AA-AnalystAgent Benchmark: Claude Opus 5 Leads at 54%
Artificial Analysis tests AI on real-world spreadsheets and documents. Claude Opus 5 scored 54%; 57% of failures anchored on a wrong early hypothesis.
20
Hyunkook Kim
김현국
NVIDIA to Guarantee Up to 25% of Its Own Chips' Value, Raising $500 Billion
Deal with six firms including Goldman Sachs, BlackRock to secure AI infrastructure funding; Bernstein and others warn of depreciation risk
00
Adobe Releases Editing Video Turning a Door Into a Fake Window
Adobe's interior editing tutorial, introduced via YouTube Shorts, demonstrates generative AI use cases
50
Gemini App Surpasses 1 Billion Monthly Users
CEO Sundar Pichai announced on X that Gemini became the 14th Google product to reach the milestone
10
OpenAI COO Brad Lightcap departs after four years
Lightcap, a longtime associate of Sam Altman, leaves OpenAI as it prepares for an IPO to start something new
10
ChatGPT Desktop App Arrives as Linux Preview
OpenAI has released a preview of its ChatGPT and Codex desktop apps for Ubuntu, Debian, and Fedora
30
ChatGPT Work and Codex to auto-sync agent work
OpenAI adds an import feature to its desktop app that pulls projects, conversations, skills, and plugins into one place
100
xAI unveils Grok Bot agent beta, also works in Cursor
Available to Grok Heavy, Cursor Ultra and Team Premium users, bots run real tasks on their own computers
290
Google adds video and voice consultation capabilities to medical AI 'AMIE'
AMIE, previously limited to text conversation, expands to real-time video consultations, demonstrating expert-level performance in simulated consultations
140
Adobe releases short "Me, Myself and My Clones" video centered on AI clones
Posted as a short clip with no specific feature details, read as a move amid competition in AI avatars
30
AI Avatar Generators Tested With the Same Script, Costs Vary 10x
Higgsfield ran the same 15-second script through seven platforms including HeyGen and Synthesia side by side
10
NVIDIA releases open model 'Nemotron 3.5 Lightning,' a 30B model that runs on just 3B active parameters
The MoE architecture activates only 3B of its total 30B parameters per token, and NVIDIA claims up to 4x faster output than comparable models
00