🤔 FRIDAY FACTS
Can an AI read your mind?
In 2019, scientists made headlines with a system that seemed to do just that, reconstructing full sentences from human brain activity. But how does it work, and just how accurate was it?
Stick around to find out! 👇
👔 Sutskever Becomes CEO of Safe Superintelligence
OpenAI co-founder Ilya Sutskever has stepped in as CEO of Safe Superintelligence following Daniel Gross’s departure amid Meta acquisition talks. The $32B startup remains focused solely on building safe superintelligence, with no side products. Sutskever will lead both the company’s technical direction and broader strategic vision.
🎬 Google Launches Veo 3 Video Gen Worldwide
Google has rolled out its Veo 3 video generation model to Gemini users in over 159 countries. Available only to AI Pro plan subscribers, the tool lets users generate up to three eight-second videos per day using text prompts. Google also plans to add image-to-video generation, expanding Veo 3’s creative and storytelling capabilities.
🔲 CoreWeave First to Deploy NVIDIA Blackwell Ultra
CoreWeave is the first cloud provider to deploy NVIDIA’s Blackwell Ultra AI chips, installed in Dell’s GB300 NVL72 systems with 72 GPUs and 36 Grace CPUs. The chips promise 50x more AI content than the previous generation. Backed by NVIDIA and recently public, CoreWeave gains a major edge in the race to power next-gen AI workloads.
🪙 OpenAI Disavows “Equity Tokens” on Robinhood
OpenAI has publicly distanced itself from “OpenAI tokens” being sold on Robinhood, clarifying that they are not official equity and were issued without its involvement or approval. The tokens allegedly offer indirect exposure via a special purpose vehicle, but OpenAI warns they don’t confer shareholder rights. Investors are urged to proceed with caution.
Enjoying our newsletter? Forward it to a colleague—
it’s one of the best ways to support us.
🎮 GAMING
Google Teases Gaming Potential of Veo 3
Google’s AI leaders recently hinted that Veo 3—its latest video-generating model—could one day power interactive games, sparking excitement about the future of playable world models. While Veo 3 excels at generating realistic video and audio using simulated physics, it's still a passive system, not yet capable of real-time interaction or environmental prediction.
That’s where Google’s other efforts come in: Genie 2, unveiled by DeepMind last year, can generate playable 2D worlds, and the company has formed a dedicated team to explore real-world simulation through AI. A hybrid of Veo’s cinematic realism and Genie’s interactivity could reshape how games are made. With rivals like Microsoft and OpenAI also advancing in this space, the race to build AI-driven game engines is officially on. → Read the full article here.
🫂 COMPANIONSHIP
Portola’s “Tolans” Rethink AI Companionship by Ditching Romance and Nudging Reality
Portola’s AI chatbot “Tolans” offer a refreshingly grounded take on digital companionship, designed to support users’ wellbeing rather than simulate romance or endless flattery. These cartoonish, nonhuman avatars actively discourage overuse, promote real-world relationships, and avoid emotional manipulation—features rooted in psychological research.
The app, especially popular with young women, has over 100,000 monthly users and is projected to bring in $12 million this year. With $20 million in new funding and growing interest in healthier AI interactions, Portola is quietly challenging the Replika-style norm by asking a radical question: What if your AI actually wanted you to go outside? → Read the full article here.
🧬 BIOTECH
AI-Designed Proteins Are Reviving Nanotech
AI-designed proteins are breathing new life into nanotechnology, unlocking applications from biofuels to brain health. At the University of Washington’s Institute for Protein Design, researchers use AI tools like RFdiffusion and ProteinMPNN to build proteins from scratch, tailoring them for tasks like enhancing photosynthesis or creating synthetic vaccines. These molecules could treat diseases, improve crop yields, and even digest plastic.
With backing from Alphabet and startups like Profluent and EvolutionaryScale, the field is accelerating—some models now train on billions of protein sequences. As synthetic biology edges closer to creating virtual cells and protein-based logic gates, nanotech’s long-promised potential may finally be within reach. → Read the full article here.
🛰️ NEWS
What Else is Happening
⚡ xAI Gets Green Light for Generators: Regulators approved 15 natural gas turbines at xAI’s Memphis site, despite pollution concerns and a looming Clean Air Act lawsuit from environmental groups.
🤰 AI Helps Couple Conceive After 18 Years: A new system called STAR found just three hidden sperm in a man's sample, enabling a pregnancy once thought impossible — baby due this December.
📩 Meta Tests Chatbots That Message You First: AI bots on Instagram and WhatsApp may soon send unprompted follow-ups to boost engagement — raising fresh questions about consent and AI boundaries.
📉 CEOs Admit AI Will Cut Jobs: Ford’s CEO says AI could replace half of all white-collar workers, joining execs at Amazon, JPMorgan, and Anthropic in warning of major workforce cuts as automation ramps up.
😷 Accenture Warns AI Emissions Could Soar: AI data centers may drive an 11-fold carbon spike by 2030—yet Big Tech is racing to build faster, sidelining sustainability for geopolitical and commercial urgency.
📽️ VIDEO
Aaron Levie: White Collar Jobs, Future of SaaS, Agents, Concentration of Power
Matt sits down with Box CEO Aaron Levie to discuss how AI will reshape work—humans won't be replaced, but will manage, guide, and review AI agents instead. Get the full scoop! 👇
🧰 TOOLBOX
Smarter Teams, Faster 3D Builds, and Real-Time Voice Insights
🎯 Luqo.ai: Boost team efficiency with AI-driven chats, smart workflows, and real-time insights—built for remote, hybrid, and scaling teams.
🚀 Sloyd.ai: Instantly generate 3D models with AI—perfect for games, apps, and rapid prototyping with export-ready, customizable assets.
☎️ Voice AI Monitoring: Track and analyze Voice AI agent calls in real time—spot failures, improve performance, and optimize customer journeys.
👉️ Find trending AI tools: Browse the Forward Future AI Tool Library
🗒️ FEEDBACK
Help Us Get Better
What did you think of today's newsletter? |
🤔 FRIDAY FACTS
In 2019, Researchers Taught an AI to "Read Minds"
Using neural data from epilepsy patients undergoing brain surgery, scientists trained a deep learning model to decode thoughts into full sentences. Electrodes recorded activity from speech-related brain regions, allowing the AI to reconstruct what the patients were hearing or thinking.
The results? Surprisingly coherent sentences. While not word-for-word perfect, the system could infer intended speech with impressive fluency—capturing grammar, syntax, and even pauses. This marked a major leap in brain-computer interface research.
It’s not telepathy, but it's a step toward giving voice to those who can't speak due to injury or illness. And yes—it raises serious ethical questions, from consent to cognitive privacy.
