👋 Yann LeCun to Leave Meta for Startup
Meta’s chief AI scientist Yann LeCun plans to exit and start a new company focused on “world models,” per the Financial Times. His departure follows Mark Zuckerberg’s AI strategy overhaul and the $14.3B Scale AI deal that brought Alexandr Wang on board to accelerate Meta’s superintelligence push.

💸 SoftBank Sells Entire $5.8 Billion NVIDIA Stake
SoftBank has sold its full 32 million-share stake in NVIDIA as part of a broader reallocation toward its AI ventures. SoftBank says the move frees capital for major bets, including a planned $30 billion investment in OpenAI and its “Stargate” data-center project. Analysts are split on whether this signals profit-taking or deeper caution around AI infrastructure growth.

📈 Lovable Nears 8M Users, Eyes Enterprise Growth
Stockholm-based AI coding startup Lovable has surged to nearly 8 million users—up from 2.3 million in July—and sees 100,000 new products built daily. CEO Anton Osika said more than half of Fortune 500 firms use the platform as the company expands hiring and targets corporate adoption.

🗣️ Meta Open-Sources ASR Model for 1,600+ Languages
Meta unveiled Omnilingual ASR, an open-source speech recognition suite supporting over 1,600 languages—and extensible to 5,400 via zero-shot learning. Licensed under Apache 2.0, it marks Meta’s return to open-source AI, empowering communities and enterprises to transcribe speech globally without restrictions.

Enjoying our newsletter? Forward it to a colleague—
it’s one of the best ways to support us.

Apple’s deliberate approach to artificial intelligence is gaining new validation, according to reports from Bloomberg and 9to5Mac. The company is reportedly finalizing a $1 billion-per-year deal with Google to integrate a custom AI model into Siri’s next major update in iOS 26.4—far less than the $20 billion Google pays Apple annually for default search placement.

Analysts now see Apple’s restraint as a strength: by leveraging partners’ models instead of building costly infrastructure, Apple can deliver AI features to hundreds of millions of devices without the financial volatility facing peers like Microsoft, Meta, and OpenAI. Wall Street, once critical of Apple’s slower AI push, is starting to reassess. → Read the full article here.

Researchers from the University of Sydney have fixed a blurring issue in the James Webb Space Telescope’s imaging system using artificial intelligence—no hardware changes or space missions required. The problem affected the telescope’s Aperture Masking Interferometer (API), which helps the Near-InfraRed Imager and Slitless Spectrograph (NIRISS) hunt for faint exoplanets.

Former Ph.D. students Max Charles and Louis Desdoigts developed a neural network called AMIGO (Aperture Masking Interferometry Generative Observations) that detects and corrects pixel distortions caused by electrical interference. The AI correction restored image sharpness comparable to Hubble’s post-repair clarity, allowing API to capture detailed views of distant stars, exoplanets, and celestial phenomena. → Read the full article here.

AI startup Goodfire.ai has uncovered clear evidence that memorization and reasoning operate through separate neural pathways in large language models. In a preprint released in late October 2025, researchers showed they could disable a model’s ability to recall training data while preserving nearly all logical reasoning functions. Using curvature analysis of the “loss landscape” in the Allen Institute’s OLMo-7B model, the team removed low-curvature components—responsible for memorized data—without degrading reasoning tasks.

Strikingly, basic arithmetic relied on the same neural circuits as memorization: once those were deleted, math accuracy fell by 34%, while reasoning tasks stayed near 100%. The finding could one day help AI developers selectively erase sensitive or copyrighted content without harming model performance. → Read the full article here.

🧔 Michael Burry Calls Out AI Giants: Investor Michael Burry accused major AI hyperscalers of inflating profits through accounting maneuvers. His warning echoes his pre-crash skepticism before 2008’s financial crisis.

🖌️ Google Photos Adds Voice Edits to iOS: iPhone users can now describe photo changes using Google’s Nano Banana AI, enabling smarter facial fixes and creative style transformations.

💰 Wonderful Raises $100M for AI Agents: Israeli startup Wonderful secured $100M Series A led by Index Ventures to scale multilingual AI customer service agents across global markets.

💸 CoreWeave Shares Drop on Data Center Delays: NVIDIA-backed CoreWeave cut its annual revenue forecast after infrastructure setbacks, erasing $5B in market value despite strong AI demand.

🫂 Former Doctor Launches Empathic AI ‘Robyn’: Ex-Harvard physician Jenny Shao unveiled Robyn, a $19.99-a-month AI companion designed to offer emotional support without replacing therapy.

Find jobs where your AI skills give you an edge, even when it's not in the job description.

❤️ Love Forward Future? Spread the word & earn rewards! 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.

📥 Got a hot tip or burning question? Drop us a note! The best reader insights, questions, and scoops may be featured in future editions. Submit here.

🛰️ Want more Forward Future? Follow us on X for quick updates, subscribe to our YouTube for deep dives, or add us to your RSS Feed for seamless reading.

Thanks for reading today’s newsletter—see you next time!

Matthew Berman & The Forward Future Team

🧑‍🚀 🧑‍🚀 🧑‍🚀 🧑‍🚀