Latest / Chief in Tech Deep Dive Shortcast / Chief in Tech S1E50 - Sarah Mehler : Reclaiming Leadership in the Era of Collective Intelligence

Transcript

Chief in Tech S1E50 - Sarah Mehler : Reclaiming Leadership in the Era of Collective Intelligence · Chief in Tech Deep Dive Shortcast

• 0:00
  So if you take a step back and, you know, just
• 0:02
  look at your day -to -day routine lately, you've
• 0:05
  probably noticed something. Oh, I'm sure they
• 0:07
  have. Right. Like, artificial intelligence has
• 0:09
  made your standard tasks just much, much faster.
• 0:12
  Yeah, unbelievably fast. I mean, you can summarize
• 0:14
  a 50 -page report in, what, a matter of seconds.
• 0:18
  You draft this difficult email before you've
• 0:20
  even had your first sip of coffee. Or you untangle
• 0:23
  that really messy broken spreadsheet without
• 0:25
  even breaking a sweat. Exactly. It feels like
• 0:27
  a superpower. But here is a question that might
• 0:31
  make you pause. Has all of that incredible speed
• 0:34
  actually made your problem solving any more imaginative?
• 0:38
  That is the million dollar question right there.
• 0:41
  Because, you know, have you invented something
• 0:42
  entirely new today, or did you just do the exact
• 0:45
  same things you always do, only faster? Yeah,
• 0:48
  it really is the ultimate modern paradox, isn't
• 0:51
  it? It really is. We have the most powerful analytical
• 0:54
  tool in human history sitting right on our desktops
• 0:57
  just waiting for a prompt. And for the most part,
• 1:00
  we use it to clear out our inboxes. Yes. We use
• 1:03
  it to clear our inboxes so we can go to another
• 1:05
  meeting. We are essentially using a supercomputer
• 1:09
  as like a slightly better broom. OK, let's unpack
• 1:12
  this because that tension right there is exactly
• 1:15
  what we are focusing on today. Definitely. We
• 1:17
  are exploring this fascinating session description
• 1:20
  we've got, and it's titled Reclaiming Leadership
• 1:23
  in the Era of Collective Intelligence. It's by
• 1:26
  Sarah Miller. Which is such a great title, by
• 1:28
  the way. It is. And our mission for this deep
• 1:31
  dive is to really figure out how to stop using
• 1:34
  AI merely as this, you know, superficial speed
• 1:37
  booster and instead start using it to fundamentally
• 1:40
  expand human capacity. We are looking at a roadmap
• 1:44
  to get out of the inbox and back into actual
• 1:46
  innovation. Which is precisely the space Sarah
• 1:49
  Mailer has spent her entire career navigating.
• 1:51
  Right, she's the founder and CEO of Left Field
• 1:53
  Labs. Yeah, a creative technology firm. For nearly
• 1:56
  two decades she's been running this. Wow. We
• 1:58
  are talking about someone who has spent years
• 2:01
  guiding some of the most massive tech giants
• 2:04
  on the planet. I mean, companies like Google,
• 2:06
  Amazon, Meta... Salesforce, Qualcomm. The heavy
• 2:08
  hitters. Exactly. And her entire focus, especially
• 2:12
  recently, has been moving these massive organizations
• 2:16
  away from simple automation and pushing them
• 2:19
  toward true augmentation. So looking at the session
• 2:22
  description, it highlights this massive problem
• 2:25
  right out of the gate. It points out that across
• 2:28
  virtually every single industry, the exact same
• 2:32
  story is unfolding. AI efficiency is somehow,
• 2:36
  paradoxically, killing human imagination. Like,
• 2:39
  we are getting significantly faster at our job,
• 2:42
  sure, but the problems we are choosing to solve
• 2:44
  are becoming, well, less ambitious. Yeah, they're
• 2:48
  shrinking. Right. So why is that happening? Why
• 2:50
  does a tool that gives us so much leverage suddenly
• 2:53
  make us think smaller? Well, it happens because
• 2:55
  of how we naturally define productivity in a
• 2:58
  corporate environment. Okay. How so? Think about
• 3:01
  it. When a new tool arrives, Our immediate instinct
• 3:04
  is to ask, how can this do my current job faster?
