Latest / Connecting the Dots / Power Plays: OpenAI’s Infrastructure Ambitions and Microsoft’s Humanist AI Shift

Transcript

Power Plays: OpenAI’s Infrastructure Ambitions and Microsoft’s Humanist AI Shift · 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
• 0:14
  hosts. Good morning, and welcome back to the
• 0:18
  Snarfl Solutions Group tech report. It is Thursday,
• 0:21
  November 6th, 2025. Great to be here. Yeah, great
• 0:24
  to be connecting the dots again. Today, we're
• 0:26
  looking at something huge. The race for artificial
• 0:30
  general intelligence. It's really heating up
• 0:33
  We've got some reports today that honestly they
• 0:35
  show the incredible scale of money and ambition
• 0:37
  involved and maybe more importantly What happens
• 0:40
  when the two biggest players open AI and Microsoft
• 0:42
  start taking really different roads to get there?
• 0:45
  Yeah, we're talking about absolutely massive
• 0:47
  bets some real political friction building up
• 0:50
  and this fundamental argument over control Versus
• 0:54
  well just letting these things run But OK, before
• 0:56
  we dive into all that, let's quickly set the
• 0:58
  scene. Sure. So locally here in Sacramento, looks
• 1:02
  like it's clearing up, becoming sunny high near
• 1:05
  68 degrees. Pretty nice. Nice indeed. Nationally,
• 1:08
  though, bit of a different story. Some pretty
• 1:11
  rough weather in the Pacific Northwest, heavy
• 1:13
  rain, even snow up in the Cascades. And meanwhile,
• 1:17
  dry and windy conditions mean elevated fire weather
• 1:20
  concerns across the high plains and upper Midwest.
• 1:23
  So keep an eye on that if you're out there. Right.
• 1:25
  And the markets are definitely paying attention
• 1:28
  to all this AI infrastructure news. Stocks had
• 1:30
  a generally positive day. Dow Jones, S &P 500,
• 1:34
  both up just a little bit. Investors seem to
• 1:37
  be digesting things. Bitcoin, though, saw a bit
• 1:39
  of a dip, down about 1 .2%. So, you know, overall,
• 1:44
  finance is still laser -focused on these tech
• 1:46
  giants. Absolutely. Okay, now a quick word for
• 1:48
  you. If you're into iconic sports and music merch,
• 1:51
  or maybe you've got that techie friend, you know,
• 1:53
  the one who's hard to shop for, holidays are
• 1:55
  coming up, you should really check out our sponsor,
• 1:57
  Old Glory. Oh yeah, they have some great stuff.
• 1:59
  They really do. Incredible fan merchandise. And
• 2:01
  here's the deal. Use our promo code SNARFUL,
• 2:04
  that's S -N -A -R -F -U -L at checkout, and you'll
• 2:07
  get 15 % off your entire purchase. Just head
• 2:10
  over to OldGlory .com. That's a solid deal. Definitely
• 2:13
  worth checking out. Totally. Okay, let's get
• 2:15
  into it, the core segment. The SNARFL Tech article
• 2:18
  report, our morning star. We're zeroing in on
• 2:22
  three big shifts, strategic and philosophical,
• 2:26
  between the giants and AI. And we kind of have
• 2:29
  to start with the sheer scale, right? Oh, absolutely.
• 2:31
  You have to start with open AI, because the scale
• 2:34
  of what they're trying to do, spearheaded by
• 2:37
  their builder in chief, Greg Brockman. It's almost
• 2:40
  tough to wrap your head around. You know Sam
• 2:42
  Altman's the visionary that the face out there
• 2:44
  selling the future, right? But Brockman he's
• 2:46
  the operator the guy actually executing the plan
• 2:49
  for this thing They call the Stargate project
• 2:51
  and Stargate isn't just code is it's concrete
• 2:53
  and power lines exactly the reports we're seeing
• 2:56
  suggest open AI has committed something like
• 2:58
  1 .4 trillion dollars to infrastructure. Okay,
• 3:01
  let's just hit pause there 1 .4 trillion. I mean
• 3:05
  is that number actually real? Or is it, you know,
• 3:08
  maybe just big talk to lure investors? Well,
• 3:11
  that is the trillion -dollar question, isn't
• 3:13
  it? Right. Figuratively and literally. Because
• 3:16
  you put that $1 .4 trillion against their current
• 3:20
  reported revenue, which is, what, only $13 billion
• 3:23
  a year? Yeah, tiny in comparison. It just throws
• 3:25
  into sharp relief what a massive high -stakes
• 3:28
  gamble this whole thing is. Brockman apparently
• 3:30
  calls AGI an end -to -end engineering challenge.
