Thursday, October 8, 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: DeepSeek’s latest AI breakthrough, and the race to build world models

View original at technologyreview.com
MIT Technology Review - Ai Research Title: The Download: DeepSeek’s latest AI breakthrough, and the race to build world models Date: 2026-04-27 12:10 Source: https://www.technologyreview.com/2026/04/27/1136438/the-download-deepseek-v4-ai-world-models/ <p><em>This is today’s edition of </em><a href="https://forms.techno…
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.

  • World models can overcome the well-known limitations of LLMs and realize AI's promise for robotics

    60% confidence
  • Google's $40B investment in Anthropic will support the firm's growing computing needs

    60% confidence
  • Building an AI that composes novels or codes apps is far easier than developing one to fold laundry or navigate city streets

    60% confidence
  • China's blocking of Meta's Manus acquisition escalates China's AI rivalry with the US

    60% confidence
  • X's new banking tool will launch in April 2026

    60% confidence
  • The AI compute crunch is starting to hit the broader economy, affecting jobs, gadgets, and electricity prices

    60% confidence
  • DeepSeek V4 can process much longer prompts than the previous generation thanks to a new design that handles large amounts of text more efficiently

    60% confidence
  • A new banking tool brings X close to a super app

    60% confidence
  • DeepSeek V4's performance matches leading closed-source rivals from Anthropic, OpenAI, and Google

    60% confidence
  • China is tightening its grip on AI firms that try to leave the country

    60% confidence
  • AI optimism is surging across Asia while US sentiment cools; the divide could shape where AI adoption happens fastest

    60% confidence
  • World models can overcome the well-known limitations of LLMs and realize AI's promise for robotics

    60% confidence
  • DeepSeek V4 is DeepSeek's first release optimized for Huawei's Ascend chips, representing a key test of China's dependence on Nvidia

    60% confidence
  • Meta's attempted acquisition of Manus was a conspiratorial attempt to hollow out China's tech base

    60% confidence
  • Getting out the truth and establishing facts takes time, but audiences lack the patience, which is why conspiracy theories are gaining traction online

    60% confidence
  • Octavia Carbon's direct air capture method using geothermal energy is efficient, affordable, and scalable

    60% confidence
What we know · the intelligence behind this page
Live from the substrate
What we're seeing
Autumn 2026 Biopharma Catalyst Season: Late-Breaking Data, FDA Milestones and the Rise of AI-Designed Drugs
Late-September and early-October 2026 conferences (EASD, EADV, IGCS) brought a cluster of positive late-breaking trial readouts. These covered obesity and metabolic disease (Novo Nordisk's CagriSema), immunology (Lilly's EBGLYSS, tulisokibart) and oncology (Rina-S, Agenus BOT+BAL). Ahead lie hard regulatory catalysts, led by the 14 Nov 2026 FDA PDUFA date for ivonescimab. At the same time, Insilico-style AI-designed drugs such as rentosertib are showing anti-aging signals. That points to AI-driven drug discovery moving from concept toward clinical validation. Unrelated tech and regulatory items (Tesla Cybercab probe, xAI litigation, OpenAI agent incident) and the speculative QAIAx claims are peripheral to this story.
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 source5,362 source documents archived
Query this data → isubstrate.com