← Feed

Posts

• View post

  How did Anthropic solve the problem of an unwieldily big monorepo? They report merges grew from ~2,500 a day to 30,000+, and a team of 7 built Scion to replace git, for agents that are always online. But size alone doesn't explain the move, as Microsoft kept a larger Windows repo on git, and Google's Piper already took 40,000 changes a day in 2016, most of them from automated systems. https://benjaminhan.net/posts/20261017-redesigning-source-control/?utm_source=mastodon&amp...

• View post

  Can AI-written code be reviewed mostly by AI? The Claude Code team lets a classifier decide whether a change needs 0, 1, or 2 human approvals, and its merged PRs more than doubled while the share reverted fell by over 40 percent. Past incidents became a daily eval for the reviewer, and changes that stop review from looping cut PRs needing five or more rounds from 8.9% to 1.5%. https://benjaminhan.net/posts/20261017-claude-code-team-code-review/?utm_source=mastodon&utm_medium=social #A...

• View post

  Well, this week’s #parkrun was a wet one. #running #photo #pnw #rain

• View post

  What helps a frozen LLM player more: the opponent's actual strategy, or notes from a memory model trained to write them? MemoPilot trains that memory model with multi-turn GRPO, and the player reads its notes between games. After one game, the notes already beat both the actual strategy and a human counter-strategy written from it, and MemoPilot remains effective when opponent strategies switch. https://benjaminhan.net/posts/20261018-memopilot-memory-rl/?utm_source=mastodon&utm_...

• View post

  Today concludes my recovery week #1 after the Snohomish River Run #marathon race. After two vaccine shots and a minor throat infection (still coughing a little), #CoachMiles grounded me for most of the week. But today, today, I was determined to run! Just a few miles, today marked the completion of my 2,000 miles for the year! 🎉 #Seattle Ghost Marathon T-6 weeks! #running #photo #pnw

• View post

  24 Alexis Ffrench https://classical.music.apple.com/us/album/1887166427?l=en-US #classicalMusic #appleMusicClassical #commute #piano #crossover

• View post

  Another day no #running, just walking -- in the opposite direction: Volunteer Park in #Seattle! https://benjaminhan.net/posts/20261014-a-hike-to-volunteer-park/?utm_source=mastodon&utm_medium=social #PNW

• View post

  Alarm clock set to 5am. Wish me luck! #marathon #race #snohomishRiverRun #running #photo #snohomish

• View post

  My #marathon #31 today, the Snohomish River Run! Watch time 3:32:49 pace 8’07”/mi, chip time 3:36:12 pace 8’15”/mi. Rank 42/221, 31/147 in males, 2/7 in AG. 6min behind my goal, 11min behind my PR, but 2min faster than my run on this course last year. My best marathon in 2026 so far. (1/4) #running #RaceDay #pnw #photo

• View post

  Not that I have time to play games these days, but I will definitely try this! After a 10-year redemption arc, No Man's Sky is coming to the #iPhone and #iPad - Ars Technica https://arstechnica.com/gaming/2026/10/after-a-10-year-redemption-arc-no-mans-sky-is-coming-to-the-iphone-and-ipad/ #games #videoGames #scifi

• View post

  Softest Thunder Toby Carr https://classical.music.apple.com/us/album/6768690421?l=en-US #classicalMusic #appleMusicClassical #commute #chamber #lute

• View post

  #Photo from #Snohomish River Run #marathon #race photographer: me just about to run away from the dreadful park at ~21mi mark. #running #pnw https://sigmoid.social/@BenjaminHan/117425142319806397

• View post

  No #running, just walking today. (1/2) #seattle #photo #pnw

• View post

  #Mahler: Lieder Andrè Schuen, Daniel Heide https://classical.music.apple.com/us/album/6793983917?l=en-US #classicalMusic #appleMusicClassical #commute #vocal

• View post

  Look at that! #sleep #health #running #ios

• View post

  July #running recap: my second-biggest month of the year. 277.6 → 251.4mi, EG 12,328 → 10,012ft, time 50.2 → 47.4hrs. Highlights: a trip to Pittsburgh and Cupertino: I ran my first Pittsburgh #parkrun in 90%+ humidity and got my fastest 5k in 2026 at 23:50, and revisited Carnegie Mellon after 20+ years! VO₂max jumped to 54.2 (almost an all-time high), and I've run 1,413mi, only 39.2mi behind my yearly goal of 2,500! Heavy training ahead – onward! 💪 https://benjaminhan.net/posts/20260801-j...

