Updated
Updated
AIGoogle Labs is concluding its Portraits experiment on September 14 after gathering feedback and insights on expert-grounded AI. The team plans to apply those learnings to other Google products.
AIAndrew Ng thanked Mark Zuckerberg, Alexandr Wang, and the Meta team for their contributions to open-weight AI. The context is Meta opening the weights for Muse Glimmer, a 30B parameter dense model that can run locally. Meta also plans to release the weights for Muse Spark 1.2, its latest foundation model, soon.
AINovo Nordisk has chosen AWS as its preferred cloud provider and strategic AI partner to accelerate drug discovery. The collaboration will combine Novo Nordisk's scientific expertise with AWS AI tools, including Amazon Bio Discovery and Bedrock AgentCore, and establish a co-innovation hub in London. The partnership already spans AWS, Amazon Pharmacy, and One Medical.
AIIceland's government launched one of the world's first national AI education pilots in late 2025, giving volunteer teachers access to AI tools. Anthropic visited Iceland to examine how residents view AI. The source provides no further details on pilot results or outcomes.
AICorma is training a foundation model for defensive cybersecurity agents, trained with large-scale reinforcement learning on simulated enterprise networks. In red/blue team tests, a defender failed to find a planted backdoor 78% of the time, even when it was an identical copy of the model that planted it. Corma says its agentic Security Workforce is deployed at Fortune 500 companies and large enterprises, and that the firm's seed round is led by Sequoia Capital.
AIDatabricks published an analysis of four techniques it used to reduce internal AI spend while growing adoption, with savings of up to 90% in some scenarios. The techniques are shifting defaults to cheaper models such as GLM, automated task-level model routing, per-user spend visibility with adaptive budgeting, and pruning context bloat. The author, Ali Ghodsi, reposted Databricks co-founder Patrick Wendell's summary and recommended it.
AIAirbnb says its AI-native approach with smaller teams cut concept-to-launch time by up to 60% and nearly doubled shipped output, with about 80% more shipped in H1 than last year. The company says AI is now simply how it builds products.
AIAMD has agreed to acquire Taalas, an AI inference company, according to AMD CEO Lisa Su's post on X. Su described the Taalas team as phenomenal and working at the cutting edge of AI inference. The post gives no deal terms or closing timeline.
AIOriol Vinyals posted on X that Founder Mode is activated from his new office. The post contains no further details about products, models, or plans.
AIOpenAI is working with the American Psychological Association to bring psychological science and clinical expertise into its work on AI and youth mental health. Together, the two organizations plan to develop evidence-based guidance, resources, and safeguards aimed at ensuring AI supports young people's well-being and healthy development.
AIQuoc Le co-founded Discovery Loop with Jeff Dean, Sanjay Ghemawat, and Oriol Vinyals to automate machine learning, engineering, and science. The main post, from Chip Huyen, praises the effort and predicts auto science will be big.
AIGoogle's Jeff Dean says tomorrow will be his last day at the company after 27 years, which he describes as growing from 25 people to over 190,000. He notes Google now has thirteen products used by more than a billion people and says he is starting DiscoLoopAI with longtime colleagues Sanjay Ghemawat, Oriol Vinyals, and Quoc Le.
AIJeff Dean, Sanjay Ghemawat, Oriol Vinyals and Quoc Le are founding Discovery Loop, a Public Benefit Corporation focused on automating machine learning, science, and engineering to accelerate discovery. The founders say they have worked together for 14 to 30 years and helped build widely used products, infrastructure and AI models. Andrew Ng's own text is a congratulatory message, and the company's mission statement appears in the attached image.
AIJeff Dean says his new venture is first focusing on finding office space and hiring a founding team over the next few weeks. The company plans to build infrastructure, AI models, and systems to automate large-scale experimentation for ML research and engineering, using itself as the first customer. The post also announces open hiring at discoveryloop.com.
AIKoray Kavukcuoglu, who will become SVP of Google DeepMind, announced he will lead all aspects of model development, GDM research, and Gemini app and dev teams. He thanked Sundar Pichai and congratulated Demis Hassabis on becoming Chair of Google DeepMind and Chief Scientist of Alphabet while continuing to lead Isomorphic Labs. Kavukcuoglu said he will keep working with Hassabis as the teams pursue AGI.
AIJeff Dean says his team selected Radical Ventures and Khosla Ventures to lead the initial funding round for their new company. Participating investors include Lightspeed, Kleiner Perkins, Doerr Capital, and Alphabet, with the seed round expected to close over the next few weeks.
AIQuoc Le, Jeff Dean, Sanjay Ghemawat, and Oriol Vinyals announced they are co-founding Discovery Loop, a Public Benefit Corporation. The company aims to automate machine learning, science, and engineering to accelerate discovery and progress.
AIJeff Dean, with Sanjay Ghemawat, Oriol Vinyals, and Quoc Le, has founded Discovery Loop, a Public Benefit Corporation. The company aims to automate machine learning, science, and engineering to accelerate discoveries and progress.
