Latest / Connecting the Dots / Ads Come to ChatGPT, Runway Bets Big, and Markets Stay Mixed

Transcript

Ads Come to ChatGPT, Runway Bets Big, and Markets Stay Mixed · Connecting the Dots

• 0:00
  Today's tech news moves fast, but understanding
• 0:03
  it shouldn't be hard. That's why Snarfl 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 hosts.
• 0:16
  Good morning. It is Tuesday, February 10th, 2026.
• 0:20
  This is Connecting the Dots brought to you by
• 0:22
  the Snarfl Solutions Group. And today we are
• 0:24
  looking at a tale of two very different bank
• 0:28
  accounts. On one side you have companies desperately
• 0:31
  trying to figure out how to pay for the AI tools
• 0:34
  you're already using. Right. And on the other
• 0:36
  you have investors backing up the armored trucks
• 0:38
  to fund the next generation of intelligence.
• 0:41
  AI that doesn't just talk but actually sees.
• 0:44
  It is essentially the cost of intelligence edition
• 0:46
  today. The bill is coming due for the chatbots
• 0:49
  while the checkbook is wide open for the physics
• 0:51
  simulators. Exactly. But before we get into the
• 0:54
  silicon and the strategy, let's talk about the
• 0:56
  physical reality outside your window. Let's open
• 0:59
  up the Morningstar Weather Report. We have to
• 1:01
  start up north. I mean, if you are involved in
• 1:03
  logistics or transport in Alaska, you are looking
• 1:06
  at a nightmare scenario. Wow. There is a massive,
• 1:09
  powerful storm system just hammering the state
• 1:11
  right now. And we aren't just talking about a
• 1:14
  little snow. This is serious. We're talking dangerous
• 1:17
  seas, blizzard conditions, and bitter cold. that
• 1:23
  is going to lock things down right through the
• 1:25
  middle of the week. So if you have supply chains
• 1:27
  running through Anchorage, you should definitely
• 1:29
  expect delays. And that system has a tail, doesn't
• 1:31
  it? It does. Down here in the lower 48, specifically
• 1:35
  out west, that Pacific storm is finally moving
• 1:38
  inland across California. OK. If you're in the
• 1:41
  valleys, it's going to be a slick, rainy commute.
• 1:44
  If you're heading up into the mountains, however,
• 1:45
  you're getting a serious dusting of snow, which
• 1:48
  is fantastic for the ski resorts. Absolutely.
• 1:50
  But. Maybe less fantastic if you're trying to
• 1:53
  drive a semi -truck over the pass. True. Moving
• 1:56
  east, the Central Great Basin needs to watch
• 1:58
  out for high winds today. And for our listeners
• 2:01
  on the East Coast, the Northeast is dealing with
• 2:03
  a clipper system. Ah, the clippers. Yeah, it's
• 2:05
  fast, it's bringing snow, and the gusty winds
• 2:08
  are going to linger well into tomorrow, Wednesday.
• 2:11
  And for us, right here in Sacramento, it is looking
• 2:15
  like a classic gray February start. Mostly cloudy
• 2:18
  this morning. But you definitely want to grab
• 2:20
  the umbrella because it is turning to rain later
• 2:22
  on. We're looking at a high of 57 degrees Fahrenheit.
• 2:25
  Standard fare for us. Now, let's look at the
• 2:27
  scoreboard. How did the financial markets treat
• 2:29
  your portfolio yesterday? It was a pretty green
• 2:32
  day for the industrialists. The Dow Jones had
• 2:34
  a very solid session. OK, it added just over
• 2:36
  200 points. That's up point four percent. closing
• 2:40
  at 50 ,394 .12. So it broke that psychological
• 2:45
  50k barrier and it's actually holding it. That
• 2:47
  feels significant for market confidence. It really
• 2:50
  does. The S &P 500, however, didn't quite catch
• 2:53
  the same updraft. It took a bit of a breather,
• 2:56
  dipping slightly by about 14 points. Right. But
• 2:58
  the real volatility, as is often the case, was
• 3:01
  in crypto. Yeah, if you're watching the charts
• 3:03
  last night, you saw a lot of red candles. Bitcoin
• 3:05
  had a rough 24 hours. How rough? It's down over
• 3:08
  2%, shedding about, what, 1 ,420 points. It's
• 3:11
  currently trading around $68 ,688. Oof, that
• 3:15
  is a sharp slide. It is. And honestly, staring
• 3:19
  at red candles on a crypto chart is enough to
• 3:22
  drain anyone's energy, which, uh... brings us
• 3:25
  to something that might help with that sluggish
• 3:27
  feeling. It does. We are really happy to welcome
• 3:30
  a new partner to the show today, Strong Coffee.