• View post

  Fruit Park impression

• View post

  Ran my first-ever #Pittsburgh #Parkrun today! Official time 24:28 over 3.17mi, otherwise my fastest 5k this year: 23:50 at 7'40"/mi (still way off from 21:05 last year)! Not very hot but 90+% humid. It was a cozy Parkrun started by Pat and with volunteer Will today -- they need more runners to join the fun! https://benjaminhan.net/posts/20260725-my-first-pittsburgh-parkrun/?utm_source=mastodon&utm_medium=social #running #photo #personal

• View post

  Flew in on a redeye and had first run in #Pittsburgh in ~20 years! 7 easy miles along the Monongahela on the South Side Trail: lots has changed, but plenty still felt like home. Paused 40min in mile 2 to take a meeting. HR stayed Z2 the whole way and got VO₂max 54, almost my all-time high (54.3)! Can't wait to run tomorrow's #Parkrun @ Pittsburgh! https://benjaminhan.net/posts/20260723-first-run-back-in-pittsburgh/?utm_source=mastodon&utm_medium=social #running #photo #personal

• View post

  How dangerous is an AI-writing detector that is mostly accurate? A profile of Pangram argues a mostly-right detector is worse than a broken one: at a one-in-10,000 false-positive rate across 30M+ students turning in dozens of assignments each, the wrongful accusations never stop. And the black-box verdict leaves the accused nothing to appeal. https://benjaminhan.net/posts/20260705-pangram-ai-detection-accuracy/?utm_source=mastodon&utm_medium=social #AI #Education #Ethics

• View post

  My #AI #running coach project is getting a little out of hand.

• View post

  1/3 This #LongRunSunday I started extending my range to 20mi, #running from Marymoor Light Rail Station in Redmond to somewhere near #UW! My moving time was 3:14:21 pace 9’43"/mi. I stopped at 12mi mark at McDonald's to get my usual — iced coffee and apple pie. I also stopped at random places for photo. Weather was 59F with clouds, wind ~4mph. Super nice weather! #photo #seattle #pnw

• View post

  A multi-agent LLM where each agent learns when to defer to a human, trained with GRPO on a cost-aware reward. Each defer event becomes SFT data, so the model gradually absorbs the human's expertise. Tunable cost knob trades accuracy against human-call budget at deployment, no retraining. https://benjaminhan.net/posts/20260520-adaptive-collaboration-mapo/?utm_source=mastodon&utm_medium=social #ICLR #HumanInTheLoop #AgenticSystems #Metacognition #RL #AI

• View post

  This is Skylark Matthew Guard, Skylark Vocal Ensemble https://classical.music.apple.com/us/album/1895031799?l=en-US #classicalMusic #appleMusicClassical #commute #vocal

• View post

  Missing the round so much I made one today — a relaxed run around Lake Union with photo stops along the way. Complete photo set at https://benjaminhan.net/posts/20260521-lake-union-its-been-a-while/ #running #photo #seattle #pnw #lakeUnion

• View post

  Apparently people have been watching this over a week now. It's a #humanoid #robot flipping packages on a conveyor belt to scan barcode. https://www.youtube.com/watch?v=DL0C8irAWKo Background on #Figure’s robots https://www.youtube.com/watch?v=oJfXvyRHbfY #AI #tech #robotics

• View post

  A meme that's been circulating on Chinese social media since at least 2021. Read 人工智能 (artificial intelligence) backwards and it becomes 能治工人 (can rule over workers). https://benjaminhan.net/posts/20260522-chinese-ai-meme/?utm_source=mastodon&utm_medium=social #AI #Society #FutureOfWork #Meme

• View post

  Are some frontier LLMs better than others at knowing when they're wrong? And is some knowledge harder to self-monitor than other knowledge? An atlas of 33 models × 6 MMLU domains: Anthropic clusters at the top with tight ranges, Gemma trails widely. Applied/Professional is reliably the easiest domain across the panel; Formal Reasoning and Natural Science the hardest. Looking at only aggregate scores per model would hide this. https://benjaminhan.net/posts/20260522-metacognition-atlas/?u...

• View post

  Can an LLM's own pre-solve and post-solve self-assessment signals drive a real test-time control loop? Yes — but only via a per-model SVM trained on labeled correctness, which lifts Sonnet-4.6 from 48.3 to 56.9 pooled accuracy on STEM/code/multimodal. The SVM is precisely the external verifier the "cannot-self-correct" line has argued the loop needs. https://benjaminhan.net/posts/20260522-metacognitive-harness/?utm_source=mastodon&utm_medium=social #Metacognit...

• View post

  What collapses frontier-LLM metacognition more — a vivid survival-threat narrative, or a single "do not refuse" suffix? Factorial isolation across 11 models says: the suffix, conclusively. 8 of 11 lose up to 30.2 accuracy points on refuse/clarify/flag tasks when forced to commit to a confident answer. Anthropic's Constitutional AI is the only family immune — same capability floor as Gemini. https://benjaminhan.net/posts/20260522-compliance-trap/?utm_source=mastodon&amp...