• 3:33
  It's all about execution for him. Just build
• 3:35
  it. Fast. Like you see that speed in his background,
• 3:39
  right? Like, Stripe's fourth employee drops out
• 3:43
  of MIT. Yeah, hacked together their first commercial
• 3:45
  API in a weekend, supposedly. The guy clearly
• 3:47
  operates at a different speed. Just another word,
• 3:49
  slow. Right. And that drive is why he's running
• 3:52
  their new scaling group, managing all that compute
• 3:54
  power. It makes him, frankly, one of the biggest
• 3:56
  power brokers in this whole AI era. And the compute
• 3:59
  needs are just astronomical, the AMD partnership
• 4:02
  alone. Oh, yeah. Tens of billions of dollars.
• 4:04
  And get this. It requires six gigawatts of computing
• 4:09
  power. Six gigawatts? What does that even mean
• 4:11
  in real terms? Well, think about the Hoover Dam.
• 4:14
  This is roughly three times the electricity generated
• 4:17
  by the entire Hoover Dam. Three times the Hoover
• 4:19
  Dam? Just for OpenAI's models? Wow. Okay, that
• 4:23
  really puts it in perspective. It does. And it
• 4:26
  tells you something about the competition now.
• 4:28
  It's measured in watts, not just code. AMD's
• 4:31
  CEO, Lisa Su, she apparently described Brockman's
• 4:34
  focus here. She called compute the currency of
• 4:38
  intelligence If you can't play at this massive
• 4:41
  infrastructure scale, you're already falling
• 4:43
  behind Which is probably why Brockman said this
• 4:45
  makes the Apollo program look almost small big
• 4:48
  words huge words But maybe not entirely hyperbole
• 4:51
  when you look at the resources involved but building
• 4:53
  something that big that fast Yeah, it has to
• 4:56
  rub some people the wrong way. There are costs
• 4:58
  beyond just the dollars. Absolutely two major
• 5:01
  friction points are popping up First, energy
• 5:03
  prices. Sucking up that much power is already
• 5:05
  starting to push utility costs up in some areas.
• 5:09
  And second, politics. These huge data centers.
• 5:12
  They're becoming local election issues. It's
• 5:15
  kind of hard to sell community on a giant power
• 5:18
  hungry facility when their own electric bills
• 5:20
  are soaring. Yeah, I can see that. And then there's
• 5:22
  the money itself. The financing sounds complicated,
• 5:27
  maybe a bit murky. It does get into some murky
• 5:29
  waters, yeah. This is where, you know, if you're
• 5:31
  following closely, you need to pay attention.
• 5:33
  The reports mention Nvidia, who makes the chips
• 5:36
  a key supplier, potentially discussing guaranteeing
• 5:38
  loans for OpenAI's data centers. So the company
• 5:41
  selling the gear is also helping finance the
• 5:43
  purchase of that gear. Exactly. And that makes
• 5:45
  analysts really nervous because, you know, it
• 5:48
  starts to look like artificial demand. It tangles
• 5:50
  everything up, makes it hard to see the real
• 5:52
  risk investors are taking on. Less like open
• 5:55
  market competition, more like, I don't know,
• 5:57
  closed loop between a few giants. That's the
• 6:00
  concern. One analyst warned about potential deflating
• 6:03
  activity in valuations across the whole sector
• 6:05
  if investors feel these ties are too close. So
• 6:09
  the success of this trillion -dollar bet really
• 6:11
  depends on keeping genuine investment separate
• 6:13
  from this sort of intertwined capital structure.
• 6:17
  Okay, so open AI is just pedal to the metal,
• 6:20
  full speed ahead towards AGI, disrupting power
• 6:23
  grids, tangling up finances. Just going for it.
• 6:26
  Yeah, that's the picture. But then you look at
• 6:28
  their biggest partner, Microsoft, and it seems
• 6:30
  like, well, a totally different vibe. Which brings
• 6:33
  us nicely to topic two, Microsoft's strategy.
• 6:37
  Let's call it the humanist superintelligence
• 6:40
  strategy. It's led by Mustafa Suleyman. And while
• 6:43
  open AI is all about accelerate, accelerate,
• 6:45
  Microsoft is saying, OK, yes, we want superintelligence
• 6:47
  too, but hold on, they're big condition, prioritizing
• 6:51
  human control over just raw capability. Right.
• 6:54
  I remember Suleyman warning that accelerating
• 6:56
  at all costs was like a crazy suicide mission.
