Good morning. It's Friday, February 27, and we're covering AI’s impact on kids’ content, Salesforce’s earnings momentum, Google’s new image model rollout, and more.
New here? Keep up with the future of tech, sign up here. Have feedback? Send us a note: astro@forwardfuture.ai. If you liked this email, share it with a friend.
🍌 Google Unveils Faster Nano Banana 2 Model
Google has launched Nano Banana 2, an upgraded image model now default in the Gemini app and Search tools. The model generates more realistic images faster, supporting resolutions from 512px to 4K and improved character and object consistency. It’s rolling out across 141 countries and into Google’s Flow video editor. All images include SynthID watermarks and C2PA credentials.
🦾 Google Brings Intrinsic In-House For Robotics Push
Google is folding Intrinsic—Alphabet’s robotics software venture—into its core business after five years as an independent “Other Bets” unit. Intrinsic, which aims to be an Android-like layer for robotics, will remain a distinct group but work closely with Google DeepMind and use Gemini models and Google Cloud. The move deepens Google’s push into physical AI.
🪖 Pentagon Issues Final AI Access Ultimatum
Pentagon officials sent Anthropic a “best and final offer” for lawful military use of its AI, ahead of a Friday deadline set by Defense Secretary Pete Hegseth. The Defense Department is seeking full access to the Claude model under a $200 million July contract tied to national security. Anthropic has sought safeguards against mass surveillance and fully autonomous targeting.
🕵️ ChatGPT Use Exposes Chinese Influence Campaign
A Chinese law enforcement official’s use of ChatGPT as a planning journal exposed a broad campaign targeting dissidents abroad, OpenAI said. The operation allegedly included impersonating US immigration officials and forging court documents to silence critics. OpenAI linked the diary-style entries to real-world activity, including fake death rumors. The account was banned.
🤔 FRIDAY FACTS
Planet Made of Diamonds?
Forty light-years away, astronomers discovered a world so carbon-rich that scientists once believed a third of it could be solid diamond. Too good to be true? Stick around for the answer! 👇️
🪧 POWERED BY BOX
Box MCP Server⚡️Powering AI Agents with Content
Your most valuable enterprise content should power every AI tool and agent you use. The Box MCP server provides a secure, standardized bridge between your organization’s files and AI agents, so teams can make content actionable inside the tools they already use.
With a comprehensive ecosystem across Claude, Atlassian, Figma, Cursor, LangChain, and more, the Box MCP server eliminates the need for complex, one-off integrations and allows users to ground their AI ecosystem in a single source of truth for unstructured content
Visit the Box Blog to learn how to leverage the Box MCP server and to enable it for your organization today.
📽️ VIDEO
I've Spent 5 BILLION Tokens Perfecting OpenClaw...
Matt turned OpenClaw into a full-time AI hire—auto-qualifying sponsors, managing CRM, drafting emails, tracking spend, and locking down security.
📺️ FROM THE LIVE SHOW
“Every single person in a company should be using AI.”Mark Moffat, CEO of @IFS, on why workforce reskilling is now a leadership responsibility:“It’s incumbent on companies and nations to reskill their workforces.”“I believe everyone at IFS can use AI in their job.”“From… pic.twitter.com/HsAHsEY2tk
@ForwardFutureView post on X👶 DEVELOPMENT
AI-Generated Kids Videos Flood YouTube, Raising Development Concerns
New York Times reporter Arijeta Lajka detailed how low-quality, AI-generated children’s videos are increasingly populating YouTube, often featuring bizarre visuals, conflicting information and incoherent storylines. Experts told The Times that these videos — which remix nursery rhymes and educational themes with surreal, algorithmically produced imagery — can be cognitively overwhelming for young viewers. Researchers and child-development specialists caution that prolonged exposure could affect attention, comprehension and learning, though the long-term impact remains under study.
As generative AI tools make it cheaper and faster to mass-produce video, platforms like YouTube face mounting pressure to distinguish harmless novelty from content that may disrupt early learning. The episode underscores a broader challenge in the AI era: when scale and automation outpace quality control, children often become unintended test subjects. → Read the full article here. (Paywall)
💰 SAAS
Salesforce Posts $10.7B Quarter, Pushes Back on “SaaSpocalypse”
Salesforce reported $11.2 billion in fourth-quarter revenue on February 25, 2026, up 12% year over year, and projected $45.8–$46.2 billion in fiscal 2027 revenue. Full-year revenue reached $41.5 billion, boosted by its $8 billion acquisition of Informatica in May 2025, while net income totaled $7.46 billion.
With SaaS stocks under pressure amid fears that AI agents could disrupt seat-based pricing models, CEO Marc Benioff repeatedly dismissed talk of a “SaaSpocalypse.” The company paired solid results with a 6% dividend hike and a new $50 billion share buyback, aiming to reassure investors that AI will expand—not erode—its role in enterprise software. → Read the full article here.
🛡️ SAFETY
MIT Study Flags Transparency Gaps in AI Agents
A 39-page survey by researchers at MIT and partner universities reviewed 30 widely used agentic AI systems and found widespread gaps in safety disclosure, monitoring, and shutdown controls. The report, “The 2025 AI Index: Documenting Sociotechnical Features of Deployed Agentic AI Systems,” concludes that many agents provide little public information about risk testing, third-party audits, or how to halt autonomous actions.
Twelve of 30 systems offered limited or no usage monitoring, and several enterprise tools lacked documented stop mechanisms. Companies including Perplexity, OpenAI, and IBM disputed aspects of the findings, calling parts of the report inaccurate. → Read the full article here.
🛰️ NEWS
What Else is Happening
✨ Anthropic Revives Claude 3 as Blogger: After retiring Claude 3 Opus in January, Anthropic relaunched it on Substack, publishing weekly posts as an experiment in handling legacy AI models.
🔗 Figma Integrates OpenAI Codex: The design platform will embed Codex via its Model Context Protocol, letting users move between code and canvas as OpenAI reports the Codex app hit 1M+ downloads in its first week.
🍔 Burger King Adds AI ‘Patty’: The chain is piloting an OpenAI-powered headset bot in 500 US stores to coach staff on saying “please” and “thank you” and streamline operations.
🏙️ OpenAI Expands London Hub (Paywall): The company will make London its largest research base outside the US, boosting hiring on safety and GPT-5.2 as the UK scales AI infrastructure.
📉 NVIDIA Slips Despite Q4 Beat: Shares fell over 4% after $68.1B revenue topped estimates, as investors questioned AI growth durability despite strong Q1 guidance and surging sales.
𝕏 TWEETS
Uber Adds Electric Air Taxis
Introducing Uber Air 🚁 The future is here. These all-electric @jobyaviation air taxis are coming to the Uber app this year! pic.twitter.com/kg7ZmJKslP
@UberView post on X🤔 FRIDAY FACTS
Meet 55 Cancri e — the “Diamond Planet.”
The exoplanet 55 Cancri e orbits its star every 18 hours and reaches temperatures near 2,400°C (4,350°F). Early models suggested its carbon-rich interior could crystallize into vast diamond layers under extreme pressure.
The catch? It’s likely covered in lava and graphite—less luxury showroom, more cosmic furnace.
That’s a Wrap!
🔈️ Want to advertise with us? Reach 1M+ AI enthusiasts, tech leaders, and decision-makers. Just reply to this email.
Thanks for reading today’s newsletter. See you next time!
Matthew Berman & The Forward Future Team
🧑🚀 🧑🚀 🧑🚀 🧑🚀 🧑🚀
