Chinese Tech — Week of 2026-03-31 to 2026-04-03#

Week in Review#

The dominant theme across the Chinese tech ecosystem this week was the sudden acceleration of AI agent workflows, unexpectedly catalyzed by Anthropic’s colossal source code leak. While frontier labs transition from consumer-facing demos to highly profitable enterprise infrastructures, the developer community is fiercely debating the right architectural boundaries for autonomous agents. Simultaneously, a noticeable counter-culture is emerging in consumer tech, with users rejecting hyper-processed AI outputs in favor of analog imperfections and human “taste.”

Simon Willison — Week of 2026-04-11 to 2026-04-17#

Highlight of the Week#

This week’s most striking revelation came from Simon’s infamous “pelican riding a bicycle” SVG generation benchmark, where a 21GB quantized local model (Qwen3.6-35B-A3B) unexpectedly outperformed Anthropic’s brand-new Claude Opus 4.7 flagship. Running locally on a MacBook Pro via LM Studio, Qwen generated a better bicycle frame and even won a secret unicycle backup test, leading Simon to conclude that his joke benchmark’s long-standing correlation with general model utility has finally broken down.

AI@X — Week of 2026-04-18 to 2026-05-01#

The Buzz#

The enterprise software paradigm is undergoing a seismic shift from human-centric, seat-based SaaS to “headless,” consumption-based API platforms driven by autonomous agents. As agents become the primary software users who “yolo straight to the tokens,” developers are realizing that traditional graphical user interfaces are increasingly obsolete for deep operational workflows. This pivot to an agent-first ecosystem is vastly expanding the total addressable use-cases for systems of record, while aggressively rendering recent LLMOps wrappers and visual interfaces completely obsolete.

Simon Willison — Week of 2026-05-08 to 2026-05-15#

Highlight of the Week#

The standout development this week is Simon’s rapid adaptation to the latest frontier model capabilities, most notably releasing llm 0.32a2 to expose and visualize the new interleaved reasoning tokens of GPT-5 class models directly in the terminal. This perfectly pairs with his hands-on explorations of embedding LLM calls deeply into developer workflows, such as executing prompts via script shebangs and leveraging models to output rich HTML rather than just Markdown.

Chinese Tech Daily — 2026-04-03#

Top Story#

Google’s release of the Gemma 4 open-source model series marks a pivotal shift toward true “local AI” by moving to the commercially permissive Apache 2.0 license. The lineup ranges from edge-optimized E2B and E4B models—capable of running completely offline on smartphones and Raspberry Pi devices—to highly efficient 26B MoE and 31B Dense models that rival much larger parameter counts in complex reasoning benchmarks. By engineering these models with native function calling, multimodal inputs, and 128K+ context windows specifically tailored for autonomous agent workflows, Google is drastically lowering the barrier for edge device AI integration while preserving data sovereignty.

Simon Willison — 2026-04-16#

Highlight#

The most fascinating takeaway today is a surprising win for local AI: a 21GB quantized Qwen3.6 model running on a laptop beat Anthropic’s brand-new Claude Opus 4.7 at Simon’s “pelican riding a bicycle” SVG generation benchmark. This result leads Simon to conclude that his joke benchmark’s long-standing correlation with a model’s general utility has finally broken down.

Posts#

Qwen3.6-35B-A3B on my laptop drew me a better pelican than Claude Opus 4.7 · Source Simon put the day’s two major model releases—Alibaba’s Qwen3.6-35B-A3B and Anthropic’s Claude Opus 4.7—through his infamous “pelican riding a bicycle” SVG generation benchmark. Running locally on a MacBook Pro via LM Studio, the quantized Qwen model produced a better bicycle frame than Opus, and even won a “secret backup test” generating a flamingo riding a unicycle. Simon admits this breaks the historical correlation between his SVG benchmark and a model’s general usefulness, noting he highly doubts the 21GB local model is actually more capable than Anthropic’s proprietary flagship.

The Vibe-Coding Backlash, Microsoft’s OpenAI Pivot, and AI’s “Hindenburg” Moment — 2026-04-27#

Highlights#

The AI community is fiercely debating the fallout of “vibe coding” disasters, with experts warning that deploying autonomous coding agents without traditional software engineering safeguards is a recipe for catastrophic data loss. At the same time, the strategic landscape is shifting massively as Microsoft and OpenAI renegotiate their exclusivity, signaling a new, highly competitive era for cloud-AI partnerships and antitrust positioning.

Infrastructure Reality Checks & The Agentic Era — 2026-04-28#

Highlights#

Today’s discourse reveals a profound tension in the AI ecosystem: massive infrastructural and ethical anxieties are colliding with surging end-user capabilities. While OpenAI faces severe internal financial pressures and Google draws intense ethical scrutiny over autonomous weapons contracts, the developer community continues to accelerate into the “agentic era” with the release of GPT-5.5, escape-velocity code generation, and a shift away from human-centric software design.

The OpenAI Trial Fallout and Enterprise Agent Expansion — 2026-05-04#

Highlights#

Today’s discourse is largely consumed by dramatic revelations emerging from the Musk v. OpenAI trial, with sworn testimony unearthing the stark financial realities behind OpenAI’s pivot from a nonprofit to a capped-profit entity. Simultaneously, the technical frontier is rapidly shifting toward enterprise-grade AI agents, highlighting a critical moment where AI integration moves past basic coding and forces sweeping modernization in corporate IT workflows.

Simon Willison — 2026-05-06#

Highlight#

The highlight of today is Simon’s candid reflection on how highly reliable coding tools like Claude Code are blurring the line between professional “agentic engineering” and hands-off “vibe coding”. He raises important questions about accountability, the loss of traditional software evaluation metrics, and how the bottlenecks of the entire software development lifecycle are radically shifting.

Posts#

Vibe coding and agentic engineering are getting closer than I’d like Simon expands on a recent podcast conversation to discuss how he is increasingly treating AI agents like Claude Code as semi-black boxes, trusting them to write unreviewed production code. He notes that because AI can generate comprehensive tests and beautiful readmes in minutes, traditional signals of software quality are losing their value, making actual usage the most important metric. Furthermore, he observes that as coding speeds up exponentially, upstream bottlenecks like cautious, extensive design processes are being fundamentally challenged. Despite these shifts, he isn’t worried about the future of software engineering careers, emphasizing that these tools are simply amplifiers for a discipline that remains fiercely difficult.