smrtinsert's favorites

1.		I spent 50 hours drawing a line graph (dougmacdowell.com)
		664 points by dougdude3339 20 days ago \| 105 comments
2.		Constraint Decay: The Fragility of LLM Agents in Back End Code Generation (arxiv.org)
		287 points by wek 17 days ago \| 197 comments
3.		Indexing a year of video locally on a 2021 MacBook with Gemma4-31B (50GB swap) (simbastack.com)
		471 points by asenna 20 days ago \| 142 comments
4.		Show HN: Forge – Guardrails take an 8B model from 53% to 99% on agentic tasks (github.com/antoinezambelli)
		687 points by zambelli 22 days ago \| 252 comments
5.		DeepClaude – Claude Code agent loop with DeepSeek V4 Pro (github.com/aattaran)
		678 points by alattaran 38 days ago \| 281 comments
6.		Show HN: Apple's SHARP running in the browser via ONNX runtime web (github.com/bring-shrubbery)
		185 points by bring-shrubbery 38 days ago \| 46 comments
7.		Ask HN: Who is using OpenClaw?
		342 points by misterchocolat 56 days ago \| 383 comments
8.		DuckDB 1.5.2 – SQL database that runs on laptop, server, in the browser (duckdb.org)
		166 points by janandonly 49 days ago \| 52 comments
9.		All your agents are going async (zknill.io)
		134 points by zknill 51 days ago \| 78 comments
10.		OpenClaw isn't fooling me. I remember MS-DOS (flyingpenguin.com)
		307 points by feigewalnuss 51 days ago \| 334 comments
11.		Scan your website to see how ready it is for AI agents (isitagentready.com)
		113 points by WesSouza 54 days ago \| 178 comments
12.		Exploiting the most prominent AI agent benchmarks (rdi.berkeley.edu)
		588 points by Anon84 60 days ago \| 143 comments
13.		Stanford study reveals AI vision models invent images they never see (arxiv.org)
		49 points by LionTurtle13 73 days ago \| 1 comment
14.		How the AI Bubble Bursts (martinvol.pe)
		372 points by martinvol 72 days ago \| 523 comments
15.		We rewrote JSONata with AI in a day, saved $500k/year (reco.ai)
		276 points by cjlm 76 days ago \| 256 comments
16.		$500 GPU outperforms Claude Sonnet on coding benchmarks (github.com/itigges22)
		489 points by yogthos 76 days ago \| 284 comments
17.		Go hard on agents, not on your filesystem (stanford.edu)
		635 points by mazieres 75 days ago \| 329 comments
18.		Show HN: Feather – a fresh Tcl reimplementation (WASM, Go) (feather-lang.dev)
		31 points by dhamidi 5 months ago \| 5 comments
19.		From zero to a RAG system: successes and failures (andros.dev)
		322 points by andros 79 days ago \| 103 comments
20.		Show HN: Atomic – Self-hosted, semantically-connected personal knowledge base (github.com/kenforthewin)
		152 points by kenforthewin 81 days ago \| 27 comments
21.		MCP server that reduces Claude Code context consumption by 98% (mksg.lu)
		570 points by mksglu 3 months ago \| 107 comments
22.		AGENTS.md outperforms skills in our agent evals (vercel.com)
		524 points by maximedupre 4 months ago \| 196 comments
23.		Everything you need to know about act() in React tests (howtotestfrontend.com)
		15 points by howToTestFE 5 months ago
24.		How we lost communication to entertainment (ploum.net)
		701 points by 8organicbits 5 months ago \| 397 comments
25.		Solving a million-step LLM task with zero errors (arxiv.org)
		222 points by Anon84 6 months ago \| 95 comments
26.		Show HN: RowboatX – open-source Claude Code for everyday automations (github.com/rowboatlabs)
		131 points by segmenta 6 months ago \| 42 comments
27.		The Initial Ideal Customer Profile Worksheet (reifyworks.com)
		89 points by mrbbk 7 months ago \| 11 comments
28.		Production RAG: what I learned from processing 5M+ documents (abdellatif.io)
		551 points by tifa2up 7 months ago \| 114 comments
29.		Show HN: Playwright Skill for Claude Code – Less context than playwright-MCP (github.com/lackeyjb)
		189 points by syntax-sherlock 7 months ago \| 45 comments
30.		Tau² benchmark: How a prompt rewrite boosted GPT-5-mini by 22% (quesma.com)
		197 points by blndrt 8 months ago \| 65 comments
		More