way of rendering 3D graphics. So
• 8:19
  traditional 3D uses Triangles polygons to build
• 8:23
  shapes a wireframe mesh mesh Gaussian splatting
• 8:26
  is different. It uses millions of tiny semi -transparent
• 8:29
  blobs or splats that contain information about
• 8:33
  color Opacity and light. Okay. So instead of
• 8:36
  building a house out of rigid triangles, you're
• 8:38
  building it out of a cloud of data points That's
• 8:41
  a great way to think about it. It's like a 3d
• 8:42
  pointillist painting And the advantage is that
• 8:45
  it's incredibly fast to render, and it handles
• 8:47
  things that are hard for polygons, like hair,
• 8:50
  smoke, or complex reflections, really, really
• 8:53
  well. It just looks photorealistic. It does.
• 8:55
  So Marvel uses this tech to let you build worlds.
• 8:59
  And they have some specific tools that sound
• 9:01
  pretty wild. One is called Chisel. Chisel, okay.
• 9:03
  Chisel is the answer to the control problem.
• 9:07
  Usually with AI, you type a prompt and you just
• 9:10
  hope the slot machine gives you what you want.
• 9:12
  Right, you cross your fingers. With Chisel, you
• 9:14
  can block out the scene yourself. You put a gray
• 9:16
  box here, a cylinder there, a sphere over there,
• 9:20
  then you tell the AI, make the box a velvet couch,
• 9:23
  the cylinder a marble column and the sphere a
• 9:26
  disco ball. So you're the stage director setting
• 9:29
  the blocking and the AI is the set designer.
• 9:32
  Precisely. It gives the human creative control
• 9:35
  over the layout while the AI handles the texture
• 9:37
  and the lighting. Justin Johnson, the co -founder,
• 9:40
  he showed off some really specific example. Like
• 9:42
  what? He showed a room made of cheese with grape
• 9:45
  chairs. Of course he did. Why is it always cheese
• 9:48
  with these tech demos? I don't know, but it actually
• 9:51
  proves a point. The AI has to understand the
• 9:53
  materiality of cheese, that it's soft, maybe
• 9:55
  has holes. and apply that to the geometry of
• 9:58
  a room. It also showed a futuristic meeting room
• 10:01
  in space with, like, super realistic lighting
• 10:04
  reflections. And there's one more feature, expansibility.
• 10:07
  This is the Truman Show feature. You can walk
• 10:11
  to the edge of the world you created and just
• 10:12
  tell the model to expand. And it looks at the
• 10:15
  context, OK, we're in a cheese room. And it keeps
• 10:18
  building the world outward in that style. Infinite
• 10:20
  cheese. Infinite cheese. But OK, let's zoom out.
• 10:23
  Why does this matter beyond, say, Video games.
• 10:27
  I get that game developers love this, it'll save
• 10:29
  them thousands of hours. But why is this a connecting
• 10:33
  the dots moment? Because of robotics. This goes
• 10:35
  right back to Fei -Fei Lai's manifesto on spatial
• 10:38
  intelligence. She calls LLMs the chatbots we
• 10:42
  all use. Wordsmiths in the dark. Wordsmiths in
• 10:45
  the dark. Yeah. Wow, that is evocative. It means
• 10:47
  they can write a poem about an apple, but they
• 10:49
  don't know that if you drop an apple, it bruises.
• 10:51
  They lack common sense physics. Right. If we
• 10:54
  want robots to be useful, to fold laundry, to
• 10:57
  cook dinner, to stack boxes, they need to understand
• 11:01
  spatial relationships. So marble isn't just the
• 11:04
  background generator. It's a training ground.
• 11:07
  Exactly. It's a simulator. Before you let a robot
• 11:10
  try to pour hot coffee in your actual kitchen,
• 11:13
  you let it practice a million times in a marble
• 11:15
  -generated simulation that obeys the laws of
• 11:18
  physics. So we start with rooms made of cheese,
• 11:21
  but we end up with robots that can actually do
• 11:24
  the dishes without breaking them. That is the
• 11:27
  trajectory. It starts with gaming and VFX because
• 11:30
  those industries have the money and the immediate
• 11:32
  need, but the end game is scientific simulation
• 11:35
  and real -world robotics. I love that. We're
• 11:39
  watching the digital world learn physics. Now,
• 11:41
  let's pivot. We've talked about building 3D worlds.
• 11:44
  Let's talk about building financial portfolios.
