← The Independent 100: a study

Taelin

#84 in The Builders on Main, the curator’s rank and section @VictorTaelin Relevance to our work: 5 of 5

Bracketed tags are evidence labels: [Verified] means we read the page, file or API response ourselves, [Reported] means a source says so and we did not check the underlying thing, and [Inference] is our own reasoning.

Credit

The Independent 100 list was curated by dan (@irl_danB) and published on 2026-09-24 at independent.prose.md (announcement). The ranks and section names are his, and number 100 is his open wild-card slot. This study is ours and is not endorsed by him or by anyone on the list.

What they publicly do

Lists Higher Order Company on his GitHub profile and describes himself as author of Kind and HVM [Verified: GitHub profile]. Top committer (870 commits) on Bend 2, a language where a human states intents as laws (LAWS.bend) and the AI must supply a proof (PROOF.bend) that the checker accepts [Verified: bendlang/bend README and contributors API]. Also publishes OptMem, a prompt-plus-script memory for coding agents [Verified: VictorTaelin/OptMem].

Public work

  • Bend 2 (bendlang/bend) repo Apache-2.0 last activity 2026-09-24 (push); v2.0.27 released 2026-09-23 · 22,669 GitHub stars on 2026-09-24

    Dependently typed, parallel language whose checker enforces LAWS.bend via PROOF.bend; compiles to C/CUDA/Metal/JS. Release v2.0.27 on 2026-09-23.

  • OptMem repo no license file last activity 2026-07-31 · 1,544 GitHub stars on 2026-09-24

    Agent memory: a 426-token AGENTS.md block plus one dependency-free Python file; append-only LOG.txt of 280-byte notes, a rebuildable summary tree, 'memo wake/note/nap/recall/zoom/forget'.

  • HVM1 repo MIT last activity 2026-09-16 · 11,348 GitHub stars on 2026-09-24

    Parallel optimal functional runtime in Rust (2022 generation); HVM3/HVM4 are successors.

  • AI-scripts repo no license file last activity 2026-07-24 · 695 GitHub stars on 2026-09-24

    Daily AI coding tools: holefill, refactor with compaction, 'board' multi-advisor file review, 'long' codex loop (goal, work, board review, repeat).

  • Interaction-Calculus repo no license file last activity 2025-11-13 · 955 GitHub stars on 2026-09-24

    Model of computation matching optimal lambda-calculus reduction; theory behind HVM.

  • Kind repo MIT last activity 2025-01-22 · 3,761 GitHub stars on 2026-09-24

    Proof language (earlier generation).

Licenses of GitHub repositories are from the authenticated GitHub license API on 2026-09-25, one row per repository in evidence/licenses.jsonl (“no license file” means the API found none; “license file present, not recognised by GitHub” means GitHub could not classify it). Other licenses are as the source states. Check the license file before reusing anything.

Relevance to our work

5 of 5 lenses: verification, agents, memory, research-method

Bend 2 (Apache-2.0) is an adoptable 'laws plus machine-checked proof' gate for AI-written code, and two franken-research cohort packets assess Bend-based repositories. OptMem is a compact memory design to compare with ee/cass, though it has no license.

Where it already touches our work: Two Franken Research cohort packets assess Bend-based repositories.

Scored 0 to 5 against the study’s relevance rubric: 5 means public, openly licensed work squarely in our lanes that we could study or adopt now, and 1 means little connection to our work. The score is one reviewer’s judgement of how much this public work bears on our own work. It is not a ranking of people and says nothing about the person or the quality of their work.

What we would study next

  1. OptMem's design: append-only fixed-width log, rebuildable summary tree, merges surfaced inline by 'note', one wake budget knob (github.com/VictorTaelin/OptMem). A minimal memory contract to benchmark ee/cass against (wake cost, recall fidelity, no background jobs). No license, so study the design only. Effort: small.
  2. Re-try the beads_bend packet's kill test on Bend v2.0.27 (github.com/bendlang/bend/releases/tag/v2.0.27). Taelin ships near-daily releases, so the beads_bend packet's kill test (a quantified codec round-trip law) should be re-tried on the current checker. Effort: medium.
  3. AI-scripts 'board' and 'long' tools (github.com/VictorTaelin/AI-scripts). A small multi-advisor review panel and a goal/work/review loop; comparable to multi-model review and agent-loop patterns. Effort: small.

Before reusing anything: Bend 2 is Apache-2.0 (OSI) [Verified: GitHub license API, 2026-09-25, evidence/licenses.jsonl]; install is curl | sh from bend-lang.com. OptMem, AI-scripts, Interaction-Calculus and HVM4 have no LICENSE file (all rights reserved by default): study only, do not copy code or the prompt text. Bend performance and checker-speed claims ('checks in under a second') are maintainer claims; our own beads_bend packet records 316-529 s and 11-12.8 GB proof checks at real-tool scale.

Sources

Overall confidence in this record: High. Sources were read on 2026-09-24 and every link was checked on 2026-09-25. The full record is line 84 of people.jsonl.

Limits

Not verified: Bend's C/CUDA speed and 'under a second' checker claims (not reproduced here). The claimed 256-machine Mac mini cluster running agents on Bend 2 and SupGen (own post, 2026-09-24). His post that Claude Code silently routed subagents to an older model (2026-09-25) is his report, not reproduced.

This record comes from a first-pass survey of public material by an AI agent session on 2026-09-24, edited for publication on 2026-09-25. Nothing was installed or run for it. Roles, stars and licenses change; see the study’s limits.

Request a correction

Is something about this record wrong or out of date, or would you rather not be covered? Open a correction (a GitHub issue form with this page’s address filled in). A request to remove a record is honoured without argument; accepted corrections are recorded in the changelog with the date.

Help fix it

Each opens a GitHub issue form with this page’s address filled in. You need a GitHub account.

Made by Joshua Nowak (ZestStream) with AI coding agents. Not affiliated with or funded by Jeffrey Emanuel; Joshua uses his public tools and is a paying subscriber to jeffreys-skills.md.

Source on GitHub MIT licensed