·
Finding signal on Twitter is more difficult than it used to be. We curate the best tweets on topics like AI, startups, and product development every weekday so you can focus on what matters.
PLUS: A16z scouts and 996 schedules
Claude launched Cowork for non-technical tasks, bringing the same workflow developers use with Claude Code to meeting transcripts and file management.
Amit joined Andreessen Horowitz's venture scout program to write checks into startups after getting one to 20x and another to 10x.
Some U.S. tech firms are recruiting for "996" roles that run 9 a.m. to 9 p.m., six days a week.
Apple determined that Google's AI provides the most capable foundation for Apple Foundation Models.
Ramp built Inspect, a background coding agent that wrote 30% of merged PRs last week, and open sourced the full blueprint.
Twitter users joke about timeline fatigue with a urinal meme about Claude Code productivity tips.
DeepSeek founder Liang Wenfeng's quantitative hedge fund generated returns of more than 50% last year.
OpenAI acquired Torch Health, with the team joining to build ChatGPT Health into the best AI tool for health and wellness.
Ralph is a coding technique that treats AI like a volatile process, starting fresh sessions to avoid context pollution instead of running 40-minute loops.
AI companies are scraping content and selling it back as tokens, breaking the economics of future content creation.
European developers are asking if switching from Stripe to Polar is worth it just for tax handling.
Anthropic launched Claude for Healthcare with HIPAA-ready products and expanded Claude for Life Sciences with connectors for clinical trials and regulatory operations.
Marc Lou added a review system to TrustMRR that requires KYC to increase trust.
Financial Times chart shows high social skills plus high math skills drive employment and wage growth while low social skills plateau.
Developers say Opus 4.5 is best for sync work while GPT 5.2 handles async tasks.
One-person dev teams now split between multiple agents at inference speed, heavy AI IDE use, and mostly manual coding.
New paper introduces AgeMem, a unified framework for agent memory management that achieves 13-21% improvement over fragmented heuristic pipelines.
Packy McCormick's piece on a16z is worth clearing your calendar for.
Turn hours of scrolling into a five minute read.