• 3:07
  Right. I just want to go home at five. Exactly.
• 3:09
  That is the textbook definition of automation.
• 3:12
  But Mailer's session argues that AI's true potential
• 3:15
  doesn't actually lie in its speed or its scale.
• 3:17
  OK. Its actual value is in its ability to expand
• 3:22
  what humans are capable of conceiving. So to
• 3:25
  fix this kind of shrinking ambition, she introduces
• 3:28
  this core philosophy she calls collective intelligence.
• 3:32
  Okay, collective intelligence. I have to say,
• 3:35
  that sounds a little bit like a sci -fi concept.
• 3:37
  It's a little bit, yeah. Like we're all plugging
• 3:39
  our brains into some giant mainframe or something.
• 3:41
  Right, it sounds super abstract, but her definition
• 3:44
  is actually incredibly grounded and practical.
• 3:47
  Okay, lay it on me. So collective intelligence
• 3:50
  isn't just taking an off -the -shelf AI model
• 3:52
  and forcefully plugging it into your existing
• 3:54
  workflow. Like just slapping a chat bot on it.
• 3:56
  Exactly. It is the intentional design of spaces
• 3:59
  and systems where human creativity and machine
• 4:02
  capability actively work together. Key word being
• 4:05
  intentional. The absolute key word, yes. You
• 4:07
  have to actively design the system so that the
• 4:09
  human and the machine amplify each other rather
• 4:12
  than the machine just doing the human's chores
• 4:15
  while the human, you know, checks out. Okay,
• 4:18
  so exponential games and amplifying human creativity.
• 4:21
  That all sounds great in a controlled environment.
• 4:23
  Or maybe on a stage at a conference. But I have
• 4:27
  to push back here. Let's hear it. Because if
• 4:30
  I am a middle manager staring down a Friday afternoon
• 4:33
  deadline. Yeah, let's be real. Right. I honestly
• 4:36
  do not want a co -pilot challenging my ideas
• 4:39
  or amplifying my creativity right then. You just
• 4:42
  want it done. Yes. I want the machine to do my
• 4:44
  chores so I can start my weekend. How does Mailer
• 4:47
  reconcile this lofty framework with the actual
• 4:51
  burnout and friction of modern corporate life?
• 4:54
  It's a fair point. But if we connect this to
• 4:57
  the bigger picture, the problem with purely chasing
• 5:00
  that Friday afternoon efficiency is that it has
• 5:03
  a hard ceiling. A hard ceiling. OK. Let's say
• 5:05
  you automate a process. You might save 20 % of
• 5:07
  your time, maybe even 50%. Which feels amazing
• 5:10
  at first. It does. But eventually you hit a limit.
• 5:12
  You were just doing the exact same baseline work
• 5:15
  faster. And soon, your competitors are doing
• 5:17
  it exactly that fast, too. Ah, so you lose the
• 5:20
  advantage pretty quickly. Exactly. Miller argues
• 5:23
  the true human machine choreography exponentiates
• 5:27
  outcomes. It doesn't just add a little bit of
• 5:30
  speed to your Friday. You know, it multiplies
• 5:33
  the possibilities of what the business can actually
• 5:35
  do. I see. When you design for collective intelligence,
• 5:39
  you create results that neither humans nor machines
• 5:42
  could ever achieve alone. Oh, wow. Yeah. You
• 5:45
  unlock entirely new categories of products, total
• 5:48
  new revenue streams, and completely different
• 5:50
  ways of operating. You solve problems that a
• 5:53
  faster spreadsheet could just never fix. Okay,
• 5:55
  so relying on efficiency just accelerates a race
• 5:58
  to the bottom where everyone is equally fast
• 6:00
  and equally average. Will said, yeah. But augmentation
• 6:03
  actually changes the nature of the race itself.
• 6:06
  That makes sense conceptually, but if realizing
• 6:10
  you are stuck on the automation treadmill is
• 6:12
  only half the battle. The question is how you
• 6:14
  map your escape. Right, the practical steps.
• 6:16
  Exactly. If I'm looking at my own daily habits
• 6:18
  right now, how do I know if I'm building collective
• 6:21
  intelligence or just building a faster treadmill?