• 3:34
  Now I know when people hear instant coffee, they
• 3:37
  usually think of that. you know, bitter sludge
• 3:39
  you get in a hotel room. I was skeptical too,
• 3:42
  I'll admit it. Right. But this is different.
• 3:43
  It's premium instant, so it tastes like a proper
• 3:46
  traditional brew, but the formulation is what
• 3:48
  really caught my eye. It's not just caffeine.
• 3:50
  Not at all. It is packed with 15 grams of protein,
• 3:53
  MCTs, and adaptogens. Specifically ashwagandha.
• 3:56
  Right, which is huge for stress modulation. So
• 3:58
  instead of that jittery spike and crash you get
• 4:01
  from sugary energy drinks or, you know, just
• 4:03
  chugging espresso. You get a clean, sustained
• 4:06
  burn. It's fuel, not just a stimulant. Precisely.
• 4:09
  And because you're listening to us, you get a
• 4:11
  break on the price. You can get 20 % off your
• 4:13
  entire purchase. Just use the promo code SNARFL,
• 4:16
  that is S -N -A -R -F -U -L -L, at strongcoffeecompany
• 4:20
  .com. It is definitely worth trying. Your brain
• 4:22
  will thank you. All right, let's open up the
• 4:24
  Morningstar and get into the main event. We have
• 4:27
  to start with the story that is just dominating
• 4:30
  the tech feeds this morning. Yeah. The free lunch
• 4:33
  era of AI is officially ending. As of today,
• 4:37
  ads have arrived in chat GPT. You know, we have
• 4:41
  been waiting for the shoe to drop for a long,
• 4:43
  long time. It was inevitable, but the timing
• 4:46
  of it. The timing is incredibly ironic. It really
• 4:49
  is. We're talking about the Super Bowl. I am.
• 4:51
  Just yesterday, during the Super Bowl, Anthropic
• 4:54
  aired that commercial. You probably saw it. Basically
• 4:57
  planting a flag and bragging about no ads in
• 4:59
  AI. There's a direct shot across the bow. A huge
• 5:02
  shot. And then, less than 24 hours later, OpenAI
• 5:05
  rolls this out. It does feel like a coordinated
• 5:08
  battle for the soul of the interface. But let's
• 5:10
  get into the specifics because ads can mean a
• 5:12
  lot of things. If I log in today, am I getting
• 5:14
  pop -ups? Is it going to look like those recipe
• 5:17
  blogs where I have to scroll past five videos
• 5:20
  to get the answer? Thankfully, no. They're being
• 5:23
  fairly cautious with the implementation, it seems.
• 5:26
  According to the release, these are contextually
• 5:30
  relevant ads that appear at the bottom of the
• 5:32
  chat bot's answers. Okay, at the bottom. At the
• 5:34
  bottom. and they're clearly labeled as sponsored.
• 5:38
  So it's reading the context of my chat to decide
• 5:40
  what to show me. Exactly. The example they gave
• 5:42
  is pretty straightforward. If you're asking chat
• 5:44
  GPT for a dinner recipe, it might generate the
• 5:47
  recipe. And then at the very bottom, you'd see
• 5:49
  a sponsored link for a grocery delivery service.
• 5:52
  Or a specific brand of pasta or something. Right.
• 5:55
  It's triggered by the intent of the conversation.