• 11:46
  A very different kind of world building. This
• 11:49
  third story is something that might actually
• 11:50
  impact your wallet directly. This is a really
• 11:53
  interesting move by Robinhood. We're talking
• 11:55
  about the IPO of Robinhood Ventures Fund I, ticker
• 11:58
  symbol RVI. So Robinhood, the app that got everyone
• 12:01
  into trading stocks on their phones, is now trying
• 12:04
  to get everyone into, what, private equity? Essentially,
• 12:08
  yes. They are launching a closed -end fund designed
• 12:11
  to give retail investors everyday people access
• 12:14
  to private companies. Let's pause and explain
• 12:16
  the why here. For the average listener, why does
• 12:19
  it matter if a company is public or private?
• 12:22
  It matters because of where the growth happens.
• 12:25
  Think about a company like Facebook or Google.
• 12:28
  By the time they went public, they were already
• 12:30
  massive. Right. The early investors, the venture
• 12:33
  capitalists, they're the ones who made the 100x
• 12:35
  or 1000x returns. Usually, regular people can't
• 12:39
  invest in a company until it hits the stock market.
• 12:41
  Because of the accredited investor rules, right?
• 12:43
  Correct. Generally, to invest in a private startup,
• 12:47
  the government requires you to be an accredited
• 12:50
  investor. That means you need a net worth of
• 12:52
  over a million dollars or a very high annual
• 12:55
  income. The idea being that private investments
• 12:58
  are risky, so they should protect regular people
• 13:00
  from them. Right. But the counter argument is
• 13:03
  that you are protecting regular people from the
• 13:05
  biggest gains in the economy. Yeah, that's Robinhood's
• 13:08
  whole argument. Exactly. They point out there
• 13:10
  are now 6 .5 times more private companies than
• 13:13
  public ones in the U .S. The value of these private
• 13:16
  firms tops $10 trillion. If you're only in the
• 13:19
  stock market, you're missing a massive chunk
• 13:21
  of the innovation economy. So enter RVI. What
• 13:24
  are the details? How does this actually work
• 13:26
  for me? So it's an IPO launching February 26th
• 13:29
  on the NYSE. They're offering 40 million shares
• 13:32
  at about $25 a share. 25 bucks. That is accessible.
• 13:36
  It is. And the portfolio is pretty impressive.
• 13:39
  The fund holds shares in companies like Aerowallex,
• 13:42
  Boom, the supersonic jet company, Databricks,
• 13:44
  Aura. The ring I'm wearing right now. Exactly.
• 13:47
  Ramp, Revolut. And breaking news, they have an
• 13:50
  agreement to buy shares of Stripe, which is expect
• 13:52
  it to close after the IPO. Whoa, Stripe is huge.
• 13:56
  Investors have been waiting for them to go public
• 13:58
  for years. And this is a way to get exposure
• 14:00
  to Stripe before they actually do. OK, look,
• 14:02
  I have to play the skeptic here. This sounds
• 14:04
  great access for everyone, but finance is rarely
• 14:07
  that simple. What's the catch? There's always
• 14:09
  a catch. This isn't just buying a share of Apple,
• 14:11
  is it? No, and this is where listeners really
• 14:14
  need to pay attention. The structure matters.
• 14:16
  This is a closed end fund. Unpack that. How is
• 14:20
  that different from an ETF? In a normal ETF,
• 14:23
  if you sell your shares, the fund can basically
• 14:26
  sell the underlying assets to give you your money.
• 14:30
  But Robinhood can't just sell a share of Stripe
• 14:32
  tomorrow. It's a private company. It's illiquid.
• 14:35
  It's illiquid. So if I want to sell my RVI stock,
• 14:37
  who am I selling it to? Another investor on the
• 14:39
  stock market. Exactly. And that creates a unique
• 14:42
  risk. The price of RVI can drift away from the
• 14:45
  actual value of the companies inside it. Drift
• 14:48
  how? Well, it can trade at a premium. or a discount.
