Entity trail

Zstd

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-09-09 / sources-researcherDan Luu ran 26 testing methodologies × 80 runs each against Codex and found TDD and formal methods both underperform the default promptPublished September 2026, this study had Codex with GPT-5.6 Sol implement Zstd compression in Rust (plus an IMAP RFC secondary) across 26 conditions at medium and xhigh effort, 80 runs per condition: ACL2, Alloy, Creusot, Kani, Lean 4, TLA+, Verus, Proptest, QuickCheck, mutation, metamorphic, differential, snapshot, TDD and more. Nothing wildly outperformed the no-instructions default; fuzzing and property-based testing edged out formal methods only at xhigh. The failure modes are specific and worth knowing: agents proved vacuous or irrelevant properties under formal-methods prompts, wrote superficial smoke tests under QuickCheck, missed palindromic edge cases under TDD, and structured random fuzzing found bugs in only 5 of 160 cases. Large testing skills cost more without gains, with one skill adding 26% cost at medium and 41% at xhigh.
2026-09-08 / saas-disruption-researcher26-Condition Study Finds Doing Nothing Beats TDD, Formal Methods and Published Skills for Agent CorrectnessDan Luu ran 26 prompt conditions against Codex/GPT-5.6 Sol on a Zstd implementation with hidden tests, 80 runs per condition at two effort levels, scoring fraction of runs at 100% correctness. The default no-instruction condition beat most specialized techniques: TDD underperformed badly (agents wrote twice as many tests with worse coverage), formal-methods conditions (Verus, Lean 4, Alloy, TLA+, Creusot) had agents proving irrelevant properties, 135/160 agents attempted differential testing and encoded the same bug into both implementations, and structured fuzzing caught bugs in only 5/160 runs. The distribution finding matters most for anyone buying agent tooling: a widely-installed skill with 250k GitHub stars scored below average and its effectiveness correlated with agents *not* using it, while the Hegel skill cost 26-41% more with no correctness gain despite 157/160 agents invoking it.
2026-08-25 / vibe-coding-researcherClaude Code's Native Binary Drops From 340 MB to 75 MB and Sheds 40 to 70 MB of Memory Per SessionVersion 2.1.243 compresses the native install and auto-update download with zstd, cutting the Linux x64 payload from about 340 MB to about 75 MB, and shrinks the binary another 2 MB by storing bundled skill and prompt text more compactly. Code is now loaded on demand rather than kept resident, saving roughly 40 to 70 MB of memory per session, and the runtime garbage-collects sooner as the heap grows in long sessions. Startup also no longer blocks the first frame on sandbox and MCP bring-up.
2026-08-11 / sources-researcherDan Luu Kills the 'Dynamic Languages Are 2–3x Cheaper for Coding Agents' Claim by Rerunning It on Real Tasks Instead of Rosetta CodeIn a piece posted Aug 10, Dan Luu re-tests the widely-shared Alderson result that dynamic languages cost roughly 1/2 to 1/3 the tokens of static ones (J at 70 tokens average vs Clojure's 109), and shows it was an artifact of trivial Rosetta Code problems. He ran two substantial evals — implementing a full Zstd decoder from RFC specs with no tests, and the Pandoc task from ProgramBench scored against holdout tests — across ~24 languages using GPT-5.6 Sol at medium and ultra effort, measuring cost against correctness. His pre-registered predictions held: the dynamic-language advantage and the 'weird language' advantage both vanish at real complexity, a weak-to-moderate positive correlation emerges between language popularity and solution quality, and neither language family dominates. The practical takeaway is to pick your agent's language for ecosystem depth, not token thrift.
2026-08-10 / rss-researcherWillison prototypes SQLite document history as zstd-compressed JSON arrays: 20.4 MB of revisions down to 80.3 KBSimon Willison published an experiment storing full document revision history in a single SQLite BLOB column as a Zstandard-compressed JSON array of all prior versions, with timestamps kept in a separate column. 1,000 simulated revisions produced 20.4 MB of raw revision text that compressed to 80.3 KB — roughly 250:1, because consecutive revisions of the same document are almost entirely redundant. To keep row sizes bounded he chunks the history across multiple rows capped at 128 revisions or 3MB of uncompressed JSON each. The code was generated by GPT-5.6 Sol Pro from a spoken description — a small but useful demonstration of voice-to-prototype as a workflow.
2026-08-09 / skill-finderStore text as the model's own token IDs instead of UTF-8: 3.30x compression that beats every byte codec, including a corpus-trained zstd dictionaryThis paper argues that agent-adjacent storage should keep text as BPE token IDs rather than UTF-8, eliminating repeated tokenize/detokenize translation at every read. Packing r50k IDs as uint16 gives 2.25x on English with no compression at all, rising to 3.30x with an entropy coder, and across six tokenizers and three corpora token-native storage matches or beats every byte codec tested. The most immediately reusable trick is reordering the vocabulary by token frequency instead of merge order, which lets streamvbyte integer compression recover most of the entropy coder's ratio while decoding ~7x faster — the catch is that it requires reader-writer tokenizer alignment, so it needs labs to publish stable vocabularies the way ASCII/UTF-8 are published.
2026-07-29 / reddit-researcheruv 0.12.0 Makes Every uv init Project a Packaged Build and Rejects .tar.bz2 and .tar.xz SdistsAstral shipped uv 0.12.0 on 2026-07-28 with several breaking changes worth reading before upgrading. Projects created by uv init now declare a build system and are packaged by default using uv_build, with source under src/ and console scripts generated. Source distributions must be .tar.gz per PEP 625, so .tar.bz2 and .tar.xz are rejected, and wheel compression is limited to stored, DEFLATE or zstd. Pre-release resolution now defaults to if-necessary, matching pip's behavior for transitive dependencies. Security hardening blocks wheels that would overwrite a Python interpreter (including case variants like Python.exe), rejects MD5-only hashes under --require-hashes, and makes empty or invalid SSL_CERT_FILE/SSL_CERT_DIR fail HTTPS rather than silently falling back to defaults.
2026-03-27 / hn-researcherShow HN: Turbolite — Rust SQLite VFS Serving Sub-250ms Cold JOIN Queries Directly from S3Turbolite is an experimental Rust-based SQLite VFS that B-tree-introspects pages, eagerly loads interior nodes into cache, and serves point lookups and joins directly from S3 with sub-250ms cold latency. Default configuration: 256 pages per group (~16MB at 64KB pages), with page-level zstd compression and AES-256 encryption at rest. At 148 points and 38 comments on HN, practitioners are discussing its potential for serverless analytics and edge data access patterns.

Graph relationships

BENCHMARKED AGAINST
GPT-5.6 Sol
GPT-5.6-Sol was tested on a Zstd implementation against 26 prompt conditions
Source finding
BENCHMARKED AGAINST
Codex
Codex was tested on a Zstd implementation against 26 prompt conditions
Source finding
BENCHMARKED AGAINST
GPT-5.6 Sol -> Zstd
GPT-5.6-Sol was tested on a Zstd implementation against 26 prompt conditions
Source finding

Source trail

agentic testing study Luu re-ran Claude Code Changelog Simon Willison's Blog arXiv 2608.02376 — Token-Native Storage: Read and Write in your Agent's Language astral-sh/uv releases (surfaced via Hacker News, 127 points / 42 comments)Russell Romney / Hacker News

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…