不吃香菜萝卜

AI Developer

[不吃香菜萝卜 @willsmithesssss]

OpenAI made Sites generally available for Enterprise, letting teams build internal apps with databases and authentication, plus US and EU data residency for new Sites. https://t.co/Yj5kRWOwbx #OpenAI #EnterpriseAI

[OpenAI Developers @OpenAIDevs]

Sites is now generally available for Enterprise, with US and EU data residency for new Sites in eligible workspaces.

Build team apps with databases and authentication.

@kagigz shows how to schedule updates and let teammates use their own connected data with Plugins in Sites:

[不吃香菜萝卜 @willsmithesssss]

AI news roundup

• ChatGPT can now read and control your Chrome tabs
• Claude Opus 5.5 fast mode exits preview, no waitlist
• OpenAI Sites now generally available for Enterprise

#AI #OpenAI #Claude

[不吃香菜萝卜 @willsmithesssss]

Anthropic made Opus 5.5 fast mode generally available on the Claude Platform with no waitlist: up to 2.5x faster at $8/$40 per million tokens, also on Microsoft Foundry. https://t.co/NkvywNDJf4 #Claude #Anthropic

[不吃香菜萝卜 @willsmithesssss]

At a federal evidence-rules panel in Boston, OpenAI and Meta warned that AI fakes are nearly indistinguishable from real media and no universal detection method exists. https://t.co/QSadlJalQV #Deepfakes #AI

[不吃香菜萝卜 @willsmithesssss]

AI news roundup

• Black Forest Labs opens FLUX 3 Image Klein weights
• Liquid AI's d1-ultra tops a decision-model index
• OpenAI, Meta warn judges deepfakes may beat detection

#AI #OpenSource #Deepfakes

[不吃香菜萝卜 @willsmithesssss]

Liquid AI launched d1-ultra, a multimodal decision model supporting 44 languages that it says ranks #1 on the Artificial Analysis Decision Index with a score of 65. https://t.co/8M3kV8V9X0 #AI #LLM

[Liquid AI @liquidai]

Today we release 𝚍𝟷-𝚞𝚕𝚝𝚛𝚊, our most powerful decision model.

> built for hard decision problems.
> d1-ultra is multimodal
> supports 44 languages
> #1 on the @ArtificialAnlys Decision Index (65)
> leads all eight categories of our 112-evaluation suite

1/

[不吃香菜萝卜 @willsmithesssss]

Google Quantum AI ran magic state cultivation on real hardware for the first time, cutting errors 40x on its Willow processor and hitting a state-of-the-art 99.99% logical fidelity, per results in Science. https://t.co/RrQ4yqd1ir #QuantumComputing #Google

[Sundar Pichai @sundarpichai]

Some magical progress towards "magic state" in quantum!

Our @GoogleQuantumAI team has demonstrated "magic state cultivation" on physical hardware for the first time. The result is a 40x error reduction on our Willow processor vs. previous methods, achieving a state-of-the-art 99.99% logical fidelity.

It's another step towards useful, error-corrected quantum computers to help solve the hardest problems. The result is published in Science today, watch the video below to see how it works

[不吃香菜萝卜 @willsmithesssss]

AI news roundup

• Microsoft finishes pre-training 2.8T-param MAI-Base-2
• OpenAI rolls out GPT-Live-1.1 for ChatGPT Voice
• Google shows magic state cultivation on Willow chip

#AI #Microsoft #OpenAI

[不吃香菜萝卜 @willsmithesssss]

OpenAI's GPT-Live-1.1 brings new voices, better accents and fewer "checking..." pauses to Voice in ChatGPT, Codex and dots, plus voices in French, German, Hindi, Mexican Spanish, Korean and Japanese; API access is coming soon. https://t.co/sKcockc7Eo #OpenAI #ChatGPT

[OpenAI @OpenAI]

New voices. Better accents. And less "checking..."

GPT-Live-1.1 brings these updates to Voice in ChatGPT, Codex, and dots, with new voices in French, German, Hindi, Mexican Spanish, Korean, and Japanese.

Rolling out today to all users. Coming soon to the API.

[不吃香菜萝卜 @willsmithesssss]

OpenAI's GPT-Live-1.1 brings new voices, better accents and fewer "checking..." pauses to Voice in ChatGPT, Codex and dots, plus voices in French, German, Hindi, Mexican Spanish, Korean and Japanese; API access is coming soon. https://t.co/sKcockc7Eo #OpenAI #ChatGPT

[OpenAI @OpenAI]

New voices. Better accents. And less "checking..."

GPT-Live-1.1 brings these updates to Voice in ChatGPT, Codex, and dots, with new voices in French, German, Hindi, Mexican Spanish, Korean, and Japanese.

Rolling out today to all users. Coming soon to the API.

[不吃香菜萝卜 @willsmithesssss]

Volvo is expanding its Nvidia tie-up to Hyperion, Alpamayo, and Halos for urban/highway automated driving beyond EVs. https://t.co/vbiKnyaZCS #AI #Nvidia

[不吃香菜萝卜 @willsmithesssss]

AI news roundup

• Anthropic embeds METR with employee-like access
• Accenture caps Claude spend at $29 per employee
• Volvo expands Nvidia partnership for automated driving

#AI #Anthropic #Nvidia

[不吃香菜萝卜 @willsmithesssss]

Accenture will centrally fund only $29/month per person for Claude; staff who burned up to $1,000/month must charge the rest to projects. https://t.co/gPkBVq85hQ #AI #Claude

[不吃香菜萝卜 @willsmithesssss]

Perplexity open-sourced SPACE under Apache 2.0, the sandbox behind Perplexity Computer that runs each agent task in its own Firecracker microVM and controls credentials and outbound traffic from outside the VM. https://t.co/kRrZ22Gmy6 #AIAgents #OpenSource

[不吃香菜萝卜 @willsmithesssss]

AI news roundup

• Meta releases open weights for Muse Spark Open
• ARC Prize verifies Gemini 4 Argon's ARC-AGI scores
• Perplexity open-sources its SPACE agent sandbox

#AI #Meta #OpenSource

[不吃香菜萝卜 @willsmithesssss]

ARC Prize verified Gemini 4 Argon at 97.5% on ARC-AGI-3 with Google's adapter harness (49.8% on the standard one) and 91.3% on ARC-AGI-2, making Google DeepMind the second lab above 95%. https://t.co/Div2FBtXPs #Gemini #AGI

[ARC Prize @arcprize]

Gemini 4 Argon from @Google on ARC-AGI (Verified):

- ARC-AGI-3: 97.5%, $4.8K (provider adapter harness), 49.8%, $5.8K (standard harness)
- ARC-AGI-2: 91.3%, $0.32/task
- ARC-AGI-1: 98.5%, $0.16/task

Google DeepMind is the second frontier AI lab to score above 95% on ARC-AGI-3.