Loading activity…
@ofirpress
SWE-bench, ProgramBench, SWE-sweep. Postdoc @Princeton. PhD @nlpnoah @UW.
AI
The presenter says it covers benchmark results, harness design and his experiments with economic swarms.
AI
A post asks whether the method could identify the best coding-agent checkpoint without post-training every base model.
AI
The benchmark’s creators say it evaluates that ability through 4D inverse graphics.
AI
A post describes the benchmark as focusing on visual bugs, including misrendered maps and misplaced buttons.
AI
A user says the model writes board-analysis code with access to 200,000 real-game board states and engine analysis.
AI
A researcher says an NYC workshop talk will cover when coding-agent teams outperform single agents and where collaboration helps.
AI
A post argues that formalizing this gap is getting harder, but remains important to making AI more useful.
AI
A post argues that human tasks are now too easy to use as AI benchmarks in many domains.
AI
The anniversary post predicts engineers will write less code, listen more to users and direct AI agents.
AI
A poster jokes about “it” getting as smart as a housecat while quoting an OpenAI post and tagging @ylecun.
AI
One user doubts we’ve reached “Navier-Stokes-level coding” yet, while another thinks we already have.
AI
OpenAI says it drew on an independent mathematics advisory group’s advice about how to release the results.
AI
A user jokes that seeing the whole trace could someday sound like a thing of the past.
AI
Reflection says Beam uses 23B active parameters and plans to publish its full weights later in October 2026.
AI
Its creators say the benchmark evaluates that ability through 4D inverse graphics.
AI
A user also claims Astra solved five Erdős problems, two under budget.
AI
A post introducing SWE-sweep says top models score below 5% on the benchmark and calls it one of its creators’ most challenging.
AI
A post warns that scammers are using fake Calendly links while posing as someone it identifies as Reuters journalist Tifafny Wu / Keyaki Yuu.
AI
A post sharing a Microsoft Research paper says Agensh coordinates coding agents through shared state. Agents claim subtasks, share findings, verify results and merge changes asynchronously.
AI
Instagram creator rpn says he challenged five AI models to design and 3D-print bridges that could hold as much weight as possible. Each bridge had to span two feet, use no more than 500 grams of filament and take under 18 hours to print.
