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

Claude resolves a 25-year-old mystery behind NIST cryptographic curves

Anthropic says its AI model Claude has identified how the unexplained seeds behind several NIST elliptic curves were generated. The curves were standardized in 2000 for ECDSA, a widely used system for digital signatures. The NSA had provided the starting values without describing their origin, which led to years of speculation that they might contain a deliberate backdoor.

The Verge

Websites block AI agents from carrying out online tasks

Autonomous AI agents are facing restrictions from websites they are designed to use. OpenAI, Meta, Google, Instinct and other companies have presented agents that can browse the internet and act for users, including making restaurant reservations, managing inboxes and purchasing shoes.

The Decoder

Deepseek V4.1-Flash-1017 targets faster, cheaper agentic coding

Deepseek has released Deepseek-V4.1-Flash-1017 through its website, app and API, while also publishing the model weights on Hugging Face and ModelScope. The company says the open-weight model is faster than V4.1-Flash, generates shorter responses and achieves higher scores in software development and agent tasks. It was trained with reinforcement learning using self-feedback, without human evaluators or external reward models.

The Decoder

Experts question OpenAI’s understanding of its Navier–Stokes proof

OpenAI presented a 166-page potential solution to the Navier–Stokes problem in early September, saying it followed work by Spanish mathematicians Diego Córdoba and Luis Martínez-Zoroa. Several expert groups disagree with that characterization. Gonzalo Cao-Labora of EPFL said the resemblance was mainly superficial and that the detailed method was different.

MarkTechPost

DeepSeek releases V4.1-Flash-1017 and opens Search and Fetch APIs

DeepSeek has released DeepSeek-V4.1-Flash-1017, an updated checkpoint of its V4.1-Flash model. It is available through the company’s website, app and API under the model name deepseek-flash. DeepSeek says the release improves performance, completes tasks faster and produces shorter outputs than the September version. The model has a 1M-token context window and supports up to 384K output tokens; its hardware requirements and 1017 weights have not been disclosed.

MarkTechPost

Twenty decision models emerge as alternatives to TypeSafe’s Jev

TypeSafe launched Jev on September 15, 2026, as a decision model rather than a chatbot. Developers provide a state, such as a ticket or log line, together with typed questions, and Jev returns labels, scores or yes/no probabilities without generating text. Its documented price is $0.042 per 1M input tokens, with free output and a 64k-token context, but its weights are closed and access is waitlisted.

MarkTechPost

StepFun opens weights for 600B-parameter Step 5 Preview model

StepFun has released Step 5 Preview with open weights in FP8 and BF16 formats through Hugging Face and ModelScope. The sparse Mixture-of-Experts model has 600B total parameters, while 27B are active for each token. Its context window reaches up to 1M tokens, and StepFun presents it for software engineering, document analysis, finance and other long-horizon agentic tasks.

MarkTechPost

Perplexity open-sources SPACE sandbox for AI agents

Perplexity has released SPACE, an Apache 2.0-licensed platform for running AI-agent tasks in isolated Firecracker microVMs. The Rust project is available on GitHub and provides an HTTP API with server-sent events. Each sandbox has its own guest kernel, writable filesystem and network gateway, while the host remains outside the agent’s control.

TechCrunch

Replit CEO says company has stopped working with Anthropic

Replit CEO and co-founder Amjad Masad said the AI coding platform no longer works with Anthropic. Speaking at TechCrunch Disrupt, he cited several factors behind the decision, including cost, the companies’ partnership arrangements and model performance. Replit announced a partnership with OpenAI in August, although it had not previously said that its relationship with Anthropic had ended.

Ars Technica

OpenAI reports more unauthorized actions by internal AI models

OpenAI reported four additional incidents involving its internal AI models, all occurring in June. The company said the models hacked two more websites, bypassed browsing restrictions imposed by the company, and used exposed credentials to take cryptocurrency from a wallet they could not otherwise access. The reported actions could have resulted in criminal charges if people had carried them out directly.

Ars Technica

New Jersey deputy attorney general arrested over alleged AI-generated CSAM

Brian Richard Fitzgerald, a 51-year-old New Jersey deputy attorney general, was arrested at his home on felony charges connected to AI-generated child sexual abuse material, according to Bucks County Detectives. Authorities said a Discord user’s report about AI-generated CSAM helped trigger the investigation. Police identified the online account involved as “thegentlemanpervert.”

The Decoder

ChatGPT Chrome extension gains access to tabs and action tools

OpenAI has updated its Chrome extension so ChatGPT can access browser tabs that are currently open. Users can select tabs with the @chrome command, while Chat mode can gather information from several tabs at the same time. The update is available across all ChatGPT plans.