The core tension in AI today is between OpenAI's closed-ecosystem, API-centric business model and Hugging Face's open-repository model that allows developers to download and run models locally, accord…
Knowledge distillation transfers reasoning from a large teacher model like GPT-4 or Claude 3.5 Sonnet to a smaller student model such as Llama 3 8B, aiming for 90%+ of the teacher's performance while …
Security researchers are using prompt engineering techniques to bypass AI guardrails and generate exploit code for vulnerability discovery, according to a technical guide. The approach involves contex…
An engineer built LLM Latency Tracker, an independent, provider-neutral tool that measures AI API latency and uptime from four global regions. The open-source project covers ~45 providers and is desig…
Mixture of Experts (MoE) is a machine learning architecture that divides a model into specialized subnetworks called experts, with a gating network routing each input token to only the most relevant s…
Open source AI models like Llama 3 and Mistral give developers full control over deployment, avoiding reliance on third-party APIs that may change behavior or pricing, according to a technical guide. …
Microsoft will fund Mistral's European AI expansion in a multibillion-dollar deal, the companies announced on July 21, 2026. The investment aims to accelerate Mistral's development of large language m…
Microsoft released Fara1.5-27B, a vision-only browser agent fine-tuned from Alibaba's Qwen3.5-27B, signaling that US cloud players find Chinese base models strong for specialized research. AMD is inve…
France must build 12GW of compute capacity under a proposed 'Plan Prométhée' to create a sovereign frontier AI lab within three years, according to an analysis prompted by U.S. export controls on Anth…
AMD announced a $5 billion investment in Anthropic, while Microsoft released Fara1.5-27B, a vision-only browser agent fine-tuned from Alibaba's Qwen3.5-27B model. The US government formally accused Mo…
Upstage, a Korean AI company, released Solar Open 2, an open-weight large language model with 250 billion parameters that activates only 15 billion at a time using a mixture-of-experts structure, desi…
A study by researchers on arXiv (2607.20301v1) finds that PortLLM, a training-free and data-free scheme for adapting large language models after continual pretraining, maintains long-term temporal por…
Cory Doctorow, the tech gadfly who coined 'enshittification,' called the push for sovereign AI 'nonsense' in a Wednesday post, arguing that countries should instead focus on building sovereign apps an…
Arrakis, a London and Paris startup building an AI operating system for industrial companies, emerged from stealth with $38 million in total funding and a thesis that the biggest AI returns will come …
Andreessen Horowitz (a16z) announced the inaugural cohort of 65 Forward Deployed Engineer Fellows, an 8-week program for engineers deploying AI in enterprises, with fellows from companies including Op…
Staan has launched the first European Search API, offering web search, news, and AI-powered answer features with pricing up to 9 times cheaper than competitors like Google, Bing, and Brave. The API pr…
A new analysis argues that the open-source versus proprietary AI debate has calcified into a lazy narrative, with many organizations reflexively choosing open-weight models for data sovereignty withou…
A developer built Mistral.Agents.Net, an open-source .NET library that wraps Mistral's Agents API to enable enterprise multi-agent workflows. The library supports persistent agents, stateful conversat…
OpenAI is launching OpenAI Presence, a new product that helps companies deploy AI agents by connecting them to internal data, policies, and workflows, aiming to automate customer support, sales, and o…
Samsung is in talks to invest up to one billion euros in French AI startup Mistral, which would push the company's valuation to around 20 billion euros, according to The Decoder.…