OpenAI commits $1.4T to infrastructure; Nvidia projects $500B in chip sales. Same day, same pitch: AI requires sovereign-scale capital. The tension: revenue is still catching up to the rhetoric, and someone will hold expensive capacity if demand falters.
OpenAI's restructure puts a $130B nonprofit in control of a $500B for-profit. Microsoft locked in 27% and access through 2032. But a year of regulatory review produced new numbers, not new transparency—and Microsoft's accounting gaps remain unaddressed.
Good Morning from San Francisco, Adobe stopped pretending one AI model will rule them all. The company built a switchboard
Explore the shift from Big Tech to more personal digital spaces. Understand user motivations, consequences, and the future of online interaction.
A new study by The Verge and Vox Media's Insights and Research team exposes a seismic shift in online behavior. Their December 2024 survey of over 2,000 US adults, conducted with Two Cents Insights, reveals that Google's search results feel less useful to 42% of users.
Social media giants watch helplessly as their audiences seek refuge in smaller, more intimate online spaces. This shift signals more than just platform fatigue – it marks a fundamental change in how people want to connect online.
Users aren't just leaving big platforms. They're reimagining what online communities should be. The data shows a clear pattern: people crave authentic connections in smaller, purpose-driven groups where trust and shared values thrive. They're tired of feeling like "numbers in a giant algorithmic machine," as 60% of survey respondents described their social media experience.
The content creation landscape tells an equally compelling story. In most online communities, 90% of members never post. Another 9% contribute occasionally. Just 1% of users generate 80% of the content. But here's the surprise: lurkers still feel connected. They don't need to post constantly to feel part of the community. Sometimes, silent presence creates meaningful bonds.
AI adds another layer of complexity to this digital transformation. While artificial intelligence powers new tools and experiences, many users approach it with caution. Nearly half of survey participants prefer communities that ban AI-generated content entirely. They want real human interaction, not automated engagement.
The research reveals a striking paradox in how we build online spaces. Traditional social media promised to connect the world, but instead created echo chambers that divided us. Now, users seek smaller, more focused communities – and these intimate spaces might actually help bridge the divisions that larger platforms created.
TikTok's rise amid this shift isn't just about short-form video. It represents a broader hunger for authenticity and direct connection. Users flock to platforms that feel real, even as they navigate the blurry line between genuine content and calculated performance.
Trust emerges as the critical currency in this new landscape. Legacy platforms hemorrhage user confidence while niche communities gain momentum. These smaller spaces succeed by prioritizing human connection over algorithmic engagement, meaningful interaction over viral reach.
Google's declining usefulness marks a turning point. When 42% of users find the internet's primary gateway less helpful, it signals a deeper problem with how we organize and access online information. This creates opportunities for new platforms that better understand how people want to discover and connect.
The Verge's research suggests we're entering a "post-social media era." But this doesn't mean the end of online connection. Instead, it points to a future where digital communities more closely mirror real-world relationships: smaller, deeper, and built on genuine shared interests.
Why this matters:
Read on, my dear:
AI assistants fail basic accuracy checks on news queries nearly half the time, but users don't just blame the AI—they blame the news outlets it cites. As adoption climbs, newsrooms face reputational damage for errors they didn't commit and can't fix.
Anthropic wires Claude into lab systems for documentation speed while rivals burn billions chasing AI-discovered drugs that don't exist yet. The strategy: sell efficiency today, skip moonshot risk—but if discovery suddenly works, infrastructure looks conservative.
Security teams assumed attackers needed to taint a percentage of training data. New research shows a fixed number of documents can backdoor models regardless of scale—upending detection strategies built around dilution assumptions.
How can an AI master a complex game without ever playing it? DeepMind's Dreamer 4 learns by watching, then trains in imagination. This shift from big data to efficient world models could be key for real-world robotics and autonomous systems.
Get the 5-minute Silicon Valley AI briefing, every weekday morning — free.
