AI Salaries Drive San Francisco House Prices to Record $1.76M Median
San Francisco home prices have surged to a record median of $1.76 million as of May 2026, driven by massive wealth from AI workers at companies like OpenAI and Anthropic. The city has overtaken San Jose as the most expensive US homebuying market, with prices up 14-19% year-over-year since March.
A three-bedroom apartment in Duboce Triangle was listed for nearly $3 million, with the seller willing to accept shares in OpenAI or Anthropic instead of cash. An OpenAI employee who viewed the property said he plans to ask his bosses about the stock transfer option.
Key data points from Redfin:
- Median sale price in SF: $1.76M (May 2026) vs ~$400K nationally
- SF prices up 19% YoY in March, 14.5% in April, 14.1% in May
- National prices rose only 1.4-2% in same period
The price boom is attributed to extraordinary AI worker compensation. Last October, over 600 current and former OpenAI employees sold combined shares worth $6.6 billion — an average of $11 million per participant. Anthropic workers also recently sold shares totaling ~$6 billion. Both companies are expected to go public later this year or next, minting more multi-millionaires.
Real estate agent Rachel Swann noted that today's bidding wars will likely be seen as bargains in hindsight.
📖 Read the full source: HN LLM Tools
👀 See Also

Claude Code v2.1.117 Release: Subagent Forking, Plugin Improvements, and Performance Fixes
Claude Code v2.1.117 enables forked subagents on external builds via CLAUDE_CODE_FORK_SUBAGENT=1, improves plugin dependency handling, and fixes Opus 4.7 context window calculations. The release includes faster startup with concurrent MCP connections and replaces Glob/Grep tools with embedded bfs/ugrep on macOS/Linux.

Claude Opus 4.6 effort=low parameter causes lazy agent behavior
When using effort=low with Claude Opus 4.6, agents made fewer tool calls, were less thorough in cross-referencing, and ignored parts of system prompts about web research. Switching to effort=medium resolved the issues.

Claude Code v2.1.157: Auto-Load Plugins from .claude/skills, Improved Agents & Worktrees
Claude Code v2.1.157 automatically loads plugins from .claude/skills, adds claude plugin init scaffolding, honors agent setting in settings.json, and fixes numerous bugs across agents, worktrees, and terminal integration.
Cognitive Surrender: When AI Agents Write Code You Don't Understand
AI coding agents can produce code authors don't understand, leading to unreviewed PRs. The article explores 'cognitive surrender' and how to avoid it in your workflow.