🤖 AI drug discovery hasn't lifted pharma hit rates — the M&A bid is falling
Chamath's contrarian read off the BioNTech/InstaDeep deal: exit values for ML-in-pharma companies have gone DOWN even as the technology got far better (Flatiron Health sold for $1.9B in 2018 and is still the high-water mark), because none of it has been shown to improve the design-space guess that actually drives pharma's batting average. Every pharma gets the capability as an adjunct and pays less for it, so the listed AI-drug-discovery cohort is overvalued; the only model that captures real value is giving the tools away for a back-end royalty (Royalty Pharma is his template).
0
CONVICTION
peaked 38.0 EMBER
DORMANT
band · contested
▼ SHORT RXRX
expression · bearish tech
MISS
outcome · R -11.3% · α -35.0%
2024-01-20
window closed
⚖ Why this verdict
fully deterministic — evaluate.py replays this from daily closes; nothing below is editable or hand-set
- Window: 2023-01-20 → 2024-01-20 — first mention + 12-month horizon, then the window locks.
- The call: ▼ SHORT RXRX (primary play). RXRX rose 11.3% over the window → direction-adjusted R = -11.3% (the call lost money).
- Benchmark: holding SPY over the same window returned +23.7%
→ α = -11.3 − (+23.7) = -35.0%
— what this call made or lost against just owning the index. This is the number the verdict uses.
Stock-selection read: +12.4% — did RXRX move the predicted way relative to the market. The two only differ on a SHORT: this call was right about direction relative to the index but still cost money versus holding it.
- Rule fired:
· HIT — R ≥ +10% AND α ≥ +5· PARTIAL — R ≥ +5% OR α ≥ 0▶ MISS — everything else
- Credit: supporters of a MISS earn 0 each, opposers the inverse — this feeds the scoreboard weights. supported: Chamath | opposed: Friedberg
Conviction timeline
bands: green ≥ 65 · watch ≥ 45 · ember ≥ 15
Plays vs SPY · % since first mention (2023-01-20)
Plays
| expression | symbol | kind | relevance | rationale |
|---|---|---|---|---|
| ▼ SHORT | RXRX | stock | PRIMARY | Recursion is the listed AI-drug-discovery pure play — the cleanest expression of the cohort whose acquisition and equity value he says keeps compressing |
| ▼ SHORT | SDGR | stock | adjacent | Schrodinger is the computational/in-silico discovery platform on the same thesis |
Mention log
Chamath
tech w=0.90 · n=83
· E112 (2023-01-20) · explicit_prediction · strength 3
▶ 1:25:52
SUPPORT
“And so what's really happened is the value of acquisition in M&A has gone down, even as the technology capability has gone way, way up. And why is that? It's because this stuff has yet to be proven to actually meaningfully improve the hit rate for these drug companies.”
Friedberg
tech w=1.08 · n=65
· E112 (2023-01-20) · explicit_prediction · strength 2
▶ 1:17:22
OPPOSE
“So if you take 99% of the junk out of the top of your pipeline and you only focus on the 1% that the software predicts will be more successful, you much more quickly get drug discovery through the pipeline and you have a much higher hit rate.”
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 -20.7 over the idea's life — that's time passing, attributed to no one
Score events
| episode | kind | Δ | after | note |
|---|---|---|---|---|
| E112 2023-01-20 | init | +38.0 | 38.0 | E112 born by Chamath (explicit_prediction x3) [w=0.90] |
| E112 2023-01-20 | oppose | -12.9 | 25.0 | E112 Friedberg opposes x2 [w=1.08] |
| E113 2023-01-27 | decay | -1.0 | 24.0 | E113 silent |
| E114 2023-02-04 | decay | -1.0 | 23.1 | E114 silent |
| E115 2023-02-11 | decay | -0.9 | 22.1 | E115 silent |
| E116 2023-02-17 | decay | -0.9 | 21.3 | E116 silent |
| E118 2023-03-03 | decay | -0.9 | 20.4 | E118 silent |
| E119 2023-03-11 | decay | -0.8 | 19.6 | E119 silent |
| E120 2023-03-17 | decay | -0.8 | 18.8 | E120 silent |
| E121 2023-03-24 | decay | -0.8 | 18.1 | E121 silent |
| E122 2023-03-31 | decay | -0.7 | 17.3 | E122 silent |
| E123 2023-04-07 | decay | -0.7 | 16.6 | E123 silent |
| E124 2023-04-14 | decay | -0.7 | 16.0 | E124 silent |
| E125 2023-04-21 | decay | -0.6 | 15.3 | E125 silent |
| E126 2023-04-28 | decay | -0.6 | 14.7 | E126 silent |
| E128 2023-05-12 | decay | -0.6 | 14.1 | E128 silent |
| E129 2023-05-19 | decay | -0.6 | 13.6 | E129 silent |
| E130 2023-05-26 | decay | -0.5 | 13.0 | E130 silent |
| E131 2023-06-02 | decay | -0.5 | 12.5 | E131 silent |
| E132 2023-06-10 | decay | -0.5 | 12.0 | E132 silent |
| E133 2023-06-16 | decay | -0.5 | 11.5 | E133 silent |
| E134 2023-06-24 | decay | -0.5 | 11.1 | E134 silent |
| E135 2023-07-01 | decay | -0.4 | 10.6 | E135 silent |
| E136 2023-07-09 | decay | -0.4 | 10.2 | E136 silent |
| E137 2023-07-14 | decay | -0.4 | 9.8 | E137 silent |
| E139 2023-07-27 | decay | -0.4 | 9.4 | E139 silent |
| E140 2023-08-04 | decay | -0.4 | 9.0 | E140 silent |
| E141 2023-08-11 | decay | -0.4 | 8.7 | E141 silent |
| E142 2023-08-18 | decay | -0.3 | 8.3 | E142 silent |
| E143 2023-08-25 | decay | -0.3 | 8.0 | E143 silent |
| E144 2023-09-01 | decay | -0.3 | 7.7 | E144 silent |
| E146 2023-09-22 | decay | -0.3 | 7.4 | E146 silent |
| E147 2023-09-29 | decay | -0.3 | 7.1 | E147 silent |
| E148 2023-10-07 | decay | -0.3 | 6.8 | E148 silent |
| E149 2023-10-13 | decay | -0.3 | 6.5 | E149 silent |
| E150 2023-10-20 | decay | -0.3 | 6.2 | E150 silent |
| E151 2023-10-27 | decay | -0.2 | 6.0 | E151 silent |
| E152 2023-11-03 | decay | -0.2 | 5.8 | E152 silent |
| E156 2023-12-08 | decay | -0.2 | 5.5 | E156 silent |
| E157 2023-12-16 | decay | -0.2 | 5.3 | E157 silent |
| E158 2023-12-23 | decay | -0.2 | 5.1 | E158 silent |
| E159 2023-12-29 | decay | -0.2 | 4.9 | E159 silent |
| E160 2024-01-06 | decay | -0.2 | 4.7 | E160 silent |
| E161 2024-01-13 | decay | -0.2 | 4.5 | E161 silent |
| E162 2024-01-19 | decay | -0.2 | 4.3 | E162 silent |