🤖 Enterprises tune and self-host open-source models instead of paying for proprietary APIs
Friedberg's counter to OpenAI's $3.4B run rate: open Llama instances already deliver 85-90% of what an enterprise needs, so once companies learn how to fine-tune and wire a model into their own internal data, they build their own LLM tooling rather than pay per seat or per token for ChatGPT. He explicitly frames it as a couple of years out and floats 'maybe OpenAI is AOL' — proprietary model revenue is a transitional artifact of enterprise inexperience, and the value accrues to whoever supplies the open weights and the inference.
0
CONVICTION
peaked 71.7 GREEN
EMBER
band
▲ LONG META
expression · bullish tech
PARTIAL
outcome · R +13.1% · α -26.9%
2026-06-14
window closed
⚠ CONFLICTING IDEAS ON THE SAME TICKER
— the board is arguing with itself; net it before trading (ticker view →)
also touching these tickers, same direction: AI compute capex supercycle has 12-24 months of runway 87 (NVDA) · Nvidia + Hugging Face makes NVDA the open-source AI champion 58 (NVDA) · Meta becomes America's open-source AI champion 27 (META) · Independent frontier-model labs lose to hyperscaler capital 14 (META) · Meta's AI assistant takes ten points of search share from Google 7 (META)
⚖ Why this verdict
fully deterministic — evaluate.py replays this from daily closes; nothing below is editable or hand-set
- Window: 2024-06-14 → 2026-06-14 — first mention + 24-month horizon, then the window locks.
- The call: ▲ LONG META (primary play). META rose 13.1% over the window → direction-adjusted R = +13.1% (the call made money).
- Benchmark: holding SPY over the same window returned +40.0% → α = +13.1 − (+40.0) = -26.9% — what this call made or lost against just owning the index. This is the number the verdict uses.
- Rule fired:
· HIT — R ≥ +10% AND α ≥ +5▶ PARTIAL — R ≥ +5% OR α ≥ 0· MISS — everything else
- Credit: supporters of a PARTIAL earn 0.5 each, opposers the inverse — this feeds the scoreboard weights. supported: Aaron Levie, Chamath, Friedberg, Jason | proxy-sensitive: NVDA→HIT (+55.8%)
Conviction timeline
bands: green ≥ 65 · watch ≥ 45 · ember ≥ 15
Plays vs SPY · % since first mention (2024-06-14)
Plays
| expression | symbol | kind | relevance | rationale |
|---|---|---|---|---|
| ▲ LONG | META | stock | PRIMARY | owns Llama, the open-weight standard he names as the substitute for the OpenAI platform |
| ▲ LONG | NVDA | stock | adjacent | self-hosted enterprise inference still has to run on accelerated compute |
Mention log
Friedberg
tech w=1.08 · n=65
· E183 (2024-06-14) · explicit_prediction · strength 3
▶ 1:02:26
SUPPORT
· horizon 24mo
“I can see like a couple years from now, all these big enterprises are going to figure the same thing out, and then you're not necessarily going to need to pay for the ChatGPT stuff if there's an internal tool and an internal LLM that you can run.”
AA
Aaron Levie
(regular guest ×1.0)
tech w=1.33 · n=6
· E208 (2024-12-20) · explicit_prediction · strength 2
▶ 1:06:06
SUPPORT
“If at any moment you know that Zuck will basically provide an open source model that is at kind of best in class benchmarks and at the frontier, then there is a limit on how much you can charge for the tokens of your hosted model”
Jason
tech w=1.05 · n=56
· E208 (2024-12-20) · sentiment · strength 2
▶ 1:07:01
SUPPORT
“Isn't it also the truth that major enterprises, Fortune 500s, 200s, 20 years ago weren't interested in open source, and now that's kind of their default. They want to buy into open source, because they don't want to be locked into a vendor.”
Chamath
tech w=0.90 · n=83
· E213 (2025-01-31) · explicit_prediction · strength 2
▶ 40:46
SUPPORT
“I think all the pressure right now, I think, is on meta. Because I think meta has to show up with the next iteration of llama that beats and exceeds Gemini, that exceeds R1. And I think that that is going to be crucial for us to have a counterweight to whatever China is going to put out after this.”
