Sign Up | Collaborate with us Welcome back! Google bundled Gemini, Wiz, Mandiant and CodeMender into one autonomous security platform and the cybersecurity sector lost billions in market cap inside a session. Meta started putting price tags on Instagram, Facebook, WhatsApp and Meta AI. Amazon MGM greenlit three AI-assisted animated series with five-week pilot deadlines. OpenAI's non-profit committed $250M to study the labor damage caused by the technology its for-profit sells.
In today's Generative AI Newsletter: Google: Why is it a terrible week for cybersecurity companies?
Claude learned to check its own code for vulnerabilities Google Cloud launched Google AI Threat Defense, an always-on platform that fuses Gemini, Wiz, CodeMender and Mandiant into a single stack. And it just happens that Zscaler dropped 31%, CrowdStrike fell 3.4%, Palo Alto Networks lost 3% and the First Trust Cybersecurity ETF closed down 2.7%. The details: Gemini handles reasoning, Wiz maps live exposure across cloud and code, CodeMender writes patches inside the developer's IDE, Mandiant supplies the response playbooks. No single frontier model finds every vulnerability, so AI Threat Defense routes the heaviest scans to internet-facing apps and authentication logic. CodeMender generates the fix and the verification tests automatically before anything ships.
Zscalerās 31% drop also tracked an earnings cut and a below-consensus FY27 guide released the night before Google's keynote. Accenture, Deloitte, PwC and Netenrich will deploy and manage the platform for enterprise customers. The exploit window has collapsed from weeks to hours and security teams cannot keep up by hand. The price the market put on that capability is whatever Zscaler, CrowdStrike and Palo Alto can no longer charge for the same work.
How to own an early piece of winning AI products 95% of enterprise AI projects produce no measurable revenue, according to MIT's 2025 GenAI Divide report. Distribution is what separates the few that scale from the ones that stall. GenAI Works has products that already earn revenue and 14M+ in owned distribution behind them. AI Academy generated $12K in its first month. 30,000+ students. 50+ courses and 30+ guides.
The Hackathon app already powers events for Oracle, Red Bull and G42 alongside our own. Company revenue grew from $1M in 2024 to $2.5M in 2025. That's 2.5x in a year. Enter from just $1,000. Commit before July 1, 2026 to earn up to 25% bonus shares. Reserve your share today In making an investment decision, investors must rely on their own examination of the issuer and the terms of the offering, including the merits and risks involved.
Works, Inc. has filed a Form C with the Securities and Exchange Commission in connection with its offering, a copy of which may be obtained here Meta puts a price tag on its AI features Meta confirmed it is rolling out paid subscription tiers across its core apps. Instagram Plus and Facebook Plus will cost $3.99 a month. WhatsApp Plus will cost $2.99. Meta AI gets two paid tiers at $7.99 and $19.99 a month. The economic angle is that Meta's ad business no longer covers the AI buildout the company has committed to.
Reality Labs losses, Llama training runs and the Wang-era model lab all have to be paid for, and Wall Street has been asking for a second revenue line since the Cambridge Analytica years. The open question is whether anyone actually pays $3.99 a month for whatever Instagram Plus turns out to mean. Amazon MGM greenlit three AI animated series Amazon MGM Studios launched the GenAI Creators' Fund and ordered three animated series for Prime Video, all built through Project Nara, a new AI production platform Amazon developed jointly with AWS. Punky Duck: from Jorge Gutierrez, writer-director of "The Book of Life." Love Diana Music Hunters: from Albie Hecht, the former Nickelodeon president who worked on "SpongeBob SquarePants." Cupcake & Friends: from BuzzFeed Studios.
Casting: all three projects use human actors and voice actors. Project Nara is the product Amazon is selling and the animated series serving as proof points to convince other studios to adopt it. Is the five-week pilot enough to produce anything watchable? Weāll have to be the judge of that.
OpenAI's non-profit funds AI layoffs research The OpenAI Foundation, the non-profit that holds a 26% stake in OpenAI's for-profit entity valued at $130B at restructuring, committed $250M for grants, partnerships and direct programs aimed at workers facing AI displacement. Grants will fund labor market research, support communities facing near-term displacement and explore new models for distributing the economic gains from AI. Some programs will run directly through the Foundation rather than via outside non-profits, and AI-powered simulations of how economies might evolve as the technology matures are explicitly on the list. Block and Standard Chartered both cited AI efficiencies for layoffs this month. The same week OpenAI is announcing $250M for displacement research, two of its enterprise customers used the technology to justify cutting headcount.
This comes shortly after Sam Altman said the āAI job apocalypse" isnāt real. The upcoming IPO and this announcement might be two of the reasons behind this statement. Tool of the Day: Cleanlab Cleanlab is the open-source Python library that finds the mislabeled rows, outliers and near-duplicates inside your training data before they reach production. Handshake AI acquired the company in January 2026 and steered the commercial product toward enterprise AI guardrails.
The cleanlab Python library on GitHub stayed open-source under an Apache 2.0 license and still does the data quality work it was built for. Try this yourself: Install with pip install cleanlab on Python 3.10 or newer. Get class probability predictions from your model on a labeled dataset. Tabular, text, image and audio all work.
Run find_label_issues to surface the rows most likely to be mislabeled, ranked by confidence score. Use the same package to flag outliers and near-duplicates in the same pass. ML engineers, data scientists and anyone running LLM evals against annotated data get the most out of it. Everything else you shouldn't miss Amazon and Snowflake signed a $6B five-year deal for AWS Graviton CPUs to power agentic compute, with Snowflake jumping as much as 35% on the news.
OpenAI and Thrive Holdings shipped Tax AI to Crete's network of 30+ accounting firms, where one senior accountant who spent 180 hours on tax prep last year spent 15 hours this season. TropiCam-AI identifies 84 taxa of tree-dwelling species with 95% accuracy from camera-trap footage, finally giving canopy ecologists the model their ground-level counterparts. KPMG signed Anthropic as its AI provider and is rolling Claude out to all 276,000 employees across 138 countries by September. Learn more about AI from the experts building it šø Follow us on Instagram for fast, visual AI updates in 30 seconds. š Subscribe to our Atlas newsletter ā trusted by 3M+ subscribers ā to stay ahead of AI news across tech, education, and business. šŗ Watch us on YouTube to hear insights directly from leading AI voices, builders, and innovators. š¦ Follow us on X for breaking AI news and real-time industry updates. š§ Learn how to build your next AI application with practical resources and expert guidance. š Explore investment opportunities in the future of AI and join our community-backed growth journey.
Get in touch.