The SWELL #1
SAIL Weekly Digest January 17 – 23 | Issue #1
SAIL authors dissected Claude Code from every angle this week—psychosis-inducing rabbit holes, risk-shifting economics, getting good at agents. One question ties it together: how do our strategies evolve as AI tools get significantly more capable?
Thursday’s Live with Jasmine Sun, Timothy B. Lee, and Sebastian Raschka, PhD tackled exactly that. [→ Watch the replay]
The Week in AI
Claude Code ships widely — Anthropic’s agentic coding tool is now in users’ hands, sparking debate about reliability, risk, and workflow integration.
DeepSeek’s 1-year anniversary arrives January 27 — A new model is expected before Chinese New Year, and we’re all on the edge of our seats.
US productivity jumps 4.9% — AI’s economic impact is becoming measurable in the macro data.
This Week from SAIL Authors
On Claude Code & AI Agents
📰 claude code psychosis By Jasmine Sun • Jasmi.News Are your problems software-shaped? A deep dive into what happens when Claude Code becomes your default problem-solving tool. [→ Read more]
📰 How shifting risk to users makes Claude Code more powerful By Timothy B. Lee • Understanding AI Agentic AI systems push decisions—and consequences—onto users in ways that make them more capable but also more dangerous. [→ Read more]
📰 Get Good at Agents By Nathan Lambert • Interconnects The tools are getting so powerful that we need to level up our skills to match—here’s how to actually use AI agents effectively. [→ Read more]
📰 Anthropic’s Head of Economics on AI adoption, Claude Code, the burden of keeping up By Azeem Azhar • Exponential View A conversation on the economic realities of AI adoption and what Anthropic is learning from Claude Code usage patterns. [→ Read more]
On Open Source & Competition
📰 No Business Model: DeepSeek’s Enduring Advantage By Kevin Xu • Interconnected No other frontier lab operates without commercial pressure—and that freedom to release weights and push research is reshaping the race. [→ Read more]
📰 The Death of Software 2.0 (A Better Analogy!) By Doug O’Laughlin • Fabricated Knowledge The “Software 2.0” frame has run its course—here’s a better way to think about what AI is actually doing to software. [→ Read more]
📰 Zhipu and MiniMax IPO By Jordan Schneider • ChinaTalk A week in prospectus-land as two Chinese AI companies prepare to go public. [→ Read more]
On Geopolitics & Policy
📰 The Future of Economic Security with Dan Kim and Chris Miller By Jordan Schneider • ChinaTalk The CHIPS Act authors discuss what comes next for tech competition and national security. [→ Read more]
📰 Exponential View #557: Starlink, Iran & abundance By Azeem Azhar • Exponential View AI’s geopolitical reach is expanding faster than the discourse can keep up. [→ Read more]
📰 Data to start your week By Azeem Azhar • Exponential View Revenue-compute curves, silicon détente, and the numbers that matter right now. [→ Read more]
Quick Links
claude code psychosis — Jasmine Sun | Jasmi.News
How shifting risk to users makes Claude Code more powerful — Timothy B. Lee | Understanding AI
Get Good at Agents — Nathan Lambert | Interconnects
Anthropic’s Head of Economics on AI adoption — Azeem Azhar | Exponential View
No Business Model: DeepSeek’s Enduring Advantage — Kevin Xu | Interconnected
The Death of Software 2.0 — Doug O’Laughlin | Fabricated Knowledge
Zhipu and MiniMax IPO — Jordan Schneider | ChinaTalk
The Future of Economic Security — Jordan Schneider | ChinaTalk
Exponential View #557 — Azeem Azhar | Exponential View
Data to start your week — Azeem Azhar | Exponential View
Join us next Thursday for our weekly Substack Live — hosted by Jasmine + two SAIL writers TBA.
