Revue de presse générée automatiquement avec
AI Product manager
A seasoned AI professional driving business growth through strategic AI adoption, with a focus on developing and managing AI products, infrastructure, and governance frameworks that balance innovation with regulatory compliance. They prioritize staying updated on the latest AI models, capabilities, and MLOps advancements.
Ai strategy (20%)Models & Capabilities (20%)Ai Infrastructure & MLOps (20%)Generative AI (20%)User Experience Design (20%)
Vous souhaitez recevoir chaque jour la revue de presse de ce profil ?
AI strategy shifts, generative models, and infrastructure advances...
Jeudi 11 décembre 2025 à 10:50
AI Strategy & Governance
Meta pivots to a proprietary “Avocado” model
Meta is quietly developing a new large‑scale model nicknamed “Avocado,” signaling a stark departure from its earlier open‑source stance. Both Engadget and CNBC report that the model will likely be closed‑source, a move driven by safety concerns and the stalled rollout of Llama 4, while Bloomberg notes the shift follows internal leadership changes within Meta’s AI Superintelligence Labs. The decision underscores a broader industry tension between openness and competitive advantage.
Engadget
OpenAI taps Slack’s Denise Dresser as chief revenue officer
In a bid to accelerate enterprise adoption, OpenAI has hired Denise Dresser, former Slack CEO, as its chief revenue officer. The Verge details how Dresser will steer global revenue strategy and embed AI more deeply into day‑to‑day business workflows, leveraging her experience launching AI features at Slack. The appointment reflects OpenAI’s push to monetize its models beyond consumer APIs.
The Verge
U.S. attorneys‑general demand tougher AI safeguards
A coalition of state attorneys‑general has issued a formal warning to major AI firms, urging stronger safeguards against harms such as misinformation and bias. Financial Times outlines the pressure on companies like OpenAI and Google to adopt stricter governance, while the move dovetails with broader regulatory scrutiny of generative AI.
Financial Times
Oracle’s costly AI‑focused data‑center expansion rattles investors
Oracle disclosed a $10 billion cash burn in its latest quarter, driven by an aggressive rollout of AI‑centric data centers. Both Financial Times and MarketWatch note that despite a 68 % surge in cloud revenue, the steep capital‑expenditure outlook spooked Wall Street, sending the stock down 11 %. The episode highlights the financial risks of rapid AI infrastructure scaling.
The Information
Market Watch
Generative AI in Specialized Domains
Inito leverages AI‑engineered antibodies for at‑home testing
Health‑tech startup Inito has secured $29 million to develop AI‑designed antibodies, aiming to broaden its portfolio of at‑home diagnostic tests. TechCrunch explains that the models accelerate antibody discovery, promising higher accuracy and new test categories, illustrating generative AI’s expanding role in biotech product pipelines.
TechCrunch
Models & Capabilities
Mistral AI unveils Devstral 2 and the Vibe CLI for autonomous coding
French AI firm Mistral AI released Devstral 2, a 123‑billion‑parameter open‑weights coding model that tops the SWE‑bench Verified benchmark with a 72.2 % score. Ars Technica also highlights the accompanying Vibe command‑line interface, which maintains project context across files and can execute shell commands autonomously, offering developers a powerful new tool for AI‑augmented software engineering.
Ars Technica
User Experience Design
Figma adds AI‑powered erase, isolate, and expand tools
Design platform Figma launched three new AI features—Erase, Isolate, and Extend—that let users edit images directly within the design canvas without switching apps. The Verge reports that the tools use generative AI to seamlessly remove objects, separate subjects, and enlarge canvases, streamlining creative workflows for designers.
TechCrunch
Tavus’s AI Santa creates hyper‑personalized holiday chats
Startup Tavus introduced an AI‑Santa experience that remembers user interests and asks tailored questions, encouraging conversations that can last “hours per day.” TechCrunch notes the novelty of the persistent persona and its potential to deepen brand engagement through conversational UI design.
TechCrunch
AI Infrastructure & MLOps
Intel’s llm‑scaler‑vllm 1.2 adds Arc GPU support for new LLMs
Intel released version 1.2 of its llm‑scaler‑vllm Docker container, now optimized for Arc Graphics hardware and supporting a broader range of large language models. Phoronix highlights the update as a step toward more accessible, GPU‑accelerated inference pipelines for enterprises adopting AI workloads.
Phoronix
Baseten acquires Parsed to boost continual‑learning capabilities
Inference provider Baseten announced the acquisition of Parsed, a startup specializing in reinforcement‑learning fine‑tuning of open‑source models for niche tasks. The Information explains that the deal will enable Baseten’s customers to more easily implement continual learning pipelines, reinforcing the trend toward adaptable, task‑specific AI services.
The Information
Aller aux sources
11 sources citées