◢ ALL-IN ALPHA research only · never trades · 2026-09-10

← pool · child of Models commoditize — proprietary data is the only AI moat

🤖 YouTube is the largest proprietary training corpus on earth - Google's insurmountable AI moat

Friedberg's sizing: Common Crawl, the backbone of GPT-3's training mix, is roughly 10 petabytes, while YouTube holds an estimated 2,000-3,000 petabytes growing 1-2 petabytes a day - about 300x bigger, and it is video, image, audio and text rather than scraped text alone. If model quality converges once everyone trains on the open internet, the only remaining edge is a corpus nobody can license, and Google's is the biggest that exists (plus Gmail and Docs as private-data reserves). Conditional on Google actually shipping, that makes it the most valuable asset in the world under the data-moat thesis.

0 CONVICTION
peaked 55.6 WATCH
WATCH
band
▲ LONG GOOGL
expression · bullish tech
PARTIAL
outcome · R +24.8% · α +3.4%
2025-02-09
window closed
⚠ CONFLICTING IDEAS ON THE SAME TICKER — the board is arguing with itself; net it before trading (ticker view →)
GOOGL · Blue cities ban or license-cap robotaxis SHORT EMBER 28.2
GOOGL · Tort floodgates open on social media SHORT DORMANT 14.6
GOOGL · ChatGPT takes Google search share and the market starts pricing the decay SHORT DORMANT 9.3

also touching these tickers, same direction: Google wins consumer AI via distribution 61 (GOOGL) · Anthropic IPO (~Oct 2026) is attractively priced at $1.5-2T 54 (GOOGL) · Personal AI agents: Google's data moat wins the form factor 29 (GOOGL) · Frontier AI labs keep durable premium pricing (not commoditized) 27 (GOOGL) · Own AWS, Azure and GCP and nothing else — the clouds capture the AI application dollars 21 (GOOGL) · Independent frontier-model labs lose to hyperscaler capital 14 (GOOGL) · Willow's error-correction scaling puts encryption on a two-to-five-year clock 9 (GOOGL) · Meta's AI assistant takes ten points of search share from Google 7 (GOOGL)

⚖ Why this verdict

fully deterministic — evaluate.py replays this from daily closes; nothing below is editable or hand-set

  1. Window: 2024-02-09 → 2025-02-09 — first mention + 12-month horizon, then the window locks.
  2. The call: ▲ LONG GOOGL (primary play). GOOGL rose 24.8% over the window → direction-adjusted R = +24.8% (the call made money).
  3. Benchmark: holding SPY over the same window returned +21.4% → α = +24.8 − (+21.4) = +3.4% — what this call made or lost against just owning the index. This is the number the verdict uses.
  4. Rule fired:
    · HIT — R ≥ +10% AND α ≥ +5
    ▶ PARTIAL — R ≥ +5% OR α ≥ 0
    · MISS — everything else
  5. Credit: supporters of a PARTIAL earn 0.5 each, opposers the inverse — this feeds the scoreboard weights. supported: Aaron Levie, Friedberg, Jason

Conviction timeline

bands: green ≥ 65 · watch ≥ 45 · ember ≥ 15

Plays vs SPY · % since first mention (2024-02-09)

Plays

expressionsymbolkindrelevancerationale
▲ LONG GOOGLstock PRIMARY owns YouTube, Gmail and Workspace - the corpora the argument is entirely about

Mention log

Friedberg
Friedberg tech w=1.08 · n=65 · E165 (2024-02-09) · explicit_prediction · strength 3 ▶ 1:04:26 SUPPORT
“which makes YouTube data repository 300 times larger than Common Crawl, which makes it bigger than anything else that anyone else has... if Google gets its act together and leverages the data repository at YouTube, it is an insurmountable moat that will only continue to extend because the quality of the YouTube experience and the network effects continue to accumulate for them. So I think it's the most valuable asset in the world today based on this thesis that AI value is going to accrue to the data owner.”
Jason
Jason tech w=1.05 · n=56 · E165 (2024-02-09) · sentiment · strength 1 ▶ 1:09:42 SUPPORT
“What Google also has is Google Docs and Gmail, things that people say privately. So they have another data resource there that they can tap”
Friedberg
Friedberg tech w=1.08 · n=65 · E170 (2024-03-15) · explicit_prediction · strength 2 ▶ 26:12 REVERSAL
“I heard from someone at Google that under the terms of service, Google's not allowed to train on YouTube data. ... I think there would be this really ironic handicap that Google's done to itself, where anyone else can access and download and watch YouTube data to train models, but Google cannot.”
Friedberg
Friedberg tech w=1.08 · n=65 · E208 (2024-12-20) · explicit_prediction · strength 3 ▶ 1:29:25 SUPPORT
“By some estimates, there's a thousand exabytes of video data on the Internet.”
AA
Aaron Levie (regular guest ×1.0) tech w=1.33 · n=6 · E208 (2024-12-20) · sentiment · strength 2 ▶ 1:31:59 SUPPORT
“which is another advantage because YouTube has obviously an incredible amount of screen sharing data and videos”
Friedberg
Friedberg tech w=1.08 · n=65 · E213 (2025-01-31) · explicit_prediction · strength 2 ▶ 1:08:13 SUPPORT
“So video, I think you can recognize that Google's YouTube content library is probably 100 to 200 times larger than the rest of the internet combined.”