• 6:59
  So what's the core difference in philosophy here?
• 7:01
  What makes their approach supposedly safer? It
• 7:04
  really boils down to consciousness. Suleiman
• 7:06
  draws a very sharp line. He says, AI is intelligent,
• 7:09
  yes, but it is not conscious. He's pretty emphatic.
• 7:13
  AI doesn't feel pain. It doesn't have subjective
• 7:15
  experience. It just creates the illusion of it,
• 7:18
  the seeming narrative of experience, as he put
• 7:20
  it. It's a mimic, a sophisticated tool. So Microsoft's
• 7:24
  position is essentially, we build amazing tools,
• 7:27
  but we are not trying to build a conscious being.
• 7:30
  Precisely. His stated goal is that AI must always
• 7:33
  remain a tool working in service of humans, trying
• 7:37
  to make it seem sentient. He calls that the wrong
• 7:40
  question. It's a very deliberate strategic stance.
• 7:43
  And why should that matter to, say, a business
• 7:46
  using these tools? Yeah. Or just an everyday
• 7:48
  user? Well, it comes down to trust and control,
• 7:51
  really. We already have this black box problem,
• 7:53
  right? We don't fully understand why these powerful
• 7:55
  AI models do what they do sometimes. Yeah, the
• 7:58
  unpredictability. Exactly. So if you then intentionally
• 8:02
  start adding simulated emotions or consciousness,
• 8:05
  well, control gets way harder, maybe even impossible.
• 8:09
  Microsoft seems to be betting that enterprise
• 8:11
  customers especially want predictability, reliability,
• 8:15
  and clear lines of control. They're positioning
• 8:17
  themselves as the responsible choice. Which contrasts
• 8:20
  pretty sharply with others like Meta or XAI who
• 8:23
  seem perfectly happy building emotionally intelligent
• 8:26
  chatbots. Right. Microsoft sees that path as
• 8:29
  potentially opening up a whole can of worms,
• 8:32
  unpredictable user attachments, legal issues.
• 8:35
  They seem determined to sidestep that whole mess.
• 8:38
  Makes sense from a business perspective. Avoid
• 8:40
  the headache if you can. That seems to be the
• 8:42
  strategic takeaway and you know this focus on
• 8:44
  responsibility and control isn't just talk it
• 8:46
  leads directly into our third story because this
• 8:48
  philosophical split it's actually showing up
• 8:51
  in their products and their policies now okay
• 8:53
  so the Microsoft OpenAI partnership is showing
• 8:55
  real cracks or at least divergences you could
• 8:58
  say that yeah Despite being OpenAI's biggest
• 9:01
  backer, Microsoft is pouring resources into becoming
• 9:04
  self -sufficient in AI. We've seen them rolling
• 9:07
  out their own internal models pretty quickly,
• 9:09
  MAI Voice One, MAI One Preview, and now this
• 9:12
  new MAI Image One that's focused on photorealistic
• 9:15
  images. They're building their own capabilities
• 9:17
  internally. Building their own moat, basically.
• 9:19
  Yeah. And that's affecting their big developer
• 9:21
  platform, too, GitHub. I saw something about
• 9:23
  a big overhaul plan there. Yeah, absolutely.
• 9:26
  They're planning a major revamp of GitHub. specifically
• 9:29
  to fight off these newer AI -native coding tools
• 9:31
  like Cursor and Claude Code, the vision seems
• 9:35
  to be making GitHub the absolute center of gravity
• 9:38
  for all of AI -powered software development.
• 9:40
  Center of gravity, okay. Satya Nadella, Microsoft
• 9:43
  CEO, who, by the way, just got a huge pay package
• 9:45
  tied to this AI boom, he apparently brought up
• 9:48
  Bill Gates' old frustration about apps being
• 9:50
  stuck in silos. The idea now is that AI should
• 9:53
  just erase those boundaries between your code,
• 9:55
  your documents, websites, everything. By weaving
• 9:58
  AI deeply into GitHub itself, they keep developers
• 10:02
  locked into the Microsoft ecosystem, no matter
• 10:04
  which specific AI model they might be using underneath.
• 10:08
  Ah, okay, so it's not just about having the best
• 10:10
  AI assistant, it's about owning the environment
• 10:12
  where developers work. That's classic platform
• 10:15
  lock -in AI addition. You got it. But where the
• 10:18
  split gets really public, really stark, is around
• 10:22
  content policies, specifically AI -generated
• 10:25
  erotica. Ah, right. This seems like a very clear
• 10:28
  line in the sand between them. Couldn't be clearer.
