Waymo paused its driverless car service in San Francisco on December 20 after a blackout caused widespread outages.
confirmed 1 article 2025-12-24
Waymo driverless car service San Francisco
Waymo says it is updating its fleet to improve navigation during widespread outages, after pausing its service during blackouts in San Francisco on December 20
CNBC · 2025-12-24 · #6
Waymo says it is updating its driverless vehicle fleet's navigation systems to improve navigation during widespread outages.
confirmed 1 article 2025-12-24
Waymo driverless vehicle fleet navigation systems
Waymo says it is updating its fleet to improve navigation during widespread outages, after pausing its service during blackouts in San Francisco on December 20
CNBC · 2025-12-24 · #6
AST launches its largest-ever satellite from India carrying the BlueBird Block-2 payload.
confirmed 1 article 2025-12-24
AST BlueBird Block-2 India
AST launches its largest-ever satellite from India, hosting the BlueBird Block-2, the first in a series of deployments to challenge SpaceX in satellite internet
Bloomberg · 2025-12-24 · #11
Dazzle plans to exit stealth in 2026.
confirmed 1 article 2025-12-24
Dazzle 2026
Marissa Mayer's new startup Dazzle, which is building next-gen AI personal assistants, raised an $8M seed at a $35M valuation and plans to exit stealth in 2026
TechCrunch · 2025-12-24 · #27
Dazzle is building next-generation AI personal assistants.
confirmed 1 article 2025-12-24
Dazzle next-gen AI personal assistants
Marissa Mayer's new startup Dazzle, which is building next-gen AI personal assistants, raised an $8M seed at a $35M valuation and plans to exit stealth in 2026
TechCrunch · 2025-12-24 · #27
GLM-4.7 is released as an open-weight (open-weights) model.
confirmed 1 article 2025-12-23
GLM-4.7
Chinese AI startup Z.ai releases GLM-4.7, an open-weight model that Z.ai says delivers significant improvements in coding performance compared to GLM-4.6
Z.ai · 2025-12-23 · #29
Z.ai designates GLM-4.7 as the default model in the GLM Coding Plan.
confirmed 1 article 2025-12-23
Z.ai GLM-4.7
Chinese AI startup Z.ai releases GLM-4.7, an open-weight model that Z.ai says delivers significant improvements in coding performance compared to GLM-4.6
Z.ai · 2025-12-23 · #29
DGX Cloud is reportedly pivoting from selling cloud services to enterprises to supporting Nvidia's internal AI development.
rumored 1 article 2025-12-23
Nvidia DGX Cloud selling cloud services to enterprises supporting internal AI development
Sources: Nvidia's DGX Cloud is merging with the engineering unit, pivoting from selling cloud services to enterprises to supporting internal AI development
The Information · 2025-12-23 · #27
Notion (NotionHQ) is building agents for knowledge work; these agents are viewed as having gains limited only by imagination and inertia.
confirmed 1 article 2025-12-23
Notion
How AI can work across scales, from individuals to organizations to economies, like steel and the steam engine before it, as AI arrives as “infinite minds”
@ivanhzhao · 2025-12-23 · #21
Chinese AI startup Z.ai releases GLM-4.7, a new model.
confirmed 1 article 2025-12-23
Z.ai GLM-4.7
Chinese AI startup Z.ai releases GLM-4.7, an open-weight model that Z.ai says delivers significant improvements in coding performance compared to GLM-4.6
Z.ai · 2025-12-23 · #29
OpenAI is building an LLM-based automated attacker to test and improve Atlas browser defenses against prompt injection attacks.
confirmed 1 article 2025-12-23
OpenAI LLM-based automated attacker ChatGPT Atlas browser prompt injection attacks
OpenAI details efforts to secure its ChatGPT Atlas browser against prompt injection attacks, including building an “LLM-based automated attacker”
TechCrunch · 2025-12-23 · #4
AI is described as arriving as 'infinite minds' that can continuously perform delegated work.
confirmed 1 article 2025-12-23
AI
How AI can work across scales, from individuals to organizations to economies, like steel and the steam engine before it, as AI arrives as “infinite minds”
@ivanhzhao · 2025-12-23 · #21
MiniMax releases M2.1, an upgrade to its open-source M2 model, claiming significantly enhanced coding capabilities in Rust, Java, and other languages.
confirmed 1 article 2025-12-23
MiniMax M2.1 M2
China's MiniMax releases M2.1, an upgrade to its open-source M2 model that it says has “significantly enhanced” coding capabilities in Rust, Java, and others
MiniMax · 2025-12-23 · #10
The essay compares AI's transformative potential to past technologies like steel and the steam engine.
confirmed 1 article 2025-12-23
AI steel steam engine
How AI can work across scales, from individuals to organizations to economies, like steel and the steam engine before it, as AI arrives as “infinite minds”
@ivanhzhao · 2025-12-23 · #21
Z.ai says GLM-4.7 delivers significant improvements in coding performance compared to GLM-4.6.
rumored 1 article 2025-12-23
Z.ai GLM-4.7 GLM-4.6
Chinese AI startup Z.ai releases GLM-4.7, an open-weight model that Z.ai says delivers significant improvements in coding performance compared to GLM-4.6
Z.ai · 2025-12-23 · #29
Adam Mosseri says Instagram may need a premium longform video offering, indicating consideration of a paid long-form content product for the platform.