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
⏳ decay drained -72.2 over the idea's life — that's time passing, attributed to no one
Score events
| episode | kind | Δ | after | note |
|---|---|---|---|---|
| E183 2024-06-14 | init | +45.2 | 45.2 | E183 born by Friedberg (explicit_prediction x3) [w=1.08] |
| E184 2024-06-20 | decay | -0.9 | 44.3 | E184 silent |
| E185 2024-06-29 | decay | -0.9 | 43.4 | E185 silent |
| E186 2024-07-04 | decay | -0.9 | 42.6 | E186 silent |
| E187 2024-07-12 | decay | -0.9 | 41.7 | E187 silent |
| E188 2024-07-19 | decay | -0.8 | 40.9 | E188 silent |
| E189 2024-07-26 | decay | -0.8 | 40.1 | E189 silent |
| E190 2024-08-02 | decay | -0.8 | 39.3 | E190 silent |
| E191 2024-08-09 | decay | -0.8 | 38.5 | E191 silent |
| E192 2024-08-16 | decay | -0.8 | 37.7 | E192 silent |
| E193 2024-08-23 | decay | -0.8 | 37.0 | E193 silent |
| E194 2024-08-30 | decay | -0.7 | 36.2 | E194 silent |
| E195 2024-09-06 | decay | -0.7 | 35.5 | E195 silent |
| E196 2024-09-20 | decay | -0.7 | 34.8 | E196 silent |
| E197 2024-09-27 | decay | -0.7 | 34.1 | E197 silent |
| E198 2024-10-03 | decay | -0.7 | 33.4 | E198 silent |
| E199 2024-10-11 | decay | -0.7 | 32.7 | E199 silent |
| E200 2024-10-18 | decay | -0.7 | 32.1 | E200 silent |
| E201 2024-10-25 | decay | -0.6 | 31.4 | E201 silent |
| E202 2024-11-01 | decay | -0.6 | 30.8 | E202 silent |
| E203 2024-11-08 | decay | -0.6 | 30.2 | E203 silent |
| E204 2024-11-16 | decay | -0.6 | 29.6 | E204 silent |
| E205 2024-11-23 | decay | -0.6 | 29.0 | E205 silent |
| E206 2024-12-07 | decay | -0.6 | 28.4 | E206 silent |
| E207 2024-12-13 | decay | -0.6 | 27.9 | E207 silent |
| E208 2024-12-20 | reinforce | +21.6 | 49.5 | E208 Aaron Levie support x2 (new voice) [w=1.33] |
| E208 2024-12-20 | reinforce | +12.0 | 61.5 | E208 Jason support x2 (new voice) [w=1.05] |
| E209 2025-01-04 | decay | -1.2 | 60.2 | E209 silent |
| E210 2025-01-11 | decay | -1.2 | 59.0 | E210 silent |
| E211 2025-01-18 | decay | -1.2 | 57.9 | E211 silent |
| E212 2025-01-25 | reinforce | +6.7 | 64.5 | E212 Jason support x2 [w=1.05] |
| E213 2025-01-31 | reinforce | +7.2 | 71.7 | E213 Chamath support x2 (new voice) [w=0.90] |
| E214 2025-02-07 | decay | -1.4 | 70.3 | E214 silent |
| E215 2025-02-15 | decay | -1.4 | 68.9 | E215 silent |
| E216 2025-02-21 | decay | -1.4 | 67.5 | E216 silent |
| E217 2025-03-01 | decay | -1.4 | 66.2 | E217 silent |
| E218 2025-03-08 | decay | -1.3 | 64.8 | E218 silent |
| E219 2025-03-15 | decay | -1.3 | 63.5 | E219 silent |
| E220 2025-03-22 | decay | -1.3 | 62.3 | E220 silent |
| E221 2025-03-29 | decay | -1.2 | 61.0 | E221 silent |
| E222 2025-04-05 | decay | -1.2 | 59.8 | E222 silent |
| E223 2025-04-11 | decay | -1.2 | 58.6 | E223 silent |
| E224 2025-04-19 | decay | -1.2 | 57.4 | E224 silent |
