May 16, 2024 · 1h 1m · bg2-pod
Ep9. GPT-4o, Astra, Multi Modal, China Tariffs, Tech Earnings | BG2 with Bill Gurley & Brad Gerstner · Bg2 Pod
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of BG², venture capitalists Brad Gerstner and Bill Gurley analyze major breakthroughs in multimodal AI models like OpenAI's GPT-4o and Google Astra, evaluating their impact on software business models, consumer interfaces, and platform ecosystem wars. They also critique US-China tariff policies and offer a macroeconomic market update on software earnings and interest rates.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Brad and Bill hold 98.1% of the talking time here. How this is scored →
Brad and Bill as informed peer 7.0 Guest teaching 0.0 Guest disagreement 0.0 Brad and Bill pushing back 0.5
speaking balance: gold is Brad and Bill, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 35:19 Bill bounds BPO disruption timeline
In a collegial co-host dynamic, Bill mildly pushes back on Brad's thesis by asserting BPO disruption warning levels are currently at a 6 rather than a 10.
Hardest push from Brad and Bill ▶ 12:43 Bill clarifies benchmark focusBill clarifies Brad's chart interpretation to point out that standard LLM reasoning metrics miss the mark on measuring voice interpretation quality.
Biggest teaching moment ▶ 38:27 Bill compares AI funding to NIH and NSF budgetsBill provides detailed context on federal research budgets, noting that a $32B AI appropriation would 4x the National Science Foundation budget.
Brad and Bill hold their own ▶ 56:50 Brad presents proprietary software earnings guidance dataBrad demonstrates deep public market expertise by citing internal tracking showing 54% of software companies lowered forward guidance despite beating adjusted consensus.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Brad and Bill as informed peer | Guest teaching | Guest disagreement | Brad and Bill pushing back | Why |
|---|---|---|---|---|---|---|
| BG² Podcast Animated Title Sequence | 5 | 0 | 0 | 0 | Co-hosts Brad and Bill casually discuss using ChatGPT as an interactive tutor for studying economics. Bill references Neal Stephenson's The Diamond Age and Sal Khan's AI education initiatives. | |
| OpenAI's GPT-4o, Google Astra, and 'Her' Realized | 4 | 0 | 0 | 0 | Brad plays audio clips comparing OpenAI GPT-4o, Google Astra, and the film Her to illustrate end-to-end multimodal voice assistants. | |
| Evaluating Voice Interfaces and Consumer Latency | 7 | 0 | 0 | 1 | Bill shares his hands-on testing of GPT-4o's latency and voice inflection, contrasting it with UI needs for visual information. Brad agrees and discusses multi-modal blended interfaces. | |
| AI Benchmarks, Inference Costs, and Multimodal Architecture | 8 | 0 | 0 | 2 | Brad examines SemiAnalysis benchmark charts comparing HumanEval scores and inference pricing curves. Bill clarifies that the industry needs dedicated voice-interpretation benchmarks rather than just standard LLM benchmarks. | |
| The Big Tech Platform War: Apple, Google, and OpenAI | 8 | 0 | 0 | 1 | Brad and Bill explore platform dynamics between Apple, OpenAI, and Google, comparing potential ecosystem dependencies to Microsoft and AOL platform struggles in prior eras. | |
| AI Business Models and Threat to 'GPT Wrappers' | 7 | 0 | 0 | 0 | Brad analyzes enterprise model pricing cliffs and Sam Altman's warning to GPT wrapper startups. Bill recounts using GPT-4o for ad-hoc Spanish language tutoring to illustrate the threat to single-feature apps like Duolingo. | |
| Competitive Advantage Period and BPO Disruption | 7 | 0 | 0 | 1 | Bill references Michael Mauboussin's competitive advantage period (CAP) theory to explain tech multiple compression, while slightly tempering Brad's alarm on immediate BPO disruption. | |
| Washington's AI Policy: Senate Roadmap and $32B Spending | 8 | 0 | 0 | 0 | Bill breaks down Chuck Schumer's Senate AI roadmap, heavily criticizing the proposed $32B annual public spending allocation in an already well-funded venture market. | |
| US-China Tariffs, Deglobalization, and Free Trade | 8 | 0 | 0 | 0 | Both hosts strongly oppose newly announced tariffs on Chinese EVs and tech, invoking classic free trade economic principles, Matt Ridley's books, and firsthand observations of Chinese manufacturing speed. | |
| Macro Market Check: CPI Print, Rates, and Software Earnings | 8 | 0 | 0 | 0 | Brad details macroeconomic indicators including dovish CPI data, 10-year Treasury yields, and deteriorating guidance across 54% of software companies to justify reducing portfolio risk. |
Statements from this episode (23)
Gurley: AI Models Succeed as Tutors Due to Infinite Patience
“One thing that seems blaringly obvious is this teacher element, and Saul Kahn, who's present in some of the demos this week has been leaning into this heavily. He's someone we should definitely pay attention to but they look real, and we talked about the model…”
Gerstner: Google's Astra Demo Felt Pre-Recorded and Clunky vs. GPT-4o
“Google yesterday, you know, at their IO event showed off Astra, their version of Her, albeit a little Recorded and maybe a little bit more clunky.”
