Emmett @ Aster

@JIACHENLIU8 This is super interesting work! Wondering about the full preservation of a search space -- while it's good to have lossless data, this seems pretty hard for agents to parse and understand Wonder if there's a good format for concisely collapsing search tree info in fewer tokens.

My bet: in the near future, 80%⬆️ of CS research will be done by AI in collaboration with humans. However, today's research ecosystem is still built around the human, not the AI scientist. For example, the 8-page paper PDF is a lossy compression of months of branching exploration into a linear story, optimized for a human reviewer to skim in 30 minutes. It hides two structural taxes: 📖 Storytelling Tax — failures, rejected hypotheses, and dead ends get stripped. On RE-Bench (24,008 runs, 21 frontier models), failed runs = 90.2% of total compute cost, with a 113× median failed-to-success token ratio. Every lab independently rediscovers the same dead ends. 🔧 Engineering Tax — the gap between reviewer-sufficient prose and agent-sufficient spec. Across 8,921 PaperBench requirements (23 ICML'24 papers), only 45.4% are fully specified in the PDF. The rest is tacit lab knowledge. Tolerable when readers were human. Critical now that agents read, reproduce, and extend. We propose ARA: the Agent-Native Research Artifact — replace the narrative PDF with an agent-executable package, in 4 layers: 🧠 structured scientific logic ⚙️ executable code w/ full specs 🌳 exploration graph (every failure preserved) 📊 evidence grounding every claim

@ycombinator @theregbase @ethshea @aresta_katz Congrats on the launch!!

Regbase (@theregbase) tracks new and proposed laws globally helping companies stay informed of new compliance requirements, and generate opportunities to sell to the government. Congrats on the launch, @ethshea and @aresta_katz! ycombinator.com/launches/Pxj-r…

@classiclarryd @photon_mz Happy to contribute!!

The NanoGPT Speedrun WR has broken below 90s, dropping from 92.1 to 89.8 from 4 recent contributions: 1. Tuned Kernels (-0.4s) from @.EmmetBicker on github & AI System Aster 2. Tuned Value Embeds (-0.4s) from @photon_mz 3. Sparse comms for bigram gradients (-0.3s) from @roeeshenberg 4. max_seq_len schedule and increased min lr (-1.2s) from @.dualverse-ai on github & AI System Station

@shivambuilds Thank you!

@asterailabs amazing

GPT 2 Speedrun Record: github.com/KellerJordan/m…

Moved to SF at 18 to join Magic and train LLMs. Two weeks ago, I left & started own AI company. Yesterday, our AI set the GPT-2 Speedrun Record. And that’s just the start. It launches tomorrow. 🚀