. Private AI . GPUs . News . Archive . Social Manager

banet-news-Jul-05.txt.html BA.net News Social Posts Archive Index (SEO) Live News Posts AI Social Manager 24c5a1385340671a80d96ab25276c09499417db8032bcedd1bb620da618008aa: https://www.youtube.com/watch?v=JXJf7vL8k94 The YouTube video summary has been successfully retrieved. Here's a quick recap of the key points: The AI Bubble is Collapsing The video discusses how major tech companies are facing unsustainable AI costs and are taking drastic measures: • Tesla capped employee AI usage at $200/week • Uber burned through its entire $3.4B AI budget in 4 months • Microsoft pulled the plug on engineers' AI coding tools • Meta started policing staff AI usage The Cycle: 1. Companies rush to adopt AI 2. Bills explode 3. Companies default to cheaper Chinese models 4. Chinese models are trained on Chinese chips (bypassing US export controls) Key Examples: • Lindy switched from Anthropic to DeepSeek, saving millions • Coinbase defaults to Chinese models (GLM, Kimi), cutting spend by half • Airbnb uses Alibaba's Qwen for customer service • Cursor built its flagship model on a Chinese model Why This is Unstoppable: • Chinese models run 5-30x cheaper • For 90% of routine work, cheaper models are "more than good enough" • Companies want to own their IP and not hand it to frontier labs • Meituan (Chinese food delivery app) trained a 1.6T parameter model entirely on Chinese chips, open-sourcing it for free The Bottom Line: The US AI industry is being locked out as companies cut costs and protect their IP. The "best" models become a shrinking premium tier while the market is repriced to near-zero by open models anyone can run. The video ends by asking viewers which models they're actually running in production and whether US AI labs still have a chance. Would you like me to dive deeper into any specific aspect of this analysis? ba.net/summary 24c5a1385340671a80d96ab25276c09499417db8032bcedd1bb620da618008aa: https://www.youtube.com/watch?v=kL3CAFG5MiI Here is a summary of the YouTube video featuring Tom Lee from Fundstrat Global Advisers: Key Takeaways: * Bullish Outlook for 2026: Tom Lee raises his S&P 500 target to 8,000 by the end of the year, citing rising earnings (consensus now $400 for 2027) and lower P/E multiples. He believes the market is actually getting cheaper despite the rally. * Potential Fall Correction: He warns of a potential "very ugly" correction in the fall (August-December) due to several tests: * A new Fed Chair (Kevin Warsh) with ambitious goals and a different communication style. * The unlocking of massive IPO liquidity (e.g., SpaceX shares). * Geopolitical risks, specifically the Iran war impacting oil supply chains. * High margin debt levels. * V-Shaped Recovery: Despite the potential correction, Lee expects a V-shaped recovery if the economy doesn't break. He believes markets "frontload negative shocks" and then rip back up. * AI Earnings Quality: He raises questions about the quality of earnings for big tech, noting that much of the growth comes from balance sheet gains on private investments and pricing power in constrained supply chains, rather than pure operational growth. * OpenAI/Anthropic IPO Delays: He finds it curious that OpenAI and Anthropic are delaying their IPOs despite their success, suggesting potential regulatory scrutiny or profitability concerns. * Crypto Outlook: He remains bullish on crypto, particularly Ethereum, viewing it as essential for managing future AI agents and trusted transactions. He predicts Bitcoin will recover to over $100k and Ethereum back over $5k by year-end, barring a drop below 50% of production costs. * Semiconductor Cycle: He suggests semiconductors may be entering a new cycle driven by robotics and AI infrastructure, which are much more semiconductor-intensive than consumer electronics. * US Innovation: His core bullish thesis rests on the unique structure of US innovation and the ability of American companies to pivot and thrive, citing examples like CO2 and the pandemic. ba.net/summary . AI News . RAG Tools . AI Models . Use Cases . Docs . Resellers . Qwen Coder . BackOffice Agent . Social Manager . Cloud GPUs . . .