Who built this conviction

each voice's total force on the score — supports and opposes from every mention, weighted exactly as the replay applied them · share = % of all mention-driven movement

Friedberg
Friedberg w=1.08
4 scoring events · 75% of moves
+73.7 / -28.3 → net +45.3
AA
Aaron Levie w=1.33 regular guest ×1.0
1 scoring event · 17% of moves
+23.2 → net +23.2
Jason
Jason w=1.05
1 scoring event · 8% of moves
+10.4 → net +10.4

⏳ decay drained -32.0 over the idea's life — that's time passing, attributed to no one

Score events

episodekindΔafternote
E165 2024-02-09 init +45.2 45.2 E165 born by Friedberg (explicit_prediction x3) [w=1.08]
E165 2024-02-09 reinforce +10.4 55.6 E165 Jason support x1 (new voice) [w=1.05]
E166 2024-02-16 decay -2.2 53.4 E166 silent
E167 2024-02-23 decay -2.1 51.3 E167 silent
E168 2024-03-01 decay -2.1 49.2 E168 silent
E169 2024-03-08 decay -2.0 47.2 E169 silent
E170 2024-03-15 reversal -28.3 18.9 E170 Friedberg flips
E171 2024-03-22 decay -0.8 18.1 E171 silent
E172 2024-03-29 decay -0.7 17.4 E172 silent
E173 2024-04-05 decay -0.7 16.7 E173 silent
E174 2024-04-12 decay -0.7 16.0 E174 silent
E175 2024-04-19 decay -0.6 15.4 E175 silent
E176 2024-04-26 decay -0.6 14.8 E176 silent
E177 2024-05-03 decay -0.6 14.2 E177 silent
E178 2024-05-10 decay -0.6 13.6 E178 silent
E179 2024-05-17 decay -0.5 13.1 E179 silent
E180 2024-05-24 decay -0.5 12.6 E180 silent
E181 2024-05-31 decay -0.5 12.1 E181 silent
E182 2024-06-07 decay -0.5 11.6 E182 silent
E183 2024-06-14 decay -0.5 11.1 E183 silent
E184 2024-06-20 decay -0.4 10.7 E184 silent
E185 2024-06-29 decay -0.4 10.2 E185 silent
E186 2024-07-04 decay -0.4 9.8 E186 silent
E187 2024-07-12 decay -0.4 9.4 E187 silent
E188 2024-07-19 decay -0.4 9.1 E188 silent
E189 2024-07-26 decay -0.4 8.7 E189 silent
E190 2024-08-02 decay -0.3 8.4 E190 silent
E191 2024-08-09 decay -0.3 8.0 E191 silent
E192 2024-08-16 decay -0.3 7.7 E192 silent
E193 2024-08-23 decay -0.3 7.4 E193 silent
E194 2024-08-30 decay -0.3 7.1 E194 silent
E195 2024-09-06 decay -0.3 6.8 E195 silent
E196 2024-09-20 decay -0.3 6.5 E196 silent
E197 2024-09-27 decay -0.3 6.3 E197 silent
E198 2024-10-03 decay -0.3 6.0 E198 silent
E199 2024-10-11 decay -0.2 5.8 E199 silent
E200 2024-10-18 decay -0.2 5.6 E200 silent
E201 2024-10-25 decay -0.2 5.3 E201 silent
E202 2024-11-01 decay -0.2 5.1 E202 silent
E203 2024-11-08 decay -0.2 4.9 E203 silent
E204 2024-11-16 decay -0.2 4.7 E204 silent
E205 2024-11-23 decay -0.2 4.5 E205 silent
E206 2024-12-07 decay -0.2 4.3 E206 silent
E207 2024-12-13 decay -0.2 4.2 E207 silent
E208 2024-12-20 reinforce +18.6 22.7 E208 Friedberg support x3 (flipped from reversal) [w=1.08]
E208 2024-12-20 reinforce +23.2 45.9 E208 Aaron Levie support x2 (new voice) [w=1.33]
E209 2025-01-04 decay -1.8 44.1 E209 silent
E210 2025-01-11 decay -1.8 42.3 E210 silent
E211 2025-01-18 decay -1.7 40.6 E211 silent
E212 2025-01-25 decay -1.6 39.0 E212 silent
E213 2025-01-31 reinforce +9.9 48.9 E213 Friedberg support x2 [w=1.08]
E214 2025-02-07 decay -2.0 46.9 E214 silent