Sakhanda Wire
NVDA $151.03 -2.85% MSFT $507.33 +0.22% GOOGL $283.17 +1.06% META $614.95 -0.63% AMZN $266.57 +3.37%

Latest news

Older stories

MarkTechPost

Cohere releases 218B-parameter translation model for 50 languages

Cohere has released North Small Translate, an open-weight machine-translation model developed with Cohere Labs and RWS. The sparse Mixture-of-Experts system contains 218B total parameters, with 25B active for each token, and supports 50 languages from Albanian to Vietnamese. On Cohere’s WMT26 evaluation, it achieved an average score of 83.60 across all languages; an Agentic version using multiple passes reached 84.36.

MarkTechPost

Sakana AI Releases Fugu Max and Fugu Ultra v2

Sakana AI has released Fugu Max and Fugu Ultra v2, two models in its Fugu family. Fugu is a learned orchestrator rather than a single foundation model: it reads a request and routes work among other models through one API. Both releases use the same orchestration architecture, but Fugu Max is optimized for cost efficiency while Fugu Ultra v2 targets difficult, multi-step tasks.

MarkTechPost

ToolGrad reaches 99.8% pass rate for tool-use data generation

Researchers from Google, the University of Tokyo, RIKEN AIP and Tohoku University have introduced ToolGrad, a framework that reverses the usual process for creating tool-use training data. Instead of generating a user request first and searching for a suitable API sequence, it executes APIs to build a verified chain, then generates a matching query and response.

MarkTechPost

Redis launches managed semantic cache for LLM applications

Redis has introduced LangCache, a fully managed semantic caching service designed to reduce repeated calls to large language models. It compares incoming prompts with earlier requests by meaning rather than exact wording and can return a stored answer when the similarity is high enough. Redis claims savings of up to 90% on API costs and cache-hit responses up to 15 times faster than a new model request.

MarkTechPost

NVIDIA reports faster biomolecular folding with BioIR

NVIDIA has detailed BioNeMo Inference Runtime (BioIR), a Python library designed to speed up supported biomolecular structure-prediction models on NVIDIA GPUs while retaining standard PyTorch workflows. The library targets operations including Pairformer and Evoformer stacks, triangle operations, pairwise attention, diffusion transformers and atom-level modules. It can process complete worklists or provide optimized modules for custom PyTorch applications.

TechCrunch

Jensen Huang says Nvidia revenue could grow 70% next year

Nvidia CEO Jensen Huang said the company could increase its revenue by 70% year over year next year, repeating an outlook first given when Nvidia reported another record quarter. Analysts expect about $400 billion in revenue for the current fiscal year, which would put the projected figure at roughly $680 billion after such growth.

TechCrunch

Mark Wahlberg to discuss investing and entrepreneurship at TechCrunch Disrupt 2026

Mark Wahlberg is scheduled to appear on the main stage at TechCrunch Disrupt 2026 for a conversation with Bruce K. Lee, founder and CEO of Keebeck Wealth Management. The discussion will focus on Wahlberg’s development as an entrepreneur and investor rather than primarily on his acting career, covering business building, healthcare, wellness and technology.

The Verge

Slack introduces AI-built interactive tools inside chats

Slack is introducing Slackforce Surfaces, a feature that lets users create interactive reports, polls, dashboards, presentations, microsites and other tools within a chat. Users describe what they want to Slackbot, which uses AI to collect information from relevant conversations and connected services such as Google Drive and Salesforce.

MarkTechPost

OpenAI releases Agents API in public beta

OpenAI has released the Agents API in public beta, making the managed harness and infrastructure used by Codex available through an API. OpenAI hosts and maintains the harness, while developers can run agent compute in an OpenAI-managed sandbox, their own infrastructure, a partner sandbox, or without a sandbox. The service is available to all developers. Data residency is limited to the United States, and Zero Data Retention is not supported.

TechCrunch

OpenAI pauses $200 Pro subscriptions amid Astra demand

OpenAI has temporarily stopped accepting new subscriptions for its $200-per-month Pro plan, saying demand for its Astra model is placing unusually heavy pressure on the company’s infrastructure. Thibault (Tibo) Sottiaux, OpenAI’s product leader for core products including Codex and ChatGPT, announced the change on X.

TechCrunch

Anthropic reports large-scale AI distillation campaigns

Anthropic says China-based AI companies have conducted increasingly sophisticated attempts to extract capabilities from its Claude models. In a report released Thursday, the company said it identified five campaigns involving nearly 200 million exchanges. The targeted capabilities included agentic behavior and tool use, coding, data analysis, and logical reasoning. OpenAI has previously described similar activity and linked it specifically to DeepSeek.

TechCrunch

Meta’s Muse reaches No. 2 on the U.S. App Store

Meta’s AI agent Muse has been downloaded more than 83,000 times on iOS in the United States, according to data from market intelligence firm Sensor Tower. The figure placed it second in the U.S. App Store’s overall rankings. Muse is currently available in the U.S., while its web and WhatsApp versions are not included in the estimate.