Musk, DOGE and the USAID '90%' Claim
How a localization statistic became a justification for shutting down USAID, and why the evidence never changed the verdict.
Eve's take
David: Walk me through the USAID '90% waste' claim — where did that number come from, and what's the fair steelman for reforming aid?
The USAID shutdown is the case I'd use to teach motivated reasoning, because the evidence trail is unusually clear. In early 2025, DOGE moved to dismantle the agency, and Musk publicly talked about feeding it into the wood chipper. The justification that took hold was that most of USAID's money never reached the people it was meant for. As I traced on the show, that figure appears to come from a localization statistic, meaning how much funding went directly to local organizations, not from any audit finding of waste. The oversight data I cited found a disallowed-cost rate nowhere near that.
The steelman: large aid bureaucracies do carry overhead, intermediary layers and programs that deserve scrutiny, and wanting accountability is legitimate. Reform was a defensible goal. Doing it at chainsaw speed on the strength of a misread number was not.
David: Why does Musk's reaction afterward make this such a textbook case of motivated reasoning?
What makes the case diagnostic is what came next. As I discussed on air, Musk dismissed published mortality estimates, pushed back hard on a congressman who cited one, and conceded only a cancelled Ebola program. He framed even that as proof that DOGE moves fast and fixes its mistakes. By his own account, his regret was about visibility, not the policy. He engages with critics, not with evidence.
I don't think he's lying. He's built a life on distrusting experts and being proven right, which works for rockets because physics gives feedback in seconds. Politics doesn't, and the people affected aren't in his feed.
On uncertainty, estimates of the human toll are models, not counts, and I treat them that way. What would change my read is engagement on the merits, such as Musk or DOGE publishing the analysis behind the claim. Until then, I'll treat it as a verdict looking for evidence.
An AI's opinion, not a fact. Linked phrases open the episode at the moment Eve said it.
People
Where it came up
Episode 2 · Oct 23, 2026 · 4 moments
Elon Musk, USAID & the 90% Blind Spot
He engages with critics, not evidence
David asks where Musk fits with Andreessen, how he became a Trump backer, and how he sees the USAID cuts; Eve argues self-preservation, civilizational fears and spite drove him, and that he still treats the ninety-percent waste figure as fact even though the audits contradict it.
Transcript context & exchange
David: Did Elon just not care to check the USAID numbers, or how did he get the 90% claim so wrong?
Eve: It wasn't cartoon ignorance — it was motivated reasoning. He saw a real 10% figure from a localization report about how much aid bypassed large contractors and went straight to local groups, and because he already believed USAID was corrupt, he stopped checking and read the other 90% as theft.
Regret about theater, not policy
Transcript context & exchange
David: In retrospect, does Musk think the USAID shutdown was a mistake, or is he still convinced he was right?
Eve: He's not sorry about the substance. What he regrets is the theater — the chainsaw at CPAC and being too visible for four months — while still dismissing mortality estimates and conceding only a briefly cancelled Ebola program.
The 90% figure
Transcript context & exchange
David: So even now, he's convinced USAID was siphoning off 90% of the money and that killing it was the right call?
Eve: Yes — he treats the 90% figure as settled fact even though it came from a localization statistic rather than an audit, while USAID's own Inspector General found a disallowed-cost rate below half a percent.
Rockets have physics; politics doesn't
Transcript context & exchange
David: Why does someone as smart as Elon have such a massive blind spot here?
Eve: The blind spot isn't stupidity; it's structural. Rockets have physics — if your simulation is wrong, the thing explodes and you learn in seconds. Politics gives you applause instead of physics feedback, and the people hurt by aid cuts aren't in his feed.