| E225 2025-04-26 | decay | -1.1 | 56.3 | E225 silent |
| E226 2025-05-02 | decay | -1.1 | 55.2 | E226 silent |
| E227 2025-05-09 | decay | -1.1 | 54.1 | E227 silent |
| E228 2025-05-17 | decay | -1.1 | 53.0 | E228 silent |
| E229 2025-05-24 | decay | -1.1 | 51.9 | E229 silent |
| E231 2025-06-13 | decay | -1.0 | 50.9 | E231 silent |
| E232 2025-06-21 | decay | -1.0 | 49.9 | E232 silent |
| E233 2025-06-28 | decay | -1.0 | 48.9 | E233 silent |
| E234 2025-07-04 | decay | -1.0 | 47.9 | E234 silent |
| E235 2025-07-11 | decay | -1.0 | 46.9 | E235 silent |
| E236 2025-07-19 | decay | -0.9 | 46.0 | E236 silent |
| E237 2025-08-01 | decay | -0.9 | 45.1 | E237 silent |
| E238 2025-08-09 | decay | -0.9 | 44.2 | E238 silent |
| E239 2025-08-15 | decay | -0.9 | 43.3 | E239 silent |
| E240 2025-08-22 | decay | -0.9 | 42.4 | E240 silent |
| E241 2025-08-29 | decay | -0.8 | 41.6 | E241 silent |
| E242 2025-09-07 | decay | -0.8 | 40.7 | E242 silent |
| E243 2025-09-19 | decay | -0.8 | 39.9 | E243 silent |
| E244 2025-09-27 | decay | -0.8 | 39.1 | E244 silent |
| E245 2025-10-03 | decay | -0.8 | 38.3 | E245 silent |
| E246 2025-10-10 | decay | -0.8 | 37.6 | E246 silent |
| E247 2025-10-17 | decay | -0.8 | 36.8 | E247 silent |
| E248 2025-10-24 | decay | -0.7 | 36.1 | E248 silent |
| E249 2025-10-31 | decay | -0.7 | 35.4 | E249 silent |
| E250 2025-11-07 | decay | -0.7 | 34.7 | E250 silent |
| E251 2025-11-14 | decay | -0.7 | 34.0 | E251 silent |
| E252 2025-11-22 | decay | -0.7 | 33.3 | E252 silent |
| E253 2025-12-06 | decay | -0.7 | 32.6 | E253 silent |
| E254 2025-12-13 | decay | -0.7 | 32.0 | E254 silent |
| E255 2025-12-19 | decay | -0.6 | 31.3 | E255 silent |
| E256 2025-12-31 | decay | -0.6 | 30.7 | E256 silent |
| E257 2026-01-10 | decay | -0.6 | 30.1 | E257 silent |
| E258 2026-01-17 | decay | -0.6 | 29.5 | E258 silent |
| E259 2026-01-30 | decay | -0.6 | 28.9 | E259 silent |
| E260 2026-02-07 | decay | -0.6 | 28.3 | E260 silent |
| E261 2026-02-13 | decay | -0.6 | 27.8 | E261 silent |
| E262 2026-02-28 | decay | -0.6 | 27.2 | E262 silent |
| E263 2026-03-06 | decay | -0.5 | 26.7 | E263 silent |
| E264 2026-03-13 | decay | -0.5 | 26.1 | E264 silent |
| E265 2026-03-19 | decay | -0.5 | 25.6 | E265 silent |
| E266 2026-03-27 | decay | -0.5 | 25.1 | E266 silent |
| E267 2026-04-03 | decay | -0.5 | 24.6 | E267 silent |
| E268 2026-04-10 | decay | -0.5 | 24.1 | E268 silent |
| E269 2026-04-17 | decay | -0.5 | 23.6 | E269 silent |
| E270 2026-04-24 | decay | -0.5 | 23.1 | E270 silent |
| E271 2026-05-01 | decay | -0.5 | 22.7 | E271 silent |
| E272 2026-05-08 | decay | -0.5 | 22.2 | E272 silent |
| E273 2026-05-15 | decay | -0.4 | 21.8 | E273 silent |
| E274 2026-05-22 | decay | -0.4 | 21.3 | E274 silent |
| E275 2026-05-29 | decay | -0.4 | 20.9 | E275 silent |
| E276 2026-06-13 | decay | -0.4 | 20.5 | E276 silent |