Good morning. It's Thursday, January 22, and we're covering AI in luxury resale, 10M-token LLMs, new AI constitution, and more. First time reading? Join a community of 600k+ people keeping up with the future of tech. Sign up here.
Don’t keep us a secret: Share this email with a friend. And as always, if you have feedback, we want to hear it: astro@forwardfuture.ai
📍Apple Developing AI‑Powered Wearable Pin
Apple is reportedly working on a new AI‑focused wearable device — a small pin that can attach to clothing and includes multiple cameras and microphones — according to The Information. The device is described as roughly the size of an Apple AirTag, with an aluminum and glass body, two cameras (standard and wide‑angle), three microphones, a speaker, and a physical button.
📜 Claude Publishes New AI Constitution
Anthropic released a new constitution outlining the values and behaviors guiding its AI model, Claude. The document plays a central role in training, shaping how Claude balances helpfulness, safety, ethics, and compliance. Unlike prior rule-based versions, it explains the reasoning behind principles to support better judgment. The full text is released under a CC0 license to promote transparency.
⚡ Stargate AI Campuses Commit To Communities
OpenAI says its Stargate AI infrastructure program is designed to benefit local communities as it scales U.S. capacity toward 10GW by 2029. The company commits to funding its own energy needs, avoiding higher electricity costs for residents. Projects span Texas, Michigan, Wisconsin, and New Mexico, with investments in power, water efficiency, and grid stability.
🤖 Meta’s New AI Models Advance Internally
Meta’s Superintelligence Labs delivered their first major AI models for internal use this month, CTO Andrew Bosworth said at Davos. The team was formed last year after Meta reshuffled AI leadership and invested heavily in talent and infrastructure. Bosworth said the models are “very good” but still unfinished, marking progress after criticism of Meta’s earlier Llama 4 model.
🪧 POWERED BY ADOBE
Adobe Brought Popular Apps Into ChatGPT for Free
Adobe is bringing three of its industry-leading apps — Adobe Photoshop, Adobe Express, and Adobe Acrobat — directly into ChatGPT.
With Adobe apps for ChatGPT, anyone can edit photos, create designs, and transform PDFs in the most intuitive way possible: using their words inside the chat.
Adobe apps for ChatGPT expands Adobe’s reach to one of the world’s most popular conversational AI platforms with over 800 million weekly users, introducing Adobe’s category-defining tools to people who may not have used Adobe’s apps before — through a surface they already use every day.
📺️ FROM THE LIVE SHOW
“For years, the best engineer and the worst engineer still spent ~60% of their time on the same grunt work.”@levie, CEO of @Box, on why AI makes experts more powerful:“If I’m a really good engineer, I want AI to do all the stuff I hate doing.”“I want to spend my time on… pic.twitter.com/GmnEKWIdIG
@ForwardFutureView post on X📽️ VIDEO
AI Just Killed Another Company…
Stack Overflow’s fall shows how AI reshaped coding—LLMs trained on its data replaced it, turning a developer staple into a tech-era casualty.
🛒 AUTOMATION
The RealReal Deploys AI To Speed Luxury Authentication and Cut Costs
Luxury resale platform The RealReal is deploying an AI system called Athena to accelerate how consigned goods are authenticated, described, priced, and listed for sale at its New Jersey warehouse. Athena analyzes images, measurements, and historical data to assign counterfeit-risk scores, allowing low-risk items to bypass manual checks while routing high-value goods—such as Louis Vuitton totes or Hermès Birkin bags—to senior authenticators.
As of late September, Athena handled 27% of incoming items, and executives expected that share to approach 40% by the end of 2025, though final figures were not disclosed ahead of earnings. Company leaders say the system could cut listing timelines by up to 50%, helping items reach buyers faster and improving seller retention. Over time, Athena is also expected to reduce labor needs and save millions in operating costs as resale volumes continue to grow. → Read the full article here.
🧠 REASONING
MIT Framework Lets LLMs Handle 10M Tokens Without Context Rot
Researchers at MIT CSAIL have introduced “recursive language models” (RLMs), a framework that allows large language models to reason over prompts exceeding 10 million tokens without expanding their context windows or retraining. Instead of ingesting long prompts directly, RLMs treat text as an external environment that the model inspects and processes programmatically, pulling in only relevant snippets as needed.
The approach reframes long-context reasoning as a systems problem, borrowing ideas from out-of-core computing used for datasets larger than memory. In benchmark tests, RLMs dramatically outperformed base models on long-context tasks, avoiding the performance degradation known as “context rot.” Because the framework wraps existing models, it can function as a drop-in replacement for enterprise applications handling massive documents or codebases. → Read the full article here.
🛰️ NEWS
What Else is Happening
🐌 Dimon Urges Slower AI Rollout: JP Morgan boss warns rapid automation could spark unrest without retraining, while NVIDIA’s Jensen Huang says AI will create more jobs.
📈 Anthropic Revenue Surges Past $9B (Paywall): The AI startup’s run rate more than doubled since mid-2025, fueling a new funding round reportedly valuing it near $350 billion as major VCs pile in.
👥 YouTube Adds AI Creator Likeness: Creators will soon make Shorts using AI versions of themselves, as YouTube boosts its 200-billion-views-a-day format while adding safeguards against misuse.
👷 Huang Touts AI Job Boom (Paywall): NVIDIA CEO says massive data-center buildouts will drive six-figure pay for electricians and plumbers, arguing AI fuels skilled trade demand, not just tech jobs.
🦾 Microsoft’s AI Robotics Model: Rho-alpha is a model designed to translate natural language into real-world robotic actions. Built on the Phi vision-language models, Rho-alpha adds tactile sensing and supports bimanual manipulation tasks.
🎨 Comic-Con Bans AI Art: San Diego Comic-Con reversed its policy after artist backlash, fully barring AI-generated works from the art show amid fears generative tools threaten creative jobs.
🛠️ Apple Rebuilds Siri as Chatbot (Paywall): Apple plans a generative AI overhaul, replacing Siri with a system-wide chatbot on iPhone, iPad, and Mac to compete with OpenAI and Google.
🧰 TOOLBOX
Trending AI Tools
📝 Fireflies: AI notetaker that records, transcribes, summarizes meetings, and analyzes team conversations.
🌌 Astrocade: Explore the universe with powerful astronomy tools, rich space data, and collaboration features.
📚 Coursebox AI: AI platform that automates course creation, boosts learner engagement, and manages content.
🗒️ FEEDBACK
Help Us Get Better
What did you think of today's newsletter? |
That’s a Wrap!
❤️ Love Forward Future? Share your unique referral link with friends and colleagues to unlock exclusive Forward Future perks! 👉 Get your link here.
📢 Want to advertise with Forward Future? Reach 600K+ AI enthusiasts, tech leaders, and decision-makers. Let’s talk—just reply to this email.
🧑💻 The AI-Fluent Job Board: Hire talent that understands AI. Find jobs that actually value your AI skills. → Learn more.
Thanks for reading today’s newsletter—see you next time!
Matthew Berman & The Forward Future Team
🧑🚀 🧑🚀 🧑🚀 🧑🚀