• 5:57
  OK, that sounds familiar. That is essentially
• 5:59
  the Google search model adapted for chat. But
• 6:02
  here's the big question. Who is seeing this?
• 6:06
  Because I pay 20 bucks a month for the plus subscription
• 6:09
  right if I see an ad I'm gonna be canceling you
• 6:12
  are safe open AI was very explicit about this
• 6:15
  plus Pro business and enterprise users are all
• 6:18
  exempt. Okay good This is strictly for the free
• 6:21
  tier users and the users on chat GPT go their
• 6:24
  lower cost plan Yeah, the $8 a month plan. Wait,
• 6:27
  hold on the $8 plan still sees ads that feels
• 6:31
  like a tough sell You're paying, but you're still
• 6:34
  the product. It's a hybrid model. I mean, it's
• 6:37
  similar to what we see with some streaming services
• 6:39
  like Netflix or Hulu, right? You pay a little
• 6:41
  for access, but the ads subsidize the rest. I
• 6:44
  guess so. And I think this brings us to the privacy
• 6:47
  concern, because ads based on conversation sounds
• 6:51
  a little dystopian to some people. It does. It
• 6:53
  implies the model is analyzing your private thoughts
• 6:56
  to sell you stuff. Exactly. So what are the guardrails
• 6:59
  here? Because if I'm talking about something
• 7:01
  personal, I don't want a billboard for it. They've
• 7:03
  established a few pretty strict rules. First,
• 7:06
  they state that the ads will not influence the
• 7:08
  chat bot's actual answers. That is critical.
• 7:12
  Okay, so the output remains neutral. The ad is
• 7:14
  just an add -on. Correct. A brand can't pay to
• 7:17
  have chat GPT say their product is objectively
• 7:19
  the best in the text it generates. Got it. Second,
• 7:23
  they aren't selling your conversation history
• 7:25
  to advertisers. And third, and this is probably
• 7:27
  the most important, 1. Ads are blocked entirely
• 7:31
  for sensitive topics. Define sensitive. Health.
• 7:34
  Mental health. politics okay that is a big relief
• 7:37
  so if you're typing in symptoms because you think
• 7:40
  you have the flu you aren't going to get hit
• 7:42
  with a buy this medicine ad right or if you're
• 7:45
  trying to understand a complex political bill
• 7:47
  you won't see campaign ads they're also blocking
• 7:50
  ads for any account flagged as being under 18
• 7:52
  so let's zoom out and look at the strategy here
• 7:55
  we all remember Sam Altman the CEO previously
• 7:59
  warning that ads could erode trust in AI he was
• 8:02
  very vocal about that and now Here we are. What
• 8:05
  changed? The math changed. Or rather, the math
• 8:08
  became undeniable. The cost. Yeah, we have to
• 8:10
  look at the infrastructure costs. OpenAI has
• 8:13
  somewhere in the ballpark of $1 .4 trillion.
• 8:15
  With a T. With a T in infrastructure commitments.
• 8:18
  Trillion. That is a staggering number. It's hard
• 8:21
  to even visualize. It is. And while they have
• 8:23
  a reported $20 billion annualized revenue run
• 8:27
  rate, the cost to run these models, the electricity,
• 8:29
  the GPUs, the cooling is just... astronomical.
• 8:32
  So they need a scalable revenue stream that grows
• 8:34
  with usage. Exactly. Subscriptions are great,
• 8:37
  but ads are the engine of the internet for a
• 8:39
  reason. It's the Spotify model, essentially.
• 8:41
  You either pay with your wallet for the premium
• 8:43
  experience, or you pay with your attention to
• 8:46
  keep the free tier sustainable. That's it. And
• 8:49
  they've explicitly stated that the goal here
• 8:51
  is to keep the free tier viable and, you know,
• 8:55
  maybe reduce usage limits for non -paying users.