• 6:23
  Well, that is exactly where Mieler introduces
• 6:26
  this diagnostic framework. She tells leaders
• 6:29
  to audit their organizations. And frankly, Individuals
• 6:33
  can audit their own habits, too, to look for
• 6:35
  what she calls extractive patterns. Extractive
• 6:38
  patterns. Let me jump in here because this concept
• 6:41
  really struck me when we were reviewing the sources.
• 6:43
  Oh, it's so important. It really is. An extractive
• 6:46
  pattern is essentially when you use the AI to
• 6:49
  drain human insight out of a process. Yes. Let's
• 6:52
  bring this right down to the listener's level,
• 6:54
  because it is incredibly easy to fall into this
• 6:57
  trap without realizing it. So easy. Say you are
• 7:00
  prepping for a major strategy meeting. If you
• 7:02
  tell the AI, hey, write my Q3 strategy document
• 7:06
  based on these three bullet points, and then
• 7:08
  you just take whatever it spits out, format it,
• 7:11
  and present it. Yeah, that is extractive. Completely
• 7:13
  extractive. You extracted a finished product,
• 7:15
  sure, but you completely outsourced your own
• 7:17
  cognitive process. The machine do the thinking.
• 7:20
  Exactly, which means the final product lacks
• 7:23
  your nuance, your specific industry experience,
• 7:26
  and your critical edge. That is the perfect example.
• 7:30
  And when you scale that extractive habit across
• 7:33
  an entire organization for, say, six months,
• 7:36
  the result is devastating. The Diagnosti Framework
• 7:40
  looks for organizational signals of this exact
• 7:43
  behavior. Like what kind of signals? Well, if
• 7:46
  a workflow consistently results in employees
• 7:48
  checking out creatively, or if teams are just
• 7:51
  you know, rubber -stamping AI -generated outputs
• 7:53
  without any rigorous debate. Just going, yeah,
• 7:56
  it looks good and moving on. Exactly. That is
• 7:58
  a glaring red flag. The machine is extracting
• 8:01
  productivity, but the humans spark. That ability
• 8:04
  to identify a unique insight that the AI's training
• 8:08
  data wouldn't naturally prioritize is completely
• 8:11
  bypassed. Right. Over time, the organization's
• 8:13
  ideas just regress to a perfectly average predictable
• 8:17
  mean. Because the AI is a probability engine.
• 8:19
  Right. It is literally designed to give you the
• 8:21
  most statistically likely answer, which is, by
• 8:24
  definition, average. Yep. So if extracting is
• 8:26
  letting the AI drive you to the middle of the
• 8:28
  road, what does the co -creative pattern actually
• 8:31
  look like in that same meeting prep scenario?
• 8:33
  So in a co -creative pattern, you use the AI
• 8:37
  to elevate your original thought, not replace
• 8:40
  it. OK. How so? You might draft your own messy,
• 8:44
  imperfect strategy for Q3. You drop it into the
• 8:47
  AI, and you prompt it by saying, critique this
• 8:51
  from the perspective of a highly skeptical chief
• 8:53
  financial officer. Oh, I like that. Right. Where
• 8:56
  are the holes in my logic, and what data am I
• 8:58
  missing? In that scenario, the AI isn't doing
• 9:01
  the work for you. It is forcing you to think
• 9:03
  harder. It's a sparring partner. Exactly. The
• 9:07
  organizational signals of co -creation are friction,
• 9:09
  iteration, and debate. People bring weirder,
• 9:12
  more complex ideas to the table precisely because
• 9:15
  they've used the AI to stress test those ideas
• 9:17
  in private before bringing them to the group.
• 9:20
  Here's where it gets really interesting. Because
• 9:22
  once you understand the difference between an
• 9:24
  extractive habit and a co -creative one, you
• 9:26
  realize you can't just cross your fingers and
• 9:28
  hope your team naturally figures out how to co
• 9:30
  -create. No, definitely not. You actually have
• 9:31
  to choreograph the entire process. Mailer's session
• 9:34
  emphasizes that you have to design the human
• 9:36
  -machine feedback loop with extreme precision.
• 9:39
  Yes, and she grounds this, in reality, by pointing
• 9:43
  to highly specific workflow diagrams from Left
• 9:46
  Field Lab's own collaborations. Like which ones?
