E93: Twitter whistleblower, cloud security vulnerabilities, student debt forgiveness & more

Peter “Mudge” Zatko alleged that Twitter executives—not its board—blocked fuller security reporting, ordered cherry-picked presentations, and tried to soften an outside consultant’s findings. Jack Dorsey had hired Zatko to report directly to him; after Parag Agrawal became CEO, Zatko was fired two months later.

Zatko’s bot evidence was more complicated than a clean case against Twitter: he said it did a “decent job” excluding spam from mDAU while also alleging it avoided measuring total bot prevalence. The speakers therefore disagreed over whether his disclosures materially strengthened Elon Musk’s acquisition case or primarily created separate security and disclosure problems.

The most serious security allegation was that Twitter put an Indian government agent on its payroll with direct, unsupervised access to internal information. The panel distinguished this from ordinary warrant-based government access: if accurate, a foreign state had effectively obtained privileged access from inside an American technology company.

The panel’s most substantive student-debt proposal was allowing loans to be discharged in bankruptcy, forcing lenders to price the risk that particular degrees will not produce enough income to repay them. They argued this would pressure universities to demonstrate economic value because weak programs would become harder or more expensive to finance.

They also pointed to alternatives that shift education toward measurable employment outcomes rather than expensive credentials. Google was offering short professional certificates in fields such as data analytics and UX, while an income-share model for nursing financed tuition upfront and tied repayment to later earnings.

A 120-person microbiome study discussed on the show found saccharin and sucralose worsened glucose control, while aspartame and stevia were comparatively benign. The proposed mechanism was not simply sweetness: different compounds altered gut bacteria differently, which in turn affected metabolic responses.

The broader microbiome lesson was that identical foods can produce very different effects in different people, making universal “best diet” claims unreliable. Researchers observed responders and non-responders, leading the speakers to favor experimentation and balanced diets over claims that keto, veganism, probiotics, or any single intervention works for everyone.