Dec 17, 2023 · 35m · another-podcast
AI and Everything Else
0:00 / 0:00
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Technology analyst Benedict Evans and co-host Tony Cameron Brown dissect Evans's annual macro technology presentation, analyzing generative AI's explosive hype, conceptual frameworks, and enterprise adoption realities. They explore whether AI represents a standard computing platform shift or a radical paradigm change, emphasizing the importance of rigorous questions over premature predictions.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 85% of the talking time here. How this is scored →
The hosts as informed peer 8.5 Guest teaching 0.2 Guest disagreement 0.2 The hosts pushing back 0.3
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 15:21 Tony highlights EU regulatory speed as indicative of policy panic
Tony asserts that the EU moving swiftly on the AI Act proves authorities feel compelled to act despite lacking fundamental technical understanding.
Hardest push from the hosts ▶ 16:15 Benedict rejects direct AI regulation framing as wrong abstractionBenedict pushes back on the premise of regulating AI directly, comparing it to attempting to regulate spreadsheets because FTX conducted fraud with Excel.
Biggest teaching moment ▶ 21:44 Tony provides concrete industry examples of generative AI in film and salesTony adds practical frontline color on how video creators build full movie trailers from scratch and sales teams use LLMs daily, complementing Benedict's personal skepticism.
The host holds their own ▶ 31:45 Benedict's 2002 Mobile World Congress thought experimentBenedict masterfully illustrates the peril of early S-curve forecasting by showing that even accurate technological predictions in 2002 missed Apple and Google dominating the smartphone platform.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Introducing the 10th Annual Macro Tech Presentation on AI | 8 | 0 | 0 | 0 | Benedict introduces his 10th annual macro presentation, contextualizing how macro shifts moved from mobile eating the world to e-commerce and now AI. Tony acts as a collaborative co-host facilitating the presentation overview. | |
| Quantifying the AI Hype and Silicon Valley Zeitgeist | 8 | 0 | 0 | 0 | Benedict details specific industry metrics, such as Hacker News front-page data comparing the iPhone launch to AI, Nvidia GPU revenue surges, and cloud capex. Tony agrees and highlights the deck's opening impact. | |
| Framing AI: Standard Platform Shift Versus GUI-Level Paradigm | 9 | 0 | 1 | 0 | Benedict lays out his high-level three-tier framework: platform shift, GUI-level paradigm shift, and AGI. He illustrates the second case using Bill Gates' GUI comparison and a Singapore tax automation scenario. | |
| The AGI Thought Experiment and the Search for Analogies | 9 | 0 | 0 | 0 | Benedict explores AGI as a philosophical thought experiment rather than an engineering milestone, comparing single-purpose machines like calculators and washing machines to mammal brains and Apollo orbital mechanics. Tony listens and validates the analogies. | |
| Regulatory Responses and Evolution from Pattern Recognition to Generation | 9 | 1 | 1 | 3 | Tony notes the unprecedented speed of the EU AI Act. Benedict offers nuanced pushback, explaining that the draft pre-existed GenAI and that regulating AI directly is the wrong level of abstraction, before tracing ML history from 2013 ImageNet pattern recognition to generative models. | |
| Workplace Adoption Realities and the 'Infinite Interns' Analogy | 8 | 1 | 0 | 0 | Benedict cites McKinsey enterprise adoption statistics and introduces the 'infinite interns' analogy for narrow task execution. Tony complements this with frontline observations of creative filmmaking workflows and sales automation. | |
| Historical Precedents: From Routine Automation to Systemic Shifts | 8 | 0 | 0 | 0 | Benedict explains how technology shifts evolve from replicating existing workflows to fundamentally altering business architectures, citing SQL enabling just-in-time supply chains and smartphones scaling to five billion users. | |
| Market Structure: Foundation Model Monopoly Versus Ubiquitous Proliferation | 9 | 0 | 0 | 0 | In response to Tony's prompt about gatekeepers, Benedict presents the tension between an oligopoly of three massive foundational models versus the commoditized proliferation of millions of specialized models. | |
| Navigating the S-Curve: Historical Blindspots and UI Misconceptions | 9 | 0 | 0 | 0 | Benedict unpacks why having definite answers at the start of an S-curve is foolish, using the 2002 mobile ecosystem thought experiment to show how incumbent predictions like Nokia, Microsoft, Sun, and Adobe failed. | |
| Maturation of Technology and Sector-Specific Inquiry | 8 | 0 | 0 | 0 | Benedict and Tony conclude that when a technology platform matures, the pertinent analytical questions migrate from computer science to domain-specific industry analysis in sports media, retail, and automotive. |
Statements from this episode (14)
Evans: Generative AI Shift Is Faster and Broader Than 2013 ML Wave
“We are certainly having at least one clear major platform shift is now underway around how we think about what generative machine learning is, just as in a sense, at a kind of a bare minimum, we also had that 10 years ago, as we thought, what is this machine l…”
Evans: Nvidia Sold $15B in GPUs Last Quarter, Up From $2B-$3B
“NVIDIA sold whatever the number is fifteen billion dollars of GPUs last quarter up from, you know, two or three the same quarter last year.”
Evans: Google, AWS, and Microsoft Cloud Capex Will Near $100B in 2023
“Google AWS, Microsoft probably spent, will probably spend a hundred billion dollars on cloud capex this year.”
Evans: Replacing bespoke software with AI prompts requires AGI
“I mean, I think that only works if you have AGI, which is the third point.”
Evans: AGI is a thought experiment and concept, not a technology
“AGI, the problem we're talking about artificial general intelligence is it's a thought experiment, not a technology. Or it's a concept, not technology.”
Evans: AI field lacks theoretical mathematics to chart progress toward AGI
“We don't have any kind of equivalent mathematics or any kind of equivalent chart that we could draw that says, well, AI is there, and we're here, and LLMs are in this place, and this is how fast it's going, and if it grows at the current rate, it will hit huma…”
Evans: Regulating AI Is the Wrong Level of Abstraction
“I think regulating AI, quote unquote, is the wrong level of abstraction. To me, that's like saying, well, look, you know, FTX did all this fraud with spreadsheets, so clearly we need to regulate spreadsheets.”
Evans: Tech Still Lacks a Useful Mental Model for Generative AI
“We haven't got to the point that we've got a mental, even if you think it's a platform shift, we haven't got to a point that we've got like a kind of a useful mental model of what would, what is this? Whereas for the last wave of machine learning, we're now ki…”
Evans: Benedict Evans has no personal or professional task for ChatGPT
“I personally do not have a task where chat GPT is useful. There are other people who do. I don't.”
Evans: McKinsey finds 90% of managers tried AI, but few use it
“This is a study from McKinsey, which is, you know, how many people in big company management have at least tried this? And the answer is like, 70, 80, 90%, depending on industries. But people using it for work is sort of three, four, five, maybe 10%.”
Evans: Previous wave of machine learning effectively provides 'infinite interns'
“The analogy I used to use for the last wave of machine learning was to say, this gives you infinite interns. So you want to listen to every call coming into the call center and tell me when the customer's angry. You want to look at every credit card transactio…”
Evans: AI models will not become commoditized in 2024
“Clearly that's not gonna happen next year, because you had these models have, like, much more data, much more capital requirements in order to get them to work.”
Evans: Shein will do $40 billion in clothing sales this year
“Sheehan will do forty billion dollars in, in, in, in clothing sales this year.”
Evans: As tech matures, defining questions become domain-specific, not software-focused
“It's all about understanding how the car industry works. It's not really a software question. And that's what happens once technology has kind of matured, and it's kind of passed on the torch to the industries that it's overturned.”