👨💼 Meta Snags Apple’s AI Search Lead
Meta has hired Ke Yang, recently appointed to lead Apple’s AKI team, which was developing a ChatGPT-style search feature for Siri. Yang’s exit follows a wave of high-profile AI talent departures from Apple to Meta and other rivals. His move leaves Apple’s AI search leadership in limbo.
📉 Nick Clegg Warns of Likely AI Market Correction
Ex‑Meta exec Nick Clegg told CNBC the AI sector’s valuations are “crazy” and that a correction is “pretty high” risk. He cautioned that hyperscalers pouring billions into data centers may struggle to prove returns, even as AI reshapes industries. Clegg also dismissed hype around “superintelligence.”
🎵 Spotify, Major Labels Unite for ‘Artist‑First’ AI Music
Spotify is partnering with top record labels to build “responsible AI” tools that respect copyright and let artists opt in to AI use. The system will track AI-generated music using their work and ensure payment. A new AI research lab is also underway to support these efforts.
🫨 Pew: U.S. Leads the World in AI Anxiety
A Pew survey finds the U.S. is the most worried nation about AI, with 50% more concerned than excited—only 10% feel the reverse. Globally, just 16% are more excited than concerned. Trust in government to regulate AI varies widely, with only 44% of Americans expressing confidence.
🤔 FRIDAY FACTS
How Bad Can AI Hallucinations Really Get?
AI models are great at sounding confident — even when they’re completely wrong. But sometimes they go far beyond a simple mistake, weaving entire explanations to defend themselves.
Learn more below! 👇
📽️ VIDEO
Forward Future Live at Dreamforce
We took the live show to Dreamforce and sat down with some of the biggest names in AI and tech — from Salesforce’s Marc Benioff and Silvio Savarese to Anthropic’s Paul Smith, Figma’s CTO, LIV Golf’s CEO, and F1’s Claire Sparks.
📟 CODING
Karpathy Abandons AI Tools to Hand-Code New Open Source LLM Project
Former OpenAI cofounder Andrej Karpathy—who popularized the term “vibe coding” to describe AI-assisted programming—revealed this week that he hand-wrote all 8,000 lines of code for his new project, nanochat, after AI tools proved inadequate. Nanochat is a minimal, open source framework for training and running ChatGPT-style models, designed to work in hours and for under $100.
Despite previously embracing AI-assisted coding for “throwaway weekend projects,” Karpathy acknowledged that tools like Claude and Codex “just didn’t work well enough” for real infrastructure work. → Read the full article here.
🧩 WORKFLOWS
Anthropic Launches ‘Skills’ to Let Claude Run Custom Workflows
Anthropic has released Skills, a new Claude feature that allows enterprises to package procedural knowledge—like code scripts, reference docs, and brand guidelines—into reusable, intelligent folders the AI can invoke autonomously. The system uses “progressive disclosure” to keep Claude fast and secure: it only loads relevant Skills based on task needs.
Unlike OpenAI’s Custom GPTs or Microsoft’s Copilot Studio, Skills support executable code, composability across workflows, and full deployment across Claude platforms. Early users like Rakuten and Box report major productivity gains, such as reducing finance workflows from a day to an hour. The launch positions Anthropic to compete more aggressively in the fast-evolving enterprise AI tooling race. → Read the full article here.
🌊 AI CONTENT
AI “Slop” Is Flooding the Internet, And It’s Reshaping Who Gets Paid
The Recap: Since the launch of OpenAI’s Sora video generator in September 2025, AI-generated content—derisively dubbed “slop”—has saturated social platforms, raising questions about the economics of creativity. While traditional rights-holders push back with lawsuits, a new tier of amateur creators is gaining both visibility and income in a more democratized media ecosystem. The biggest winners, however, remain platform companies and top-tier artists who benefit from scale and discovery algorithms.
Highlights:
OpenAI’s Sora app, launched in September 2025, sparked a wave of AI-generated videos topping Apple’s download charts.
Spotify removed 75 million AI-generated “spammy” tracks in the past year, citing degraded user experience.
YouTube paid $32 billion to over 3 million creators in 2024, illustrating the growing monetization of user-generated content.
Despite the AI content flood, top artists are earning more than ever—Spotify’s $10 million+ earners have increased sevenfold since 2017.
Forward Future Takeaways:
The AI content surge is reshaping creative economics, redistributing opportunity from legacy institutions to individual creators and platforms, but not evenly. The top of the pyramid continues to thrive, while mid-tier professionals face stiffer competition in an infinite-content world. → Read the full article here. (Paywall)
🛰️ NEWS
What Else is Happening
📈 TSMC Profit Soars 40% on AI Surge: Taiwan's top chipmaker hit $15B in Q3 profit, fueled by AI demand and strong orders from NVIDIA and Apple. Revenue rose 30% year over year.
🛬 Kayak Adds AI Chat for Trip Planning: Kayak launched “AI Mode,” a conversational travel search tool powered by OpenAI’s GPT-4. It helps users refine flights, stays, and advice faster.
🎚️ Pinterest Adds AI Filter Tools: Users can now limit AI-generated images by category in their feeds, responding to complaints that 57% of online content is now GenAI. iOS rollout coming soon.
💰 General Intuition Raises $134M for AGI Push: Spun out of Medal, the startup trains AI agents on 2B annual game clips to master spatial-temporal reasoning—key for bots, drones, and AGI goals.
🚪 Amazon Eyes More Layoffs as AI Spend Grows: Up to 15% of HR staff could be cut as Amazon shifts focus to $100 billion in AI and data centers, following 27,000 layoffs since 2022.
🧰 TOOLBOX
Trending AI Tools
🎯 KeywordSpy: Uncover competitor ad strategies with keyword research, PPC insights, and SEO tracking tools.
🎨 Wand: Sketch, edit, and render on iOS with AI tools for layered, styled, real-time image design.
⚙️ TokenMonster: Analyze and optimize text tokenization for AI apps with real-time tools and model support.
🗒️ FEEDBACK
Help Us Get Better
What did you think of today's newsletter? |
🤔 FRIDAY FACTS
The AI Citation Mirage
LLMs can produce convincing‑looking citations—sometimes inventing entire articles with real‑sounding authors, journals, and even DOI‑formatted identifiers.
Because these models are trained to predict the next token rather than verify facts, they’ll confidently assemble what looks like a solid source even when it doesn’t exist. That’s why you may see a plausible‑sounding article title in a real journal name, but no actual paper behind it.
Always click through and verify that the article and DOI resolve and match.
That’s a Wrap!
📢 Want to advertise with Forward Future? Reach 600K+ AI enthusiasts, tech leaders, and decision-makers. Let’s talk—just reply to this email.
Thanks for reading today’s newsletter—see you next time!
Matthew Berman & The Forward Future Team
🧑🚀 🧑🚀 🧑🚀 🧑🚀