• 9:49
  She specifically mentions projects like Best
• 9:51
  Phones Forever and Illuminated. These aren't
• 9:54
  just, you know, theoretical ideas. They are mapped
• 9:56
  out structural interactions. Think of it like
• 9:59
  a high stakes relay race. Okay, I like this analogy.
• 10:03
  In a relay race, you can have the two absolute
• 10:05
  fastest runners in the world on your team. But
• 10:08
  if they haven't explicitly practiced the Baton
• 10:10
  Pass, they will drop it and lose the race. Every
• 10:13
  single time. Right. The magic isn't just in both
• 10:15
  parties running at maximum capacity. The magic
• 10:18
  is entirely in the mechanics of the handoff.
• 10:21
  Where does the human stop? And where does the
• 10:23
  AI start? What's fascinating here is how these
• 10:26
  work -for diagrams actually execute that Baton
• 10:28
  Pass. Let's get into it. Let's look at how a
• 10:30
  project like the Best Phones Forever campaign.
• 10:32
  which is a known, highly successful tech marketing
• 10:35
  initiative, likely structures this choreography.
• 10:38
  The diagram defines exactly what the human is
• 10:41
  responsible for. In a creative campaign, humans
• 10:44
  must set the comedic tone, the emotional resonance,
• 10:48
  and the core strategic message. Things AI is
• 10:50
  terrible at. Exactly. Then, the AI's role is
• 10:54
  defined. So perhaps it processes years of engagement
• 10:57
  data to determine the optimal pacing of a video,
• 11:00
  or it rapidly generates dozens of visual storyboards
• 11:04
  based on the human's core joke. But the crucial
• 11:07
  part is that transition. If the AI just dumps
• 11:09
  a massive, unformatted spreadsheet of engagement
• 11:12
  data onto a creative director's desk. Yeah, that
• 11:14
  director's gonna drop the baton. Right. They
• 11:16
  can't do anything with that. It kills the creative
• 11:18
  flow. Precisely. Architecting the transition
• 11:21
  means building an interface or a process where
• 11:24
  the AI's output is immediately legible and actionable
• 11:27
  for the human. Well, legibility is huge. It translates
• 11:30
  raw data into a visual map or a set of clear
• 11:33
  options. Then the human curates, refines, and
• 11:36
  polishes those options, passing the feedback
• 11:38
  back to the AI to constrain its next generation.
• 11:42
  Or consider a data -heavy project like Illumined.
• 11:45
  If you are dealing with massive datasets regarding
• 11:48
  user behavior or technical performance, the AI
• 11:51
  is doing the heavy lifting of pattern recognition
• 11:54
  across millions of data points. Right. But the
• 11:57
  handoff requires the AI to surface those patterns
• 12:00
  with context. So human engineers can apply ethical
• 12:04
  guardrails or business logic that the AI simply
• 12:07
  cannot grasp. Because it doesn't have the real
• 12:09
  -world context. Exactly. When the choreography
• 12:12
  is tight, when everyone knows exactly what they
• 12:14
  are handing off and receiving, the creative momentum
• 12:17
  accelerates rather than stalling out. You require
• 12:20
  such clear boundaries. You don't want humans
• 12:22
  trying to do the AI's job of parsing millions
• 12:24
  of data points, and you certainly don't want
• 12:25
  the AI trying to do the human's job of defining
• 12:28
  emotional resonance. That's a disaster. You want
• 12:30
  them moving in sync. So let's say a listener
• 12:32
  takes this to heart. They do the audit. They
• 12:35
  identify and eliminate their extractive rubber
• 12:37
  stamping patterns. They design this perfectly
• 12:40
  choreographed co -creative system where the baton
• 12:44
  passes between human and machine are absolutely
• 12:46
  flawless. Sounds amazing. What do they aim this
• 12:49
  powerful new system at? Because if we build this
• 12:51
  incredibly sophisticated engine just to reply
• 12:54
  to customer service emails slightly faster, I
• 12:57
  mean, that feels like a tragedy. It is a tragedy.
• 12:59
  Yeah. And it leads directly to Mailer's final
• 13:01
  mandate, bringing ambition back to AI implantation.
