corpus.blog
41554 blogs · [ { "id": "01a08780-72e0-703e-95bd-58d03a226afa", "title": "Evals Aren't Dead, But They're Low Res", "url": "https://hackbot.dad/writing/towards-high-fidelity-evals/", "published_at": "2026-09-09T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03a56def5", "title": "No Autonomy Without Scalable Oversight", "url": "https://hackbot.dad/writing/no-autonomy-without-scalable-oversight/", "published_at": "2026-04-20T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03a83ac5a", "title": "ProofJudge: Can we align vibe-proving with human taste?", "url": "https://hackbot.dad/writing/proofjudge-and-taste/", "published_at": "2026-03-29T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03b1dd8f0", "title": "The Tests All Pass", "url": "https://hackbot.dad/writing/tests-all-pass/", "published_at": "2026-03-14T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03b77ae33", "title": "Intro to GPUs For the Researcher", "url": "https://hackbot.dad/writing/intro-to-gpus/", "published_at": "2026-01-05T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03ba4d104", "title": "All Reduce Across the Atlantic: Bandwidth in Decentralized Training", "url": "https://hackbot.dad/writing/all-reduce-across-atlantic/", "published_at": "2025-12-12T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03c3568a7", "title": "Twenty Billion Tokens of What, Exactly?", "url": "https://hackbot.dad/writing/20b-tokens-of-what/", "published_at": "2025-12-01T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03c921703", "title": "Pretraining at home: 20B tokens from 222 hours to 12", "url": "https://hackbot.dad/writing/pretraining-at-home/", "published_at": "2025-11-23T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03cc9db5e", "title": "Offsec Evals: Growing Up In The Dark Forest", "url": "https://hackbot.dad/writing/offsec-evals-dark-forest/", "published_at": "2025-10-28T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03cd4d65b", "title": "DiLoCo: Data Parallelism for the Datacenter Poor", "url": "https://hackbot.dad/writing/data-parallelism-for-the-poor/", "published_at": "2025-10-03T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03ce86293", "title": "RL Needed LLMs Because Agency Requires Priors", "url": "https://hackbot.dad/writing/rl-llms-and-priors/", "published_at": "2025-08-19T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03d7d78e4", "title": "GPT-5 is Good, Actually: The Agony and Ecstasy of Public Benchmarks", "url": "https://hackbot.dad/writing/agony-and-ecstasy-evals/", "published_at": "2025-08-17T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03e358341", "title": "The Religious Devotion of Haskell", "url": "https://hackbot.dad/writing/haskell-empathy/", "published_at": "2024-07-01T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03e96121d", "title": "The Input Sanitization Perspective on Prompt Injection", "url": "https://hackbot.dad/writing/prompt-injection/", "published_at": "2023-07-02T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03eb9e9ba", "title": "Infosec's Data Problem", "url": "https://hackbot.dad/writing/infosecs-data-problem/", "published_at": "2022-06-02T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03f50ce24", "title": "Deep Reinforcement Learning for Security: Toward an Autonomous Pentesting Agent", "url": "https://hackbot.dad/writing/towards-autonomous-pentesting/", "published_at": "2020-04-28T00:00:00+00:00" }, { "id": "01a08780-72e0-703e-95bd-58d03fcb1beb", "title": "An ML Eng's Review of OSCP", "url": "https://hackbot.dad/writing/oscp-review/", "published_at": "2020-04-26T00:00:00+00:00" } ] posts Claim your blog
Back to hackbot.dad
Blog · corpus.blog/blogs/hackbot.dad/posts

hackbot.dad

hackbot.dad

2026

Evals Aren't Dead, But They're Low Resoriginal ↗
9 Sept 2026
No Autonomy Without Scalable Oversightoriginal ↗
20 Apr 2026
ProofJudge: Can we align vibe-proving with human taste?original ↗
29 Mar 2026
The Tests All Passoriginal ↗
14 Mar 2026
Intro to GPUs For the Researcheroriginal ↗
5 Jan 2026

2025

All Reduce Across the Atlantic: Bandwidth in Decentralized Trainingoriginal ↗
12 Dec 2025
Twenty Billion Tokens of What, Exactly?original ↗
1 Dec 2025
Pretraining at home: 20B tokens from 222 hours to 12original ↗
23 Nov 2025
Offsec Evals: Growing Up In The Dark Forestoriginal ↗
28 Oct 2025
DiLoCo: Data Parallelism for the Datacenter Poororiginal ↗
3 Oct 2025
RL Needed LLMs Because Agency Requires Priorsoriginal ↗
19 Aug 2025
GPT-5 is Good, Actually: The Agony and Ecstasy of Public Benchmarksoriginal ↗
17 Aug 2025

2024

The Religious Devotion of Haskelloriginal ↗
1 Jul 2024

2023

The Input Sanitization Perspective on Prompt Injectionoriginal ↗
2 Jul 2023

2022

Infosec's Data Problemoriginal ↗
2 Jun 2022

2020

Deep Reinforcement Learning for Security: Toward an Autonomous Pentesting Agentoriginal ↗
28 Apr 2020
An ML Eng's Review of OSCPoriginal ↗
26 Apr 2020