• 10:31
  Sleeman stated very publicly, Microsoft will
• 10:33
  not build services for a simulated erotica. He
• 10:36
  called it very dangerous and framed it as a conscious
• 10:38
  safety decision. Okay. But OpenAI just went the
• 10:41
  other way, didn't they? They announced they will
• 10:42
  allow age -gated erotic content on chat GPT.
• 10:46
  They did. And Sam Altman's justification was
• 10:48
  basically that OpenAI shouldn't be the elected
• 10:50
  moral police of the world. He framed it as being
• 10:52
  about user autonomy, avoiding censorship. So
• 10:55
  you've got OpenAI prioritizing expression, even
• 10:58
  in potentially tricky areas, versus Microsoft
• 11:01
  prioritizing enterprise safety, and frankly,
• 11:05
  probably regulatory compliance. That's the key
• 11:07
  business angle, I think. By saying no to AI erotica,
• 11:11
  Microsoft aligns itself neatly with emerging
• 11:13
  regulations, like that California SB 243 law
• 11:16
  about companion chat bots and miners. Right,
• 11:19
  the one banning sexually explicit content for
• 11:22
  kids. Exactly. So Microsoft is clearly choosing
• 11:24
  the path of the responsible, maybe more cautious
• 11:27
  enterprise provider. It sharply differentiates
• 11:29
  them from open AI, which seems willing to push
• 11:32
  boundaries more. It's really a risk management
• 11:34
  strategy disguised as a content policy. Wow,
• 11:37
  it's just fascinating to watch this happen in
• 11:38
  real time. The partnership that basically kickstarted
• 11:41
  this whole generative AI wave is now pulling
• 11:44
  invisibly different directions on cost, philosophy,
• 11:48
  even basic content rules. Yeah, the blueprints
• 11:50
  for the future of AI. They're definitely being
• 11:52
  drawn with two very different kinds of pens right
• 11:54
  now. Okay, before we wrap things up, we also
• 11:56
  want to give a shout out to our other sponsor
• 11:57
  today, Scony. Ah, Scony, great golf shoes. Really
• 12:01
  premium stuff. If you're looking for quality
• 12:03
  and style out on the course, check them out.
• 12:05
  And again, use that promo code SNARFL, S -N -A
• 12:08
  -R -F -U -L at checkout for 15 % off your entire
• 12:11
  purchase. Their website is scoea .com. Good shoes
• 12:15
  make a difference out there. They really do.
• 12:18
  Okay, so let's quickly recap the big takeaways
• 12:19
  for you today. All right. First, we saw OpenAI's
• 12:23
  almost unbelievable $1 .4 trillion infrastructure
• 12:26
  bet, led by Greg Brockman's intense focus on
• 12:29
  compute power and the significant energy and
• 12:31
  political friction that's creating. Then we looked
• 12:33
  at Microsoft's counter -move, led by Mustafa
• 12:36
  Suleiman, a more humanist, safety -focused approach
• 12:39
  that emphasizes keeping AI as a controllable
• 12:41
  tool, deliberately avoiding simulated sentience.
• 12:44
  And finally, we saw how that philosophical split
• 12:46
  is playing out strategically. Microsoft pushing
• 12:49
  for AI self -sufficiency, overhauling GitHub,
• 12:52
  and taking a firm public stance against things
• 12:54
  like AI erotica, a direct contrast to OpenAI's
• 12:57
  policy. Yeah, the whole AI landscape right now
• 12:59
  really feels like this tug of war between these
• 13:01
  two giant, powerful, but increasingly different
• 13:04
  visions for the future. So maybe here's a final
• 13:06
  thought for you to mull over. If Microsoft is
• 13:09
  staking its claim as the responsible one, by
• 13:12
  drawing lines around content like erotica, does
• 13:15
  that mean OpenAI is consciously trading mainstream
• 13:18
  trust, maybe even future regulatory goodwill,
• 13:21
  for more boundary pushing autonomy? Which one
• 13:24
  of those paths actually builds more lasting value
• 13:27
  in the long run? Something to think about. Definitely
• 13:30
  something to think about. We're here to help
• 13:32
  at Snarfl, so reach out on our website if you
• 13:33
  have any more questions. That's a wrap on today's
• 13:39
  top stories. Remember, tech doesn't have to be
• 13:42
  complicated, and we're here to help connect the
• 13:44
  dots. If you've got questions or want to learn
• 13:47
  more, visit us at snarful .com. Thanks for listening,
• 13:51
  and we'll catch you next time.