• 13:05
  Bringing ambition back. I love that. The core
• 13:07
  argument of her session is that once you have
• 13:10
  this capacity, you must give yourself permission
• 13:13
  to pursue massive, ambitious moonshots. you have
• 13:17
  to look past the marginal efficiency games. And
• 13:19
  there's data to back this up, right? There is.
• 13:22
  She brings hard survey data to back this up,
• 13:24
  showing that organizations explicitly focused
• 13:26
  on innovation, the ones using AI to solve complex
• 13:29
  systemic problems, heavily outperform organizations
• 13:32
  that are just using it to trim their operational
• 13:34
  budgets. So what does this all mean for you,
• 13:37
  the listener? How do we justify taking our eyes
• 13:39
  off the daily grind to actually chase a moonshot?
• 13:42
  Yeah, especially when the pressure to just survive
• 13:44
  the week is so immense. Exactly. It's tough.
• 13:48
  We justify it by looking at the scale of the
• 13:51
  problems we could actually be solving if we elevated
• 13:54
  our gaze. And Miller doesn't just preach this.
• 13:58
  Her broader projects outside of traditional corporate
• 14:00
  tech serve as the ultimate proof of concept.
• 14:03
  What kind of projects? Well the source material
• 14:05
  notes her work as a vocal advocate for responsible
• 14:08
  innovation, specifically highlighting her involvement
• 14:11
  with the Pachamama Alliance and her support of
• 14:14
  regenerative climate initiatives. Regenerative
• 14:16
  climate initiatives. OK, that is a massive pivot
• 14:20
  from optimizing a marketing campaign or building
• 14:22
  a better spreadsheet. It is. And the specific
• 14:25
  initiative mentioned is something called the
• 14:27
  New Green Loop. This is an effort to spur global
• 14:30
  climate action using a choreographed combination
• 14:32
  of AI, blockchain and DMRV technologies. Let's
• 14:36
  break that down because those are heavy technical
• 14:38
  concepts. But the way they interact is actually
• 14:41
  the perfect example of collective intelligence.
• 14:43
  Definitely. So DMRV stands for for digital measurement,
• 14:46
  reporting, and verification. In the context of
• 14:49
  a climate initiative, say, reforesting a massive
• 14:52
  tract of land DMRV involves using sensors, drones,
• 14:56
  and satellite imagery to physically measure the
• 14:58
  real world environment. It captures the raw data
• 15:01
  on the ground. Exactly. And this is where the
• 15:04
  choreography comes in. You have these massive
• 15:06
  streams of physical data coming from the DMRV
• 15:08
  sensors. The AI steps in to analyze that data.
• 15:12
  It's predicting forest health, calculating the
• 15:15
  exact carbon capture rates and identifying ecological
• 15:18
  patterns that humans could just never process
• 15:20
  fast enough. There's too much data. Right. But
• 15:23
  to make that data useful for global climate action,
• 15:26
  it has to be trusted. And that is where the blockchain
• 15:28
  comes in. Ah, right. The blockchain acts as the
• 15:31
  immutable ledger. Exactly. It takes the AI's
• 15:34
  verified calculations and locks them into a transparent,
• 15:37
  unchangeable record. So if a corporation or a
• 15:40
  government is buying carbon credits or funding
• 15:42
  this regeneration, they don't have to guess if
• 15:44
  the impact is real. Nope, no guessing. The sensors
• 15:47
  measured it, the AI verified it, and the blockchain
• 15:49
  locked it. And where is the human in all of this?
• 15:52
  That's the question. The human is setting the
• 15:54
  strategic vision, building the vital relationships
• 15:57
  with local communities on the ground, and designing
• 16:00
  the economic incentives that make the new Green
• 16:02
  Loop function. Wow. No single technology could
• 16:05
  do this alone. And humans certainly couldn't
• 16:08
  do it manually. It requires the precise intentional
• 16:12
  handoff between sensors. algorithms, decentralized
• 16:15
  ledgers, and human diplomacy. It is taking three
• 16:19
  highly complex emerging technologies and choreographing
• 16:22
  them together to tackle literally the biggest
• 16:25
  existential threat we face. It proves that when
• 16:28
  you stop using AI to just automate the mundane
• 16:31
  and you start designing systems of collective
• 16:33
  intelligence, you suddenly have the bandwidth
• 16:35
  and the capability to tackle problems that previously
• 16:37
  seemed entirely unsolvable. That's the beauty
• 16:40
  of it. It isn't about making climate reporting
• 16:42
  5 % faster. It is about architecting an entirely
• 16:45
  new, verifiable system for planetary regeneration.