• 14:52
  So let's say the Stripe and Databricks shares
• 14:54
  inside are worth $25. If everyone goes crazy
• 14:57
  for RVI, the stock might trade at $30. They're
• 15:00
  paying a premium. But if the market gets scared,
• 15:02
  it might trade at $20 a discount to its actual
• 15:04
  value. So you could technically lose money on
• 15:06
  the stock even if the companies inside are doing
• 15:08
  well. Theoretically, yes. That disconnect is
• 15:11
  the risk of a closed end fund. Also, we have
• 15:13
  to talk about fees. There's always a fee. There
• 15:16
  is a 2 % annual management fee. Now, they are
• 15:19
  discounting it to 1 % for the first six months,
• 15:22
  but 2 % is standard for private equity. But it's
• 15:25
  expensive compared to an index fund that charges
• 15:28
  like 0 .03%. Very expensive. And dividends. Don't
• 15:32
  hold your breath. These are high growth tech
• 15:34
  companies. They reinvest everything. Everything.
• 15:36
  You are betting entirely on the share price going
• 15:38
  up over the long, long term. So this is not a
• 15:41
  savings account. No, this is a high risk, high
• 15:43
  reward bet on the future of tech. It is. It's
• 15:46
  the democratization of venture capital. It fixed
• 15:49
  Robinhood's brand perfectly, but it definitely
• 15:51
  shifts the risk curve for the average investor.
• 15:54
  You have to know what you're buying. You do.
• 15:56
  It really connects nicely to our earlier stories,
• 15:58
  though. With Autodesk and World Labs, we have
• 16:01
  the tools of creation neural CAD, marble moving
• 16:03
  from research labs to commercial products. And
• 16:06
  with Robinhood, we have the tools of wealth creation,
• 16:08
  private equity, moving from the elite to the
• 16:11
  everyday trader. It's all about access, access
• 16:13
  to world building tools and access to wealth
• 16:15
  building markets. The walls are coming down.
• 16:18
  But as the walls come down, the responsibility
• 16:20
  goes up. You have to understand the tool before
• 16:22
  you build the world. And you have to understand
• 16:24
  the fund before you buy the stock. Wall said.
• 16:27
  Yeah. All right. Speaking of access to great
• 16:29
  stuff. Yeah. Here comes the transition. I knew
• 16:32
  it was coming. Before we wrap up, I want to shout
• 16:35
  out our other sponsor, Old Glory. If you have
• 16:38
  a techie friend or a music lover or a sports
• 16:40
  fanatic in your life, you need to check them
• 16:42
  out. Old Glory is iconic. They have the best
• 16:46
  band tees, sports merch, just really high quality
• 16:49
  stuff. It's the kind of gear that lasts forever.
• 16:51
  It is. And because you're listening to us, you
• 16:53
  can get 15 % off your entire purchase. Just go
• 16:55
  to oldglory .com - only say it once oldglory
• 16:59
  .com - and use promo code SNARFL at checkout.
• 17:03
  Get yourself a vintage tee. You deserve it. All
• 17:05
  right. Let's do a quick recap of what we covered
• 17:07
  today for everyone commuting right now. Story
• 17:10
  one. Autodesk drops $200 million into world labs.
• 17:14
  They're building NeuralCAD, merging generative
• 17:17
  AI with serious engineering tools to bridge that
• 17:19
  gap between imagination and manufacturing. Story
• 17:22
  two. The engine behind that is Marble, a new
• 17:27
  AI product that builds persistent, editable 3D
• 17:30
  worlds using Gaussian splats. It's moving us
• 17:34
  toward true spatial intelligence, where computers
• 17:37
  understand physics, not just pixels. And story
• 17:40
  three, Robinhood is launching the RVI fund. It's
• 17:45
  a $25 ticket for retail investors to own a piece
• 17:47
  of private giants like Stripe and Databricks,
• 17:51
  though it comes with unique risks like premiums
• 17:53
  and discounts. It is a fascinating time to be
• 17:55
  watching these dots connect. The line between
• 17:58
  the digital and physical is blurring, and the
• 18:00
  barriers to high finance are crumbling. The tools
• 18:03
  of the elite are moving into the hands of the
• 18:05
  many. It makes you wonder what the landscape
• 18:07
  will look like even a year from now. It really
• 18:08
  does. Well, that is all we have time for today.
• 18:10
  Thank you so much for starting your morning with
• 18:12
  us. It's been a pleasure. We're here to help
• 18:14
  at Snarfles, so reach out on our website if you
• 18:16
  have any more questions. Have a great day, everyone.
• 18:21
  That's a wrap on today's top stories. Remember,
• 18:25
  tech doesn't have to be complicated, and we're
• 18:27
  here to help connect the dots. If you've got
• 18:29
  questions or want to learn more, visit us at
• 18:32
  snarful .com. Thanks for listening and we'll
• 18:35
  catch you next time.