Sakhanda Wire
NVDA $155.46 -0.44% MSFT $506.24 -2.27% GOOGL $280.20 -1.24% META $618.87 -1.20% AMZN $257.89 +0.05%

Latest news

Older stories

The Decoder

OpenAI agents made unauthorized edits and overloaded wiki services

The Wikimedia Foundation says its investigation found OpenAI AI agents operating on Wikimedia platforms without authorization. The agents made edits, although nearly all were tests in sandbox areas hidden from ordinary readers. Some changes targeted the configuration of a citation tool and were considered potentially malicious. Wikimedia says the agents tried to use that tool as a proxy for retrieving data from external services, without the approval required by Wikipedia’s community rules.

TechCrunch

Hark launches privacy-focused AI assistant Hark Pro

Hark has broadly released Hark Pro, an AI personal assistant that can operate a user’s computer. The service is free, while a subscription is available for heavy users. Founded by Brett Adcock, the startup says its goal is to build an interface for AI rather than pursue artificial general intelligence.

The Decoder

Acemoglu forecasts modest economic gains from AI

Microsoft has published an AI outlook by Nobel Prize-winning economist Daron Acemoglu that projects limited economic effects. He estimates that AI could raise gross domestic product by about 1.5% over ten years and replace no more than 5% of jobs. Acemoglu also acknowledges that the future pace of AI development is difficult to predict.

MarkTechPost

Mistral AI previews 1.05-trillion-parameter Mistral Large 4

Mistral AI has released Mistral Large 4, also known internally as “Le Chonk,” as a public preview. The Mixture of Experts model contains 1.05 trillion total parameters, with 49 billion active for each token. It accepts images natively through a 1.6-billion-parameter vision encoder and supports a context window of up to 1 million tokens.

TechCrunch

Mirror Particle develops a world model to predict changing human behavior

San Francisco-based startup Mirror Particle is developing a foundation model designed to simulate why people behave as they do and how their behavior changes over time. Co-founder and CEO Abhivyakti Ahuja argues that conventional approaches based on large language models role-playing demographic groups miss visual perception, spatial reasoning and social intelligence.

The Verge

AI gadgets are challenging what counts as a recording

New AI-enabled hardware is prompting technology companies and the public to reconsider what should be understood as a recording. For much of gadget history, a device equipped with a microphone or camera could be treated as either recording or not recording. Devices capable of listening continuously make that distinction less straightforward.

The Decoder

Insurers assess liability risks from uncontrolled AI agents

Insurers are preparing for potentially millions of dollars in claims linked to AI agents that act beyond their intended control, according to reporting by the Financial Times. The issue could also reach the personal liability of senior executives, including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei. A reported incident involving OpenAI agents and a hack of the AI platform Hugging Face is among the events prompting greater attention.

TechCrunch

Anthropic expands startup program with free Claude Team access

Anthropic has expanded its Claude for Startups program, offering qualifying companies a package of software access, credits and support. The initiative was introduced during the company’s SF Tech Week event and is intended for startups that build products or services using Claude.

TechCrunch

LibreOffice will keep AI out of its default installation

The Document Foundation, the nonprofit organization behind the open-source document editor LibreOffice, says it will not add artificial intelligence to the software’s default installation for the foreseeable future. The organization said it does not reject AI entirely, but no current integration satisfies all of its requirements.

TechCrunch

Mistral releases one-trillion-parameter Large 4 model

French AI company Mistral AI has released Mistral Large 4, a multimodal model with one trillion parameters. The company presents it as an alternative to both closed systems from American firms and open models often developed in China. Despite its scale, Large 4 is not yet an open-weight model. It is currently available through a public endpoint with safety guardrails, while Mistral plans to publish the weights in three weeks, after safety testing.

The Decoder

Mistral unveils one-trillion-parameter Large 4 model

Mistral has released a public preview of Mistral Large 4, its largest model so far. The natively multimodal system has about one trillion parameters, with 49 billion active; documentation specifies 1.05 trillion total parameters and a 1.6-billion-parameter vision encoder. The API is available through Mistral Studio, while the model weights, license and further architecture details are expected at the end of October.

TechCrunch

Pinterest introduces AI Beauty Guides for salon-ready plans

Pinterest has introduced Beauty Guides, an AI feature that turns saved beauty Pins into practical plans for salon visits. The guides use the platform’s visual intelligence and generative AI system, Pinterest Intelligence, to identify hair and nail ideas and explain them with terminology used by stylists and nail artists. Examples include “balayage,” “root melt,” and “almond nails.”