• 8:58
  It brings up a real dilemma for the listener,
• 9:00
  then. if you're on the free tier, is a banner
• 9:03
  at the bottom of the screen enough to make you
• 9:05
  switch to paid? I think most people will just
• 9:08
  ignore it, you know, provided it stays at the
• 9:10
  bottom. But it marks a major shift. The pure
• 9:13
  research phase is over. This is a commercial
• 9:15
  utility now. Speaking of massive amounts of money
• 9:18
  and commercial ambition, let's pivot to our second
• 9:21
  story. Because while OpenAI is trying to pay
• 9:24
  the electric bill for the current models, other
• 9:27
  investors are pouring cash into what comes next.
• 9:30
  This is a huge story that might have flown under
• 9:32
  the radar for some because it's a bit more technical,
• 9:35
  but it's so crucial. Runway. The AI video company.
• 9:39
  Right. The New York -based startup that really
• 9:41
  pioneered AI video generation just announced
• 9:45
  a massive, series -y funding round. They raised
• 9:49
  $315 million. Wow. $315 million. And that pushes
• 9:55
  their valuation to what? $5 .3 billion. That's
• 9:59
  nearly double their previous valuation. It is.
• 10:01
  And the list of people writing the checks is
• 10:03
  a who's who of the tech sack. You have General
• 10:06
  Atlantic leading, but then you have Nvidia, Adobe,
• 10:09
  and AMD all chipping in. That lineup tells a
• 10:11
  story right there. Nvidia and AMD make the chips
• 10:14
  this runs on. Adobe makes the creative tools
• 10:16
  this integrates with. Exactly. But Runway isn't
• 10:18
  just using this money to make better deepfake
• 10:21
  videos, are they? No, and this is where we need
• 10:23
  to connect the dots for the listener. Most people
• 10:25
  know Runway for their Gen series models making
• 10:27
  cool cinematic clips from text prompts, but this
• 10:31
  funding is explicitly for building something
• 10:33
  called world models. Okay, world models. We hear
• 10:35
  this term thrown around a lot in research papers.
• 10:38
  Let's break it down. How is a world model different
• 10:40
  from what ChatGPT does? Think of it like this.
• 10:43
  A large language model like ChatGPT is essentially
• 10:47
  a probability machine. It predicts the next word
• 10:50
  in a sentence. It's text math. It's text math.
• 10:53
  It doesn't know what a cat is. It just knows
• 10:56
  that the word cat mathematically follows the
• 10:58
  fluffy. Right. It understands syntax, not reality.
• 11:01
  Exactly. A world model is trying to do something
• 11:04
  much, much harder. The quote from Runway is,
• 11:07
  to predict the next frame, a video model must
• 11:10
  learn how the world works. So it's trying to
• 11:13
  construct an internal simulation of the physical
• 11:15
  environment. Precisely. So instead of guessing
• 11:17
  a word, it's understanding physics. It knows
• 11:20
  that if I drop a glass, gravity pulls it down.
• 11:23
  And when it hits the floor, it shatters. It doesn't
• 11:25
  just draw a shattered glass, it simulates the
• 11:28
  event. Bingo. It is the difference between describing
• 11:31
  a car crash in a book and simulating the physics
• 11:34
  of metal crumpling in a computer program. And
• 11:36
  that distinction is why this matters way beyond
• 11:38
  Hollywood. This isn't just for filmmakers. Not
• 11:41
  at all. If you have an AI that understands 3D
• 11:43
  space, object permanence, and physics, You apply
• 11:46
  that to robotics. That's the holy grail. If you
• 11:49
  want a robot to navigate a messy kitchen or a
• 11:52
  construction site, you can't just give it a language
• 11:54
  model. A robot needs to understand, if I push
• 11:57
  this heavy box, it will slide. If I push this
• 11:59
  tall lamp, it will tip over. It needs a mental
• 12:02
  simulator. That is a great way to put it. It
• 12:04
  is a simulator for machines. Runway is already
• 12:07
  working with robotics and autonomous vehicle
• 12:10
  companies. So they're using these video models
• 12:11
  to create training grounds for self -driving
• 12:13
  cars. Yes. Instead of driving a real car a million
• 12:16
  miles to learn how to avoid a pedestrian, you
• 12:19
  can simulate that scenario in a world model a
• 12:22
  million times in an hour. And it goes even further,
• 12:24
  right? Medicine, climate modeling. Energy, yep.
