Articles by constantinum
1

FinePDFs: Liberating 3T of the finest tokens from PDFs (huggingface.co)

2

Sentry – No Marketing Mode (sentry.io)

1

Reducto Raises $108M to Shape the Future of AI Document Intelligence (reducto.ai)

2

Iconfinder will permanently close on November 15, 2025 (freepik.com)

2

Retab: The developer starter pack for document processing (retab.com)

1

Understanding why deterministic output from LLMs is nearly impossible (unstract.com)

1

Lumigator: The Dev Tool for AI Model Evaluation (mozilla.ai)

4

Any-agent: A single interface to use and evaluate different agent frameworks (mozilla.ai)

3

Specification Grounding: The Missing Link in Vibe Coding (unstract.com)

66

DeskHog, an open-source developer toy (posthog.com)

2

Groupon Has Become a GLP-1 Affiliate Marketing and Bootleg MS Office Racket (thecaptainslog.io)

1

There's no leader quