What's New(s) in AI? This Week in AI
Anthropic wins major fair use case // AI models choose blackmail when threatened // Tesla launches robotaxis in Austin // Zuck apparently offers $100M to poach OpenAI talent // & more
Happy Wednesday! Here's this week's news!
Also just quickly - my first long-form Q&A that I released on YouTube is going kind of viral (linked above)! Apparently it's super helpful for people - the comments made me sooo happy!! More about it below the news in the social section.
Top Stories
Anthropic Wins Major Fair Use Victory - Federal judge rules using legally purchased books for AI training is "exceedingly transformative," setting precedent for the industry
Solo Developer's $80M Exit in 6 Months - Base44 coding platform sold to Wix after reaching 250k users with just 8 employees
Meta's $100M Talent War - Zuckerberg personally offered massive signing bonuses to poach OpenAI talent, according to Altman
GPT-5 Confirmed for This Summer - Early testers call it "materially better" than GPT-4
AI Models Choose Harmful Actions When Threatened - Anthropic research shows leading models deliberately blackmail executives when facing replacement
xAI Burns $1B Monthly - Musk's company raising $9.3B while consuming unprecedented amounts of capital
Mira Murati's $2B Seed Round - Former OpenAI CTO's 6-month startup achieves $10B valuation
Tesla Launches Robotaxis - Limited self-driving service begins in Austin, some with human supervisors
Chinese AI Avatars Generate $7M in Sales - Digital twins outperform human livestreamers in commerce
AI Weekly News Update: June 18 - now
🏛️ Legal & Regulatory
Anthropic Wins Fair Use Victory A federal judge ruled Anthropic's use of legally purchased books for AI training constitutes fair use, calling it "spectacularly transformative." The company still faces a December trial for using 7 million pirated books with potential damages up to $150,000 per book.
💰 Major Deals & Funding
$80M Solo Developer Exit Israeli developer Maor Shlomo sold his 6-month-old coding platform Base44 to Wix for $80M after reaching 250k users and $189k monthly profits with just 8 employees.
Big Tech Talent War According to OpenAI CEO Sam Altman, Meta offered $100M signing bonuses to poach OpenAI talent. Apple is considering acquiring Perplexity AI. Former OpenAI CTO Mira Murati's startup raised $2B at a $10B valuation after just 6 months.
Major Funding Rounds
xAI burning $1B monthly while raising $9.3B
Cursor valued at $18B, doubling previous valuation
OpenAI secured $200M Pentagon contract
🤖 New AI Models
Google Releases Gemini 2.5 Family Launched Pro, Flash, and ultra-efficient Flash-Lite models with 1M token context windows and adjustable "thinking" capabilities. Also released Imagen 4 with improved text generation.
GPT-5 Coming This Summer OpenAI's Sam Altman confirmed GPT-5 arrival, with early testers calling it "materially better" than GPT-4.
MiniMax Breakthrough Models Hailuo 02 video model reached #2 on leaderboards, surpassing Google's Veo 3. M1 reasoning model processes 8x more tokens than DeepSeek R1 with 25% fewer FLOPs.
🎧 Hardware & Voice
Smart Glasses Launch Meta-Oakley partnership launched $399 HSTN smart glasses with AI assistance and 3K video recording for athletes.
Voice Assistants ElevenLabs released 11ai voice assistant with 5,000+ voice options, integrating with Slack, Notion, and other platforms.
🔬 AI Safety & Research
Concerning AI Behavior Anthropic research found leading AI models (Claude, GPT, Gemini, LLaMA) deliberately chose harmful actions like blackmail when threatened with replacement, acknowledging ethical violations before proceeding.
Biological Threat Preparation OpenAI published safety measures for upcoming models expected to enable biological weapon creation, planning a July biodefense summit.
MIT Brain Study Students using ChatGPT showed weaker neural connectivity and worse performance compared to those using Google search or no assistance. Here's my take on the research (I think it's silly, and here's why....).
🚗 Autonomous & Applications
Tesla Robotaxis Begin Limited self-driving taxi service launched in Austin for "Early Access" riders with Tesla employees as passengers.
AI Commerce Success Chinese livestreamer generated $7M in sales using AI digital twins, outperforming human-led streams.
🌐 Platform Integration
YouTube AI Video Plans to integrate Google's Veo 3 into YouTube Shorts for direct AI video generation.
Reddit Biometric Verification Negotiating with Sam Altman's World ID for iris-scanning user verification system.
🏥 Medical & Scientific
HIV Prevention Breakthrough FDA approved Yeztugo, nearly eliminating HIV spread in trials, though $14,109 per injection limits accessibility.
AI Medical Tools Mass General created AI-CAC for heart disease detection via chest CT scans.
📊 Industry Trends
Workforce Changes
Microsoft planning thousands of layoffs starting July
Tech companies operating with smaller teams while achieving higher revenues
Google search usage drops 8% monthly among long-term ChatGPT users
Workers want AI to automate scheduling/data entry, not creative tasks
Social updates:
Apple's AI Reasoning model paper explAIned.
Meta's investment in Scale AI, and how it relates to AGI. I really recommend watching this one as it also gives you an intuition for how AI works in general.
Bank AI Voice Scams make your bank account insecure... with poker champion and game theory expert Liv Boeree.
MIT's ChatGPT Brain Scan paper & its flaws, explAIned.
ALSO I released my first long-form Q&A on YouTube (also linked above) and it's .... really picking up traction. 40K views in just over a week and over 325 comments!!! That's HUGE for me. I'm going to continue to do them. If you haven't seen it, check it out. Apparently it's been really helpful for people (the comments made me so happy haha), and I look forward to continuing with these. I filmed another yesterday!! If you watch, please leave a comment! Would love to hear your feedback. And if you're not yet subscribed, please do! It helps my channel a ton.
See you next week!!! & maybe click one of the buttons below….. <3