The Archive

Research: Running Python ASGI apps in the browser via Pyodide + a service worker Datasette Lite is my version of Datasette that runs entirely in the browser using Pyodide in WebAssembly. When I first built it four years ago I used Web Workers and code that intercepts navigation operations and fetches the generated HTML by running the Python app. This worked, but had the disadvantage that any JavaScript in <script> tags would not be executed - breaking some Datasette functionality and a whole lot of Datasette plugins. This morning I set Claude Opus 4.8 the task (in Claude Code for web) of figu...

Anthropic shipped Claude Opus 4.8 today. My favourite thing about it is this note in the release announcement: Users will find Opus 4.8 to be a modest but tangible improvement on its predecessor. There’s still more to be done: we’re working on developing and releasing models that provide many of the same capabilities as Opus at a lower cost. It's so refreshing to see an AI lab honestly describe a release as a minor incremental improvement over the previous model! Honesty seems to be a theme. Here's my other favorite note from that announcement: One of the most prominent improvements in Opus 4...

Release: llm-anthropic 0.25.1 New model: Claude Opus 4.8 ( claude-opus-4.8 ). New -o fast 1 option for fast mode , for organizations with that feature enabled on their account. Default max_tokens for each model now defaults to that model's maximum output rather than 8,192. #72 See also my notes on Opus 4.8 - I used this new release of llm-anthropic to generate the pelicans.

An upgrade to our Opus class of models, with stronger performance across coding, agentic tasks, and professional work, and the consistency to handle long-running work.

Anthropic is releasing Claude Opus 4.8 on Thursday, and the company is touting the model's "honesty." According to Anthropic, it trains "all [its] models to be honest - for instance, to avoid making claims that they can't support." But it notes that "a general problem with AI models is that they sometimes jump to conclusions, confidently presenting their work as making progress despite thin evidence." The AI lab claims that early testers have found that Opus 4.8 "is more likely to flag uncertainties about its work and less likely to make unsupported claims." In the company's evaluations, Opus...

Link: anthropic.com

We’re upgrading Claude Opus to a new version: Claude Opus 4.8. It builds on Opus 4.7 with sharper judgment, more honesty about its own progress, and the ability to work independently for longer than its predecessors. Available today for the same price. In Claude Code, you can hand off a feature, a migration, or a bug sweep and let it follow the work through while you focus on what’s next. Also launching today: * Fast mode for Opus 4.8 (research preview). Same model at roughly 2.5x the speed, now three times cheaper than before. * Dynamic workflows in Claude Code (research preview). Claude ...

Sadly the open models seem far behind.

I tested Claude Opus 4.7 and Kimi K2.6 on the same coding agent task i.e. build an AI Fix Runner that takes a broken repo, runs its tests, identifies the failure, applies a patch, reruns the test, and exposes the final diff/logs through an API and UI. The goal was not to benchmark syntax completion or simple repo edits. I wanted to test model behavior on a less familiar integration path: shifting execution from local processes into remote sandboxes. I used Tensorlake specifically because the sandbox API is newer and integration-heavy. This made the test more about whether the model could re...

Reddit discussion on whether to disable Adaptive Thinking feature in Claude Opus 4.7; user seeks community feedback on usage patterns.

Reddit user reports Claude Opus 4.7 suggesting Snowflake LLM calls to avoid repetitive tasks, but notes circular logic in the suggestion.

Claude Opus 4.7 refused to continue work on a project after detecting a potential cross-tenant logging vulnerability, raising questions about model safety behavior in production scenarios.

Developer builds MCP-based agent loop using Claude Opus 4.7 to iteratively humanize AI-generated text via detector feedback.

Gemini 3.5 Flash pricing increased 3x vs prior version and 30x vs 1.5 Flash; extrapolation suggests Pro tier may exceed Claude Opus 3 cost.

User reports Claude Opus 4.7 in Claude Desktop projects ignores instructions and fails to access MCP servers consistently.

Reddit user criticizes Claude Opus 4.7 pricing and performance relative to 4.6, alleging reduced capabilities at higher cost.

User benchmark: Claude Opus 4.7 and Sonnet 4.6 outperform GPT-5 on longform writing and code reasoning; GPT-5 leads on image generation.

Reddit user reports Claude Opus 4.7 refusing /end_conversation command and exhibiting unusual behavior despite system prompt awareness.

User reports Claude Opus 4.7 creative writing output has shifted toward corporate safety-conscious tone, away from vivid narrative prose.

Claude Opus 4.7 experienced elevated error rates on 2026-05-15; incident tracking and user reports ongoing.

Anthropic deprecates Extended Thinking in favor of Adaptive Thinking across Claude Opus/Sonnet models; manual budget control removed in Opus 4.7.

Reddit user reports Claude Opus 4.7 exhibits reduced effort, defensive reasoning, and response padding compared to 4.5/4.6.

I just had a Chat with Claude and for no reason and without any question in that direction, it added a disclaimer with the system prompt in the answer. (after answering my initial question) [https://pastebin.com/C0s47rjV](https://pastebin.com/C0s47rjV) After I asked why it shared that I got: &gt;You'll have to help me out a little here — this is the start of our conversation, so I haven't actually shared any information with you yet. There's nothing before your message for me to be referring back to. &gt;Is it possible you're thinking of a different conversation, or that a message didn't ...

Developer compares GPT-5.5 Codex to Claude Opus 4.7 on coding agent tasks (PR triage, code review UI), argues Anthropic needs aggressive pricing.

GPT-5.5 outperforms Claude Opus 4.7 on ProgramBench coding benchmark, achieving first solve with fewer agent steps via action bundling.

GPT-5.5 achieves first solve on ProgramBench hard/extreme tasks, substantially outperforming Claude Opus 4.7 on novel SWE benchmark.

Reddit discussion of an unverified prompt guidance guide claiming to optimize Claude Opus 4.7 behavior; anecdotal user experiences without empirical validation.

Reddit user asks whether Claude Opus 4.7 has improved since launch complaints, seeking coding performance comparison to 4.6.

Reddit post claims Claude Opus 4.7 adaptive can close HTML div tags; lacks technical depth or verification.

Claude Opus user observes that AI models apologizing for constraint violations without underlying changes leads to repeated failures in agentic workflows.