Gurley: Pure voice interfaces are too tedious for comparative visual tasks
“If you want to, you know, see four different hotel choices in Taiwan, I don't think you want that read to you.
I think that will be tedious versus looking at it on the screen.”
Gerstner: AI agents will navigate existing apps rather than replace them
“So we're not going to throw out all the apps that we currently have like DoorDash and Booking.com, but where your AI will be smart enough to navigate those and then maybe take screenshots and push them to you or things like that.”
Gurley: OpenAI's voice demo appears way better than Apple's Siri
“It feels like from the demo, this is way better than Siri.”
Gerstner: GPT-4o Dramatically Lowers Inference Pricing Over Benchmark Gains
“And we plotted chat GPT for Omni on this chart, and you can see how it barely improved in terms of human level, the human eval score, but it dramatically improved in terms of pricing, you know, in terms of inference pricing.”
Gurley: Tech Giants Are Gearing Up for a Consumer AI 'Battle Royale'
“We're just gonna have six companies all running at the same prize and it's interesting, we end up talking about AI every week, but I don't know what there is else to talk about, because that's all anyone's talking about, and so it, we're gearing up for a battl…”
Gerstner: Apple cannot outsource building its core AI assistant
“I think a decision to sell the generative AI search button you know, to open AI and the way that they're doing with Google search is a far cry from outsourcing, right? Their mission to build their own version of her, which I don't think that Apple can ever out…”
Gerstner: Serving AI models is far lower margin than traditional software
“In fact, I was talking to a large data company this week who's in the business of serving models, and they were lamenting the fact that the pricing umbrella was set by open AI and there is no price, right? So that it's a much lower margin business than the tra…”
Gerstner: Frontier AI Labs Must Win Consumers to Justify Billions Invested
“So if you're going to be Anthropic or OpenAI or Mistral or anybody else, it's hard to see how you build a business, you know, with payback on the tens of billions that you're going to be required to invest unless you're in the consumer game.”
Gerstner: ChatGPT is the only standalone consumer AI brand to break out
“And the only brand right now that's really broken out in the consumer game is ChatGPT, other than the brands that already have big consumer distribution.”
Gurley: Startups design with frontier AI models but deploy with cheaper versions
“When I talk to our entrepreneurs that are using these models, they might design with the cutting edge model, but they all back off to the affordable models on implementation, inference, and runtime. They all do.”
Gerstner: Duolingo fell 4-5% following OpenAI's GPT-4o demo
“I think the day that the demo showed, Duolingo stock was down, you know, four or five percent”
Gerstner: Altimeter's list of potentially disrupted companies is longest ever
“As an investment firm, as you know, that on the public side, you know, we look at who's going to be disrupted. I think it, you know, this is the list of companies that we have that are potentially disrupted. Is as long as it's ever been, you know, we've, we're…”
Gerstner: Multimodal AI voice and video will disrupt BPO call centers
“I'll tell you a whole nother category of companies, Bill that have moved into that bucket and it's all these BPO companies, business process, outsourcing businesses, right? So, you know, because remember, it's not just about language now, right? BPO companies …”
Gurley: BPO AI disruption warning level is a 6, not a 10
“I think with language and code, this is already, you know 10 out of 10 high alert. I think on business process automation, we think it's coming. I don't think it's at 10. You know, maybe the warning signals at six. Like, it hasn't been proven it's going to run…”
Gurley: Odds of effective early AI input regulation are near zero
“I think the odds that, that any government in any country could get in this early and start messing around with inputs and be effective is near zero.”
Gerstner: US will not pass heavy AI regulation or major funding
“I don't think we're going to get a lot of regulation. I would be surprised if we see a lot of incremental dollars go into this.”
Gurley: Deglobalization is highly inflationary and harms corporate competitiveness
“You know, de-globalization, highly inflationary, and we'll make our companies less competitive, not more.”
Gurley: Automotive innovation is most alive in China's car industry
“These companies are creating more competitive products, not just on AV, not just on landed price point, but even on features consumer care about. Much bigger screens, more interesting features you've never seen before. It's abundantly clear that innovation is …”
Gurley: Protective tariffs will weaken US automakers, not strengthen them
“And pulling up the, these tariffs to help protect our companies won't make them stronger. It'll make them weaker by not exposing them to the reality on the field.”