E147

E147: TED goes woke, Canada's Nazi blunder, AI adds vision, plus: who owns OpenAI?

2023-09-29 spoken.md · speaker-labeled ▶ watch ← E146 all episodes E148 →

ideas born

ideas moved

captures · 5 voices

dissenting

+55.5

conviction added

-130.0

decay · 115 silent

Every number here is replayed from score_events — the same ledger the pool ranks on. Decay is what the 115 ideas nobody mentioned gave up this week; it applies only when an episode is processed.

Tier crossings

conviction thresholds crossed by this episode — 65 / 45 / 15 · ideas born here show where they landed. why 65 isn't always green

▲ +5.3 🏛️ Nobody's incentives point at de-escalation — the war economy ratchets up watch → green threshold 63.2 → 68.5 still watch — green gate not met

▲ +10.3 📈 Legacy journalism is obsolete - the opinion economy takes the audience watch → green threshold 55.3 → 65.6 still watch — green gate not met

▲ +9.7 🤖 Only foundation models can disrupt search - and Google is too far ahead ember → watch 38.7 → 48.4

▲ +8.1 🤖 OpenAI is 2023's biggest winner — Microsoft deal inevitable ember → watch 39.9 → 48.1

▲ +9.0 🏛️ Affirmative-action ruling propagates into corporate DEI and ESG ember → watch 36.8 → 45.8

▲ +36.0 🤖 Apple's AR headset becomes the next compute platform dormant → ember 0.0 → 36.0

▼ -45.1 🤖 Generative AI is the next Silicon Valley bubble cycle green threshold → ember 75.1 → 30.0

▲ +28.7 🤖 ChatGPT plugins make OpenAI the next great computing platform dormant → ember 0.0 → 28.7

Kill dates that landed since E146

0 hit · 0 partial · 1 miss — windows that closed after 2023-09-22 and up to 2023-09-29, auto-scored against price data and never hand-set. verdict · R · α

idea	verdict	R	α	closed
⚡ Household energy resilience — every home becomes its own virtual power plant	MISS	-57.1%	-75.9	2023-09-23

Who moved the board

each voice's force on conviction this episode — supports and opposes, weighted exactly as the replay applied them · share = % of this episode's movement

Friedberg

3 captures · 47% of movement

+35.5 / -45.1 → net -9.5

Jason

3 captures · 21% of movement

+29.6 / -6.7 → net +22.8

Sacks

4 captures · 20% of movement

+35.0 → net +35.0

Coleman Hughes guest ×0.5

2 captures · 6% of movement

+5.0 / -6.0 → net -1.0

Chamath

1 capture · 5% of movement

+8.1 → net +8.1

What got argued (9 ideas)

ordered by how hard each idea moved · quotes are verbatim from the transcript, timestamps deep-link into the episode

▲ NVDA 🤖 Generative AI is the next Silicon Valley bubble cycle closed 24 CONTESTED ▼ -45.1 75.1 → 30.0

Friedberg reversal ×2 sentiment ▶ 1:14:34

I've been using a lot of different tools the last couple of months, and I'm kind of getting to the point that I feel that much of what's happening is underhyped rather than overhyped.

▲ AAPL 🤖 Apple's AR headset becomes the next compute platform closed 22 CONTESTED ▲ +36.0 0.0 → 36.0

Friedberg support ×2 explicit_prediction ▶ 1:21:02

So we will still need some sort of visual display, whether it's a screen or an eyeglass or something that shows us a bunch of information in a way.

Jason support ×2 sentiment ▶ 1:21:52

So eventually if Siri works and then you have those Apple goggles on, I think that that is going to be the eventual interface

▲ MSFT 🤖 ChatGPT plugins make OpenAI the next great computing platform closed 25 CONTESTED ▲ +28.7 0.0 → 28.7

Friedberg support ×3 explicit_prediction ▶ 1:14:34

So if mobile really is the dominant hardware platform that humans are using for computing today, LLMs and these sorts of tools can become the dominant operating system on that hardware.

Sacks support ×2 explicit_prediction ▶ 1:20:18

But imagine if the phone perfectly understood what you were saying, then you would just say, call me an Uber, order me food, whatever, and you would just instruct it.

