Thursday, October 15, 2026
Source trace. Via News points to the documents behind its reporting and shows what we drew from each — so you can check any claim. How we source
Source document

The Download: metric weaknesses and AI elephant warnings

View original at technologyreview.com
MIT Technology Review - Ai Research Title: The Download: metric weaknesses and AI elephant warnings Date: 2026-06-29 12:10 Source: https://www.technologyreview.com/2026/06/29/1139834/the-download-metric-weaknesses-ai-elephant-warnings/ <p><em>This is today’s edition of </em><a href="https://forms.technologyreview.com/n…
Opening lines of the source · short snapshot — read the full document at the original

What we drew from this source

The claims Via News extracted from this document. We point to the source; we don't replace it.

  • External metrics and data can never capture what is truly important; worse, they inevitably redefine one's core sense of what is important, whether or not one is aware of the trap.

    60% confidence
  • A new study suggests 6.4 to 7.8 hours of sleep is the optimal range; too much sleep may be as harmful as too little.

    60% confidence
  • Google capping Meta's access to Gemini AI compute has disrupted and delayed some Meta AI projects.

    60% confidence
  • India is home to about 60% of the world's wild Asian elephants, with around 80% of their habitat outside protected areas, leading to deadly human-wildlife conflicts.

    60% confidence
  • Security researchers say Zhipu AI is poised to reset the AI race after its model matched Anthropic's Mythos in finding security bugs.

    60% confidence
  • Android phones gave Venezuelan earthquake victims between seconds and up to two minutes of advance notice before the seismic events struck.

    60% confidence
  • US restrictions on AI models have sparked alarm that they are boosting China's AI progress.

    60% confidence
  • Chipmakers are profiting off AI at the expense of everyone else.

    60% confidence
  • Severe weather is now the leading cause of loss for data centers, as extreme heat puts AI-heavy data centers under significant operational pressure.

    60% confidence
  • Appropriate safeguards are now in place for the release of Anthropic's Mythos 5 model to trusted organizations.

    60% confidence
  • Powerful AI capable of altering the landscape of cyberwarfare cannot remain solely in American hands; Chinese AI firms must match US capabilities.

    60% confidence
  • There have been some 3,000 human casualties from human-elephant conflicts in India in the last five years, and over 1,000 elephant deaths since 2014.

    60% confidence
  • Research shows Gen Z are more likely to believe and pass on misinformation when they feel a sense of common identity with the person who shared it, with social media shifting credibility from community to identity-based authority.

    60% confidence
  • Zhipu AI's model still cannot match Anthropic or OpenAI on general tasks despite matching Mythos in security bug detection.

    60% confidence
What we know · the intelligence behind this page
Live from the substrate
What we're seeing
Q4 2026 Clinical Catalyst Season: Big-Pharma Data Flow, Regulatory Milestones, and a Thin AI-Drug Fringe
October 2026 conferences (EASD, IGCS, NACFC, HFSA) are producing a dense run of late-stage readouts from Novo Nordisk, Lilly, Roche, Bayer, Vertex and Galderma. Regulatory milestones are scheduled behind them: the ivonescimab BLA PDUFA date on 2026-11-14, FDA acceptance of Roche's vamikibart BLA, and Scholar Rock's EMA resubmission. The AI-and-finance angle is peripheral. It consists of the AI-designed drug Rentosertib and QAIAx, a speculative microcap whose trial claims (the QIII pilot, anticipated 2027-01-01, and a patent update) are forecasts and filings rather than data, so they warrant skepticism.
Our read on the data ›
Signals we're tracking
EPKINLY Regulatory-Clinical Success Cascade
High probability of expanded label indications, additional combination approvals, and competitive positioning strength in follicular lymphoma market. Predicts positive commercial uptake and potential accelerated review for related indications.
Patterns we're watching ›
Where sources disagree
ING Group
Both facts record the same metric (shares_outstanding) for ING Group at the identical observation date (2025-12-31). FACT A states 2,902,437,688 shares; FACT B states 2,902 million shares (2,902,000,000). The difference is 437,688 shares (~0.015%). This is a genuine value conflict, though the discrepancy appears to result from FACT B rounding to the nearest million while FACT A provides the precise count.
We flag conflicts openly ›
Recently verified
✓ Checked against the original source
4,986
facts traced to their source — and we flag the ones that don't hold up.
101 entities tracked4,986 facts checked against source5,375 source documents archived
Query this data → isubstrate.com