rumored 1 article 2025-12-22
Adam Mosseri premium longform video Instagram
Adam Mosseri says “maybe we'll need” premium longform video for Instagram, and that TikTok's US sale turmoil has given Meta time to retune algorithms
Semafor · 2025-12-22 · #17
Colin Angle has a new robotics startup.
confirmed 1 article 2025-12-22
Colin Angle robotics startup
Q&A with iRobot co-founder Colin Angle on the company's Chapter 11 bankruptcy, regulatory pressure that killed Amazon's deal, his new robotics startup, and more
TechCrunch · 2025-12-22 · #13
Mosseri says TikTok's US sale turmoil has given Meta time to retune Instagram's algorithms.
confirmed 1 article 2025-12-22
Adam Mosseri Meta TikTok US sale turmoil
Adam Mosseri says “maybe we'll need” premium longform video for Instagram, and that TikTok's US sale turmoil has given Meta time to retune algorithms
Semafor · 2025-12-22 · #17
Google faces the innovator's dilemma, which may complicate adopting disruptive approaches despite competitive AI offerings.
confirmed 1 article 2025-12-22
Google
Google's TPUs and Gemini are competitive but Nvidia's and OpenAI's first-mover advantage makes their lead seem durable, and Google faces the innovator's dilemma
SiliconANGLE · 2025-12-22 · #11
Sources report Nvidia plans to begin shipping H200 AI chips to China before mid-February 2026, expecting initial shipments of roughly 40,000–80,000 units.
rumored 1 article 2025-12-22
Nvidia H200 chips China
Sources: Nvidia plans to begin shipping its H200 chips to China before mid-February 2026 and expects initial shipments to be ~40,000 to 80,000 H200 units
Reuters · 2025-12-22 · #2
Coverage and analysis comparing TPU and GPU architectures, framed as a modern-era showdown (likened to CISC vs RISC).
confirmed 1 article 2025-12-22
TPU GPU
A look at Google's TPU evolution, its decision to sell chips to competitors after 12 years of internal use, the TPU vs. GPU architecture showdown, and more
The Chip Letter · 2025-12-22 · #10
Google's TPUs and Gemini are competitive offerings in AI hardware and models.
confirmed 1 article 2025-12-22
Google TPUs Gemini
Google's TPUs and Gemini are competitive but Nvidia's and OpenAI's first-mover advantage makes their lead seem durable, and Google faces the innovator's dilemma
SiliconANGLE · 2025-12-22 · #11
Waymo resumes its driverless robotaxi service in the San Francisco Bay Area after a temporary suspension during a blackout.
confirmed 1 article 2025-12-22
Waymo robotaxi service San Francisco Bay Area
Waymo resumes its robotaxi service in SF after suspending it during a blackout, and says most trips finished before vehicles returned to depots or pulled over
CNBC · 2025-12-22 · #18
Nvidia and OpenAI's first-mover advantage contributes to a durable lead in AI capabilities and market position.
confirmed 1 article 2025-12-22
Nvidia OpenAI market_lead
Google's TPUs and Gemini are competitive but Nvidia's and OpenAI's first-mover advantage makes their lead seem durable, and Google faces the innovator's dilemma
SiliconANGLE · 2025-12-22 · #11
Waymo temporarily suspended its robotaxi service in the San Francisco Bay Area during a blackout.
confirmed 1 article 2025-12-22
Waymo robotaxi service blackout San Francisco Bay Area
Waymo resumes its robotaxi service in SF after suspending it during a blackout, and says most trips finished before vehicles returned to depots or pulled over
CNBC · 2025-12-22 · #18
Shopify launches a new AI commerce stack; the company's VP of Product details the just-launched products to help entrepreneurs and developers build with the latest AI and technology.
confirmed 6 articles 2025-12-16 — 2025-12-21
Shopify AI commerce stack
Josh Woodward is credited with turning around the Gemini app, growing monthly active users from 350 million in March to 650 million by October.
confirmed 1 article 2025-12-21
Josh Woodward Gemini app
A profile of Josh Woodward, the head of Google Labs who is credited with turning around the Gemini app, growing its MAUs from 350M in March to 650M by October
CNBC · 2025-12-21 · #12
Claude Opus 4.5's 50% time horizon (~4h49m) is reported to be more than double that of Claude Opus 4 released earlier this year.
confirmed 1 article 2025-12-21
Claude Opus 4.5 Claude Opus 4 Claude Opus 4
METR: Claude Opus 4.5 has a 50% task completion time horizon of about 4 hours and 49 minutes, more than double that of Claude Opus 4 released earlier this year
@metr_evals · 2025-12-21 · #2
Moore Threads announces a new generation of AI chips slated for mass production beginning in 2026, aimed at reducing AI developers' dependence on other providers.
confirmed 1 article 2025-12-21
Moore Threads new generation of chips mass production from 2026
Chinese AI chipmaker Moore Threads announces a new generation of chips slated for mass production from 2026, just weeks after its blockbuster IPO in China
Bloomberg · 2025-12-21 · #8
On METR tasks, Claude Opus 4.5 has an estimated 50% task completion time horizon of about 4 hours 49 minutes (95% confidence interval: 1 hr 49 mins to 20 hrs 25 mins).
confirmed 1 article 2025-12-21
Claude Opus 4.5 4 hours 49 minutes 95% CI: 1 hr 49 mins to 20 hrs 25 mins
METR: Claude Opus 4.5 has a 50% task completion time horizon of about 4 hours and 49 minutes, more than double that of Claude Opus 4 released earlier this year
@metr_evals · 2025-12-21 · #2