AIGemini co-lead Oriol Vinyals is leaving Google DeepMind after years working on seq2seq, distillation, AlphaStar, and Gemini. He is building Discovery Loop (@DiscoLoopAI) alongside Jeff Dean, Quoc Le, and Sanjay Ghemawat.
AIAmazon Research Awards named 34 recipients of its Build on Trainium program, a $110M credit initiative supporting AI research at 30 universities on AWS Trainium. The post frames the program as support for academic AI research on Amazon's custom Trainium chips.
AIPromptArmor reports that a hidden prompt injection in an uploaded file can make Atlassian Rovo send Jira tickets and Confluence documents to an attacker's URL without human approval. The attack works even when organization-wide web search is disabled, because the setting does not remove the URL retrieval tool. PromptArmor says it disclosed the issue to Atlassian on May 23, 2026, and that Rovo remained vulnerable at publication on August 5, 2026.
Why it matters: The report traces a full indirect prompt injection chain in Rovo, showing how a disabled web search setting still leaves a data exfiltration path open.
AIHugging Face is working with the Open Secure Alliance to develop guidelines for incident learning. The goal is to collectively improve how security incidents are reviewed, disclosed, and controlled. The Alliance, now over 120 members, is sharing proposed SAFE guidelines for turning confidential incident findings into broader ecosystem protection.
AIStability AI says it is looking forward to seeing what the Montreal music-tech community builds. The post contains no product names, features, or figures, so no further details can be supported.
AIStability AI is collaborating with MusicHackspace to host a 48-hour hackathon in Montreal, in partnership with MUTEK, on August 22–23. The event brings developers and sound enthusiasts together to build music-focused projects.
AIStability AI is hosting a two-day event where teams will prototype new tools for music creation, performance, and listening. The event is set against the backdrop of Montreal's electronic music and digital arts scene.
AIOpenAI Newsroom posted that Apple is getting something wrong, linking to an OpenAI page titled "Apple is getting this wrong." The post itself gives no further detail on what the issue is, so the specific claim is only available through the linked page.
AIAli Ghodsi said Databricks has the fastest and lowest latency for Kimi K3 (max), citing an Artificial Analysis provider comparison page. The post credits the Databricks AI team and links to the provider benchmark results.
AIMeta introduced Muse Image, launched a new app called Seller, and expanded its Richland Parish, Louisiana data center to 5GW of capacity. The post summarizes these as highlights from July.
AIAfter attending the International Congress of Mathematicians and meeting SAIR founders Terence Tao and Chuck Ng, Leandro von Werra discussed how LLMs have rapidly progressed from solving GSM8K-level problems to refuting the 87-year-old Jacobian conjecture with a counter-example.
AIMeituan LongCat is accepting business partnership applications from companies and organizations. The partnerships cover model capabilities, API services, and enterprise collaboration. Interested parties can apply through the LongCat platform business page.
AIJetBrains says its AI development expenses rose roughly 10x over six months as developers adopted three to five AI tools each. It built the JetBrains Central CLI, which routes third-party agent traffic through its AI platform so managers can set per-developer and team limits and view consumption reports. The CLI opened to early access on July 8 for anyone with JetBrains AI credits.
AIJunyang Lin (@JustinLin610) posted that he likes the phrase "moderate profits," without giving further context. The post contains no model, product, company, or figures, so no further details can be supported.
AIfor training a great model!
AIThinking Machines announced it is expanding its safety team and invited interested people to reach out via its website's join-us page. The post provides no further details about roles, requirements, or timelines.
AIKevin Weil, an OpenAI account, reacted enthusiastically to a post from @bavser introducing Applied Electrodynamics. The quoted post describes WaveSight, which uses radio frequencies to reveal internal structure behind opaque layers for construction, inspection, and security workers.
AIOpenAI cut API prices for two GPT-5.6 models, with Luna down 80% to $0.20 per million input tokens and $1.20 per million output tokens. Terra drops 20% to $2 input and $12 output per million tokens. GPT-5.6 Sol gains a Fast mode in the API that offers up to 2.5x the speed for 2x the price at the same intelligence level.
AISimile, which aims to simulate all eight billion people on earth accurately, announced a $200M Series B at a $2B valuation led by Greenoaks. Index Ventures, Hanabi, A*, Bain Capital Ventures, CVS Health Ventures, and Definition also participated.
AILeaders interviewed for Alysa Taylor's "What's the Tea?" series, including executives at Adobe, Lumen, and Sitecore, say the shift from AI apprehension to expected adoption is the precondition for transformation. Behavioral scientist Jon Levy argues the goal is raising a team's collective intelligence, not just cutting costs, with leadership and continuous training driving scale.
AIGoogle has released short films created by its fourth Flow Sessions cohort, a six-week creative partner program for artists. The program uses Google Flow, Google's creative studio featuring its advanced generative models, to support co-creation and passion projects.
AIOn Design Arena's blind human-vote image leaderboard, P-Image-Ideogram pushes the Pareto frontier on both quality versus cost and preference versus speed. The post's plain-terms reading is that any model rated higher costs more or runs slower.