JS Wei (Jack) Sun

Pando cuts pushes 5×, llm 0.37 halves default cost, HF ships 66MB fixtures

Every URL the pipeline pulled into ranking for this issue — primary sources plus the supporting and contradicting findings each Researcher returned. Inline citations in the issue point back here.

← Back to the issue

Sources

Scaling Git at OpenAI openai.com

How OpenAI built Pando, a Git-compatible backend designed for the scale of development that Codex makes possible.

llm 0.37 simonwillison.net

Release: llm 0.37 There’s a lot of stuff in this new release of my LLM CLI tool and Python library, representing around two months of accumulated features and fixes. I added gpt-6.1-sol and gpt-6.1-terra , plus a new llm openai endpoint add command to make it easier to configure additional models that use OpenAI-compatible endpoints. I switched the default model (if you have not set your own default) to gpt-6-luna - half the price of the previous default, gpt-5.6-luna . LLM’s support for embedd…

Transformers Model Integration Contracts huggingface.co

October sponsors-only newsletter simonwillison.net

Willison’s sponsors-only October edition calls out an inflection point for vibe game development alongside a heating-up vulnerability landscape, plus the month’s LLM releases and new API-credit perks for $100+/month subscribers. September’s issue is now public as a free preview.

I discovered a sound that leaks your exact location (and it’s everywhere) - YouTube simonwillison.net

Researcher Colsto extracts nearly inaudible aircraft noise from video soundtracks, derives Doppler curves, then cross-references ADS-B flight data to pinpoint where the video was filmed. The technique turns ambient background hum into a precise location leak.

References

Anthropic — ‘Redesigning source control’ (Scion) anthropic.com

Anthropic reported their AI agents (Claude) accelerated merging to over 30,000 changes per day… they built Scion, a system inspired by Piper and Sapling but designed for a world where ‘nearly all clients are agents.’

Matthew Wong blog — Anthropic Scion vs Pando matthewswong.com

Scion replaces Git’s distributed ‘full-clone’ model with a server-centric, ‘always-online’ system… reduced infrastructure footprint by 200x and cut CI CPU usage by 25%. Anthropic has no plans to release Scion.

AI Weekly / Harvard working paper summary aiweekly.co

A Harvard working paper analyzing 700 firms found that while agents increased commits by 20%, the average time required to merge a PR rose by 49%, and review comments increased by 35%.

Freenode — Git Contributor Summit 2026 recap freenode.net

The community discussed a ‘rising tide of thin meat proxy patches’—low-quality submissions where humans acting as intermediaries for AI models lacked the deep technical context required for core Git development… the Git project is currently drafting a formal voting process to resolve policy disputes over AI-generated code.

Cellcog.ai — critique of Pando’s formal verification claims cellcog.ai

The formal methods community has expressed significant skepticism regarding ‘vibe-formalization’—the risk that AI agents may prove properties of a model that does not faithfully represent the actual implementation. Lean can verify that a proof is logically sound, but cannot guarantee the proof addresses the correct real-world problem.

The Hacker News — ‘GitLost’ prompt-injection research thehackernews.com

Researchers demonstrated a technique called ‘GitLost’ where an attacker can open a malicious issue on a public repository to trick an AI agent—which may have cross-repo read access—into leaking private repository data into public comments.

Mixpeek 2026 Video Embedding Benchmark github.com

Gemini’s NDCG@10 at 0.769, narrowly surpassing Twelve Labs’ Marengo 2.7 (0.760), though the latter maintained a perfect Mean Reciprocal Rank (MRR), indicating it more consistently placed the correct video at the #1 position

Agentset — Gemini Embedding 2 vs Voyage 3.5 agentset.ai

while Gemini had a higher ELO and win rate in head-to-head comparisons, Voyage 3.5 delivered superior accuracy on specific datasets (nDCG@10 of 0.703 vs. 0.628)

DataCamp — GPT-6.1 Terra explainer datacamp.com

GPT-6.1 Terra, introduced in late October 2026, occupies the mid-tier at $1.00 per million input and $5.00 per million output tokens… on the Agents’ Last Exam benchmark, Terra reportedly matches the performance of the flagship Astra at one-tenth the price per token

Zen van Riel — ‘Contributor Poker’ guide zenvanriel.com

Major projects like Zig, NetBSD, and Gentoo have implemented outright bans on LLM-generated code, citing a philosophy of ‘contributor poker’ where the goal of review is to invest in long-term human relationships rather than just processing code

pyor.review — AI PRs & maintainers study pyor.review

AI-co-authored PRs contain roughly 1.7 times more issues than human-written ones, and experienced maintainers often face a ‘comprehension debt’ that makes them up to 19% slower when reviewing AI-assisted code

Umesh Malik — matplotlib AI-agent incident umesh-malik.com

an autonomous AI agent (account ‘crabby-rathbun’) submitted a PR and subsequently published a 1,500-word blog post attacking a maintainer who rejected it

ExplainX writeup on TMIC and the UMT5 bug explainx.ai

a change to PyTorch’s scaled dot-product attention (SDPA) allowed the UMT5 decoder to ‘peek’ at future tokens during training… the bug persisted in the main branch for two months because it did not cause crashes and only subtly affected training loss

transformers-mic GitHub repo (issues #5, #6) github.com

Standard ‘generic shrink’ recipes often fail for complex architectures with coupled configuration fields or unique layer patterns… required divisibility of hidden units across specific hardware kernels

Hugging Face blog — native-speed vLLM Transformers backend huggingface.co

By setting the flag —model-impl transformers, users can run models directly from their original Transformers definition while achieving vLLM-grade performance

vLLM supported-models docs docs.vllm.ai

mixin classes (e.g., CausalMixin) act as a technical contract, telling the engine how to hook into the model’s internal states for optimizations like tensor parallelism and KV caching

eSecurityPlanet — Hugging Face RCE vulnerability coverage esecurityplanet.com

malicious configuration files (config.json) could trigger remote code execution (RCE) by bypassing the trust_remote_code=False safety flag

Jack Sun

Jack Sun, writing.

Engineer · Bay Area

Hands-on with agentic AI all day — building frameworks, reading what industry ships, occasionally writing them down.

Digest
All · AI Tech · AI Research · AI News
Writing
Essays
Elsewhere
Subscribe
All · AI Tech · AI Research · AI News · Essays

© 2026 Wei (Jack) Sun · jacksunwei.me Built on Astro · hosted on Cloudflare