• 12:27
  These are all physical systems. You can't solve
• 12:30
  climate change just with text. You need to model
• 12:32
  complex, chaotic physical interactions. It sounds
• 12:35
  incredible, but is Runway the only player here?
• 12:38
  I mean, $315 million is a lot, but in this game,
• 12:41
  it isn't infinite. Oh, no. The race is incredibly
• 12:44
  crowded. You have World Labs, which is Feifei
• 12:46
  Lai's new company. She is an absolute legend
• 12:50
  in AI. You have Google DeepMind with their genie
• 12:53
  model. Luma AI is right there too. Everyone is
• 12:56
  racing towards this. So the consensus among the
• 12:58
  smartest people in the room is that the next
• 13:01
  frontier isn't just better chatbots. It is AI
• 13:05
  that interacts with the physical world. We are
• 13:07
  moving from an AI that talks to an AI that sees
• 13:10
  and understands. And looking at the investors
• 13:13
  in video, especially, they know that this requires
• 13:15
  massive, massive compute power. It's a symbiotic
• 13:19
  relationship. It feels like we are watching the
• 13:21
  industry grow up in real time. We went from write
• 13:23
  me a poem to simulate a robotic surgery in about
• 13:27
  three years. It is moving so fast. And frankly,
• 13:30
  it makes the ads and chat GPT story feel almost
• 13:33
  quaint. We're worrying about banner ads while
• 13:35
  these labs are trying to digitize reality. It
• 13:38
  really does put things in perspective. Before
• 13:40
  we wrap up today, I want to mention one more
• 13:41
  way to upgrade your own physical reality, specifically
• 13:44
  if you spend time on the golf course. Or even
• 13:46
  if you just like being comfortable. We are talking
• 13:48
  about Skone. Skone makes some of the most versatile
• 13:51
  golf shoes out there. And I say versatile because
• 13:53
  they don't look like traditional clunky cleats.
• 13:55
  No, they're stylish enough to wear to dinner
• 13:57
  after your round, but they have the performance
• 13:58
  you need for 18 holes. And we have a deal for
• 14:01
  you. Use the promo code SNARFL at checkout for
• 14:03
  15 % off your entire purchase. It is a solid
• 14:06
  upgrade for your game. Check them out at scony
• 14:09
  .com. So let's do a quick recap of what we covered
• 14:12
  today because these are two pivotal moments that
• 14:14
  signal where we're all going. First, the monetization
• 14:17
  engine has arrived. ChatGPT is rolling out ads
• 14:19
  to free users here in the U .S. Right. It's a
• 14:23
  necessary move to pay for the massive infrastructure
• 14:25
  costs. It is the Internet 2 .0 business model
• 14:28
  just applied to AI. And secondly, the smart money
• 14:31
  is betting on world models. Runway's huge $315
• 14:36
  million raise proves that the industry is shifting
• 14:39
  focus from just language processing to physical
• 14:42
  simulation. It is about connecting the digital
• 14:44
  brain to the physical world, and that is going
• 14:46
  to change everything from filmmaking to how robots
• 14:49
  walk through our homes. It leaves us with a provocative
• 14:52
  thought to chew on. If these world models get
• 14:55
  good enough, if they can perfectly simulate physics,
• 14:59
  weather, and human movement, at what point does
• 15:02
  the simulation become more valuable data than
• 15:05
  the real world itself? That is a question we
• 15:07
  might be answering sooner than we think. Well,
• 15:10
  that is all the time we have for today. Thank
• 15:12
  you so much for starting your morning with us.
• 15:14
  We're here to help at SNARFL, so reach out on
• 15:16
  our website if you have any more questions. That's
• 15:22
  a wrap on today's top stories. Remember, tech
• 15:25
  doesn't have to be complicated, and we're here
• 15:27
  to help connect the dots. If you've got questions
• 15:30
  or want to learn more, visit us at snarful .com.
• 15:33
  Thanks for listening, and we'll catch you next
• 15:35
  time.