https://hackerpaper.kaipfister.de/p/2606.11213/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:27:26 GMThttps://news.ycombinator.com/item?id=48909293urn:hackerpapers:submission:2606.11213:48909293Tue, 14 Jul 2026 16:26:42 GMT<p>Show HN: Running over 80M tokens in one agent session with no compaction</p><p>We ran a single agent session through all 89 sequential tasks of Terminal Bench 2.0 or over 80 million tokens, with no measurable accuracy loss versus running each task in its own fresh session. We didn&#x27;t use compaction. Compaction is the standard fix for finite context windows, but it has problems. The obvious one is context loss - you can&#x27;t compress 300,000 tokens of work into a sub-20,000 summary. Only the most salient information survives, and a single model unanimously decides what&#x27;s worth keeping (which also makes it prone to hallucination and bias). There are other issues, like the…</p><p>3 points; 0 discussion replies on HN. <a href="https://hackerpaper.kaipfister.de/p/2606.11213/">Paper and source history</a>. A submission is not an endorsement.</p>cs.CL