Structured PDF-to-JSON: A Guide to Open-Source Extraction Models in 2026
Most enterprise data still sits inside PDFs, scans, and slide decks. Large language models and agents cannot use that data until it becomes structured JSON. Open-source…
Latest
University of Oxford Is Older Than the Aztec Empire and Other Facts of History
Article URL: https://www.smithsonianmag.com/smart-news/university-oxford-older-than-aztec-empire-other-facts-will-change-your-perspective-history-1529607/ Comments URL: https://news.ycombinator.com/item?id=48790913 Points: 5 # Comments: 0
The Log Is the Agent
Article URL: https://arxiv.org/abs/2605.21997 Comments URL: https://news.ycombinator.com/item?id=48790912 Points: 5 # Comments: 0
Backon – Python retry (zero deps, circuit breaker, async native)
Article URL: https://github.com/Llucs/backon Comments URL: https://news.ycombinator.com/item?id=48790898 Points: 4 # Comments: 0
> 1. Draft with a quantized copy of the model itself – it re-syncs for free every training step yeah… that's no DSpark… but every acceleration count…
> 1. Draft with a quantized copy of the model itself - it re-syncs for free every training stepyeah… that's no DSpark… but every acceleration countsI…
Qwen’s Former Lead on What Hybrid Thinking Got Wrong — and Why He Now Backs Agents
Junyang Lin, the former technical lead of Alibaba's Qwen, walked through the model family in a talk "towards a generalist model / agent," then expanded it…
Concurrency plus nvfp4 on Blackwell
Parsed from VLLM log file ~2000 tps in aggregate performing bulk captioning on images. Above is parsed from vllm log while a client runs 30 concurrent…
Success Per Tokens
Work smart more than hard, to expand the pareto frontier (but also work hard)A Pareto Frontier is a set of nondominated (optimal) solutions in multi-objective optimization.…
Results of a small ZBiotics RCT
I ran a small single-blind ZBiotics RCT at a party I hosted recently. I prepared 30 cups, 21 of which contained a placebo and 9 of…
If you're a button, you have one job
Article URL: https://unsung.aresluna.org/if-youre-a-button-you-have-one-job/ Comments URL: https://news.ycombinator.com/item?id=48790689 Points: 19 # Comments: 3
RT DogeDesigner: "One of the most American values that I love is Optimism. We're gonna make the future good." 一 Elon Musk
RT DogeDesigner"One of the most American values that I love is Optimism. We're gonna make the future good." 一 Elon Musk
Efficacy and Safety of Psilocybin in Treatment-Resistant Major Depression
Article URL: https://jamanetwork.com/journals/jamapsychiatry/fullarticle/2846478 Comments URL: https://news.ycombinator.com/item?id=48790583 Points: 5 # Comments: 0
Jippity
Here you go here's what you maximize here it is. Maximize jippity please and thank you. That's what they meant when they were talking about "alignment"…
Getting close to 100K context on 32GB VRAM with Qwen3.6-27 at Q8
Not really a tutorial, but more of sharing my attempts at getting higher contexts on Q8 of Qwen3.6-27 with 32GB VRAM. Disclaimer: Not in-depth research. Crowd…
Record-breaking solo rower Kelsey Pfendler arrives in Hawaii
Article URL: https://www.hawaiinewsnow.com/2026/07/04/record-breaking-solo-rower-kelsey-pfendler-arrives-hawaii/ Comments URL: https://news.ycombinator.com/item?id=48790512 Points: 8 # Comments: 1
Somewhat humbling to have Claude Fable do a final review of some software that you're about to release and have it then find (and fix) FIVE release bl…
Somewhat humbling to have Claude Fable do a final review of some software that you're about to release and have it then find (and fix) FIVE…
sqlite-utils 4.0rc2, mostly written by Claude Fable
I wrote about the sqlite-utils 4.0rc1 release a couple of weeks ago. Since we only have Claude Fable on our Max subscriptions for a few more…
And speaking of Fable. This is what you pay top $ for. How hard is that to notice? Um, not at all? But the WSJ doesn't spell it out, I skimmed past it…
And speaking of Fable.This is what you pay top $ for.How hard is that to notice? Um, not at all? But the WSJ doesn't spell it…
Egg consumption inversely correlated with Alzheimer's
Article URL: https://pubmed.ncbi.nlm.nih.gov/42002260/ Comments URL: https://news.ycombinator.com/item?id=48790349 Points: 14 # Comments: 2
sqlite-utils 4.0rc2
Release: sqlite-utils 4.0rc2 See sqlite-utils 4.0rc2, mostly written by Claude Fable (for about $149.25).
Speaking of Chinese fraud in life sciences: What happened to this thing? It made a massive splash in 2021. Bold hypothesis, human genes in plants, YUG…
Speaking of Chinese fraud in life sciences:What happened to this thing? It made a massive splash in 2021. Bold hypothesis, human genes in plants, YUGE potatoes……
Using "applications" to make a smaller model more effective at bigger tasks.
deleted and reposted due to accidental inclusion of personal information in original post. Demo of an idea I had for my personal JARVIS, quickly put together…
A case for LLMs as Self-predictors
Written as part of the MATS 9.1 extension program, mentored by Richard Ngo. Additional thanks to Maria Kostylew for helpful draft feedback.IntroductionThis post advocates a perspective…
RT Ying Wang✈️ ICML: Warm take: Your world model should never stop learning Introducing AdaJEPA, an adaptive WM that plans, acts, and adapts in a cl…
RT Ying Wang✈️ ICMLWarm take: Your world model should never stop learningIntroducing AdaJEPA, an adaptive WM that plans, acts, and adapts in a closed loop. Every…