▼ NYT 📈 Legacy journalism is obsolete - the opinion economy takes the audience closed 60 ▲ +10.3 55.3 → 65.6

Coleman Hughes support ×2 sentiment guest ×0.5 ▶ 25:33

Someone like Barry Weiss, who used to be at The New York Times, her point of view on it is, look, you just got to start your own institutions.

Sacks support ×2 sentiment ▶ 56:29

Now, the New York Times is framing this as a problem because it was a propaganda coup for Putin. And presumably it was. But I think it's a problem because it's a problem, not because of just the PR optics of it.

▲ GOOGL 🤖 Only foundation models can disrupt search - and Google is too far ahead closed 39 CONTESTED ▲ +9.7 38.7 → 48.4

Jason support ×2 sentiment ▶ 1:12:42

So this idea that Google is going to be displaced or they're moving slow, that might be antiquated information.

▼ ESGU 🏛️ Affirmative-action ruling propagates into corporate DEI and ESG closed 42 ▲ +9.0 36.8 → 45.8

Coleman Hughes oppose ×2 sentiment guest ×0.5 ▶ 36:40

I mean, in a word, no, because by that point, it's an intractable problem. I've talked to CEOs that ask this question to me over and over again.

Sacks support ×2 explicit_prediction ▶ 42:31

Remember what Brian Armstrong did at Coinbase? He basically just said, listen, we have a mission here. It's around crypto. We're going to focus 100% on this mission.

▲ MSFT 🤖 OpenAI is 2023's biggest winner — Microsoft deal inevitable closed 32 CONTESTED ▲ +8.1 39.9 → 48.1

Chamath support ×2 explicit_prediction ▶ 1:06:23

Sam is the closest thing that we have to an emergent mogul in tech. And the reason is because if everything sits on this substrate, you're going to need to get a license.

▼ GOOGL 🤖 Natural-language chat disrupts Google's search box closed 0 CONTESTED ▼ -6.7 6.7 → 0.0

Jason oppose ×2 sentiment ▶ 1:12:42

So this idea that Google is going to be displaced or they're moving slow, that might be antiquated information.

▲ ITA 🏛️ Nobody's incentives point at de-escalation — the war economy ratchets up closed 50 CONTESTED ▲ +5.3 63.2 → 68.5

Sacks support ×2 sentiment ▶ 58:22

From 2015 to 2021, we could have gone along with efforts under the Minsk Accords to resolve this conflict in the Donbass peacefully, but we never did that. We never gave it any support.

Episode digest

written during extraction and stored in data/extractions/ep147.json — the auditable source of truth, including everything market-adjacent that did not earn a capture

Guest Coleman Hughes anchors 44 minutes on TED suppressing his colorblindness talk; the market-relevant read is institutional capture — Sacks prescribes the Brian Armstrong/Coinbase mission-litmus-test as the remedy while Coleman insists large captured institutions are past the point of no return, and both sides trash legacy media (NYT framing, Barry Weiss's Free Press) in the process. The Canada/Nazi-ovation and Ukraine segment is politics with one market-adjacent thread: Sacks argues US policy systematically declined available off-ramps (Minsk) and made common cause with ultranationalist groups, reinforcing his no-constituency-for-de-escalation thesis. The AI block is the real alpha: off GPT-4V multimodal and Bard's Google-services integration, Friedberg makes his biggest AI call yet — LLMs as a kernel/OS, apps collapsing into in-stream plugins, 'a wholesale rewriting of computing' — and explicitly flips his own December-2022 bubble framing to 'underhyped rather than overhyped' (logged as a reversal; worth a spot-check at 1:14:34). Jason declares the Google-gets-displaced narrative 'antiquated information'; Sacks and Jason both land on voice-plus-glasses as the eventual interface. Two things captured only here in prose: Sacks's informed-speculation thesis that the capped-profit structure means Altman owns 100% of OpenAI's residual value through the nonprofit foundation (with Chamath's corollary that late buyers at the $80-90B secondary earn debt-like ~8% returns) — no clean listed expression, so no idea coined — and the FT report that OpenAI is raising ~$1B with Jony Ive and Masayoshi Son to build 'the iPhone of AI'.