• 16:49
  That is the power of bringing ambition back.
• 16:52
  It challenges you to look at the tools on your
• 16:53
  desk and realize they are capable of so much
• 16:56
  more than what you are currently asking of them.
• 16:58
  It really does put our daily routines in perspective,
• 17:01
  doesn't it? Big time. Let's recap the journey
• 17:03
  we've been on today, because there is a lot for
• 17:06
  you to take back to your own desk. We started
• 17:08
  by identifying the trap of efficiency. The realization
• 17:12
  that using AI purely as an automation tool to
• 17:16
  speed up our chores is actually causing our ambition
• 17:18
  to shrink. Right. From there we explored the
• 17:21
  alternative. Sarah Mailer's concept of collective
• 17:24
  intelligence, which is the intentional design
• 17:27
  of spaces where human creativity and machine
• 17:29
  processing actively amplify each other. We also
• 17:32
  looked at how to measure our progress by auditing
• 17:35
  our habits for extractive patterns. The audit
• 17:37
  is key. It is. We have to be honest with ourselves
• 17:40
  about when we are just outsourcing our cognitive
• 17:43
  load to the machine and rubber -stamping its
• 17:45
  output versus when we are actively co -creating
• 17:48
  and using the AI to stress -test our best ideas.
• 17:52
  the mechanics of how to actually build those
• 17:54
  co -creative systems. It is all about the relay
• 17:57
  race. You have to clearly define where the human
• 17:59
  leads, where the machine processes, and most
• 18:02
  importantly, how to cleanly pass the baton back
• 18:04
  and forth so momentum isn't lost. The handoff.
• 18:07
  The handoff. And finally, we saw why this entire
• 18:10
  shift in mindset is so critical, so we can free
• 18:14
  up our human capacity to chase massive complex
• 18:17
  moonshots like the new Green Loop. We have to
• 18:20
  demand more from our tools, but we also have
• 18:23
  to demand more imagination from ourselves. Absolutely.
• 18:27
  The ultimate question you need to ask yourself
• 18:28
  the next time you open an AI prompt is, am I
• 18:31
  leading this dance or am I just letting the machine
• 18:34
  extract value from my time? This raises such
• 18:37
  an important question about our own agency. We
• 18:40
  live in an era of unprecedented information and
• 18:42
  capability, but knowledge is only powerful when
• 18:46
  it is actively applied. you have to be the one
• 18:49
  to design your system, or I guarantee you the
• 18:51
  system will design you. That is a powerful point.
• 18:54
  But before we wrap up this deep dive, I want
• 18:56
  to leave you, the listener, with one final lingering
• 18:58
  thought to mull over as you look at your own
• 19:00
  workflows. Okay. We spent this entire time talking
• 19:02
  about how to choreograph these perfect systems,
• 19:05
  how to iron out all the inefficiencies, and how
• 19:07
  to build these flawless feedback loops between
• 19:09
  human and machine. But consider this. If we successfully
• 19:13
  architect a future where AI perfects all of our
• 19:17
  processes, where it smooths out every single
• 19:19
  logical inconsistency, and instantly provides
• 19:22
  the most statistically correct answer to any
• 19:25
  problem. Which is the goal for a lot of people.
• 19:27
  Right. But if we do that, does the true value
• 19:30
  of human contribution actually become our capacity
• 19:33
  to make mistakes? Oh, wow. Think about it. Machines
• 19:36
  are literally programmed to avoid errors and
• 19:39
  anomalies. But historically speaking, those weird,
• 19:42
  completely irrational leaps of logic, the mistakes
• 19:45
  that fail 99 times out of 100, are almost always
• 19:49
  the exact things that spark our greatest, most
• 19:52
  unpredictable creative breakthroughs. That's
• 19:54
  a great point. So keep that in mind the next
• 19:56
  time the AI politely suggests that your new idea
• 19:58
  doesn't make logical sense.