Listen to today’s podcast: https://www.youtube.com/channel/UC-nqwUyvLDEvs7bV985k-gQ
Today’s podcast episode was created from the following stories:
AI futures model: Dec 2025 update
By Daniel Kokotajlo, Eli Lifland, Brendan Halstead, and Alex Kastner — December 31, 2025
The AI Futures Project’s updated model pushes the median timeline for fully automated coding to roughly 2031–2032, reflecting more conservative assumptions about pre-automation speedups and potential bottlenecks in compute, data, and labor. It also narrows the odds of an ultra-fast takeoff to superintelligence while emphasizing uncertainty, grounding forecasts in benchmark trend extrapolations like METR-HRS. The interactive model invites readers to tweak assumptions and see how outlooks shift.
4 Andreessen Horowitz partners share their 2026 AI predictions
By Alice Tecotzky — January 2, 2026
a16z partners see multimodality—and strong video/image performance—as a key advantage in 2026, potentially favoring Gemini if it sustains momentum. They argue ChatGPT’s app-store push and SDK could define the year, even as the assistant market trends toward “winner-take-most.” For startups, the advice is clear: avoid generic text-in/text-out—find differentiated angles where incumbents are weak.
Chinese tech stocks charged into 2026 with an AI chip IPO surge and more listings lined up
By Huileng Tan — January 2, 2026
Shanghai Biren’s Hong Kong debut jumped nearly 120%, highlighting ravenous demand for China’s homegrown AI hardware amid U.S. export controls on advanced Nvidia chips. The rally, catalyzed by the DeepSeek-R1 breakthrough, is fueling more listings—Baidu’s chip unit, MiniMax, and Zhipu AI are among those queued up. The result: capital is pouring into domestic AI semiconductors, minting new billionaires and lifting the Hang Seng Tech Index.
The AI race is causing a memory chip shortage — and that’s bad news for the smartphone and PC industries
By Henry Chandonnet — January 2, 2026
Hyperscale AI buildouts are soaking up DRAM, NAND, and HBM capacity from Samsung, SK Hynix, and Micron, creating a memory crunch for device makers. IDC now expects deeper 2026 contractions in both smartphone and PC shipments, with potential device price hikes up to 8% as Dell, ASUS, and others signal cost pressures. Bottom line: AI’s infrastructure boom is reshaping consumer hardware economics.
LLMs and a chipmaking boom: Here’s what Xi Jinping said about China’s AI wins in his New Year’s address
By Aditi Bharade — January 2, 2026
Xi Jinping celebrated China’s progress in large AI models and indigenous chips, framing the country as a rapidly rising innovation economy after a standout 2025. DeepSeek-R1’s debut helped ignite domestic momentum even as U.S. export restrictions complicated access to top-tier silicon, with selective exceptions like Nvidia’s H200 for approved buyers. Cross-border moves—such as Meta’s acquisition of Manus—underscore the sector’s intertwined competition and collaboration.
Microsoft CEO Satya Nadella really wants you to stop calling AI “slop” in 2026 — “We are beginning to distinguish between spectacle and substance.”
By Jez Corden — January 2, 2026
Nadella argues AI is shifting from spectacle to substance and will serve as a “scaffolding” for human potential as Microsoft doubles down across its product lines. The piece contrasts that optimism with ongoing skepticism over Copilot’s consumer utility and broader backlash, suggesting AI still lacks “societal permission.” Expect Microsoft to prioritize tangible, high-impact scenarios to win over users in 2026.
AI Slop: Den Augen kann man nicht mehr trauen
By Eva-Maria Weiß — January 2, 2026
As generative content floods platforms, “AI slop” is eroding visual trust; Instagram’s Adam Mosseri warns we may no longer be able to trust our eyes and calls for robust labeling. Creators are leaning into “raw” aesthetics as audiences gravitate toward known, credible sources; a Kapwing analysis suggests a sizable share of YouTube recommendations to new users are low-grade AI videos. The net effect: authenticity signals and provenance become critical for attention and trust.
La paradoja de que Meta compre Manus: su mayor ventaja es no tener modelo propio
By Javier Lacort — January 2, 2026
Meta’s reported $2–2.5B purchase of Manus signals a bet on the application layer as foundational models commoditize. Manus orchestrates third-party models into agentic workflows that ship finished work—distribution and productization, not raw model IQ, are the differentiators. The deal also highlights a playbook for globally focused Chinese AI startups to capture Western capital and exits while sidestepping domestic constraints.
Taken together, these stories chart a year where AI moves from model showdowns to real-world impact—pressuring supply chains, reshaping capital markets, and forcing platforms and policymakers to rebuild trust and distribution. Expect the winners of 2026 to pair credible capabilities with clear value, transparency, and reach—from chips and infrastructure to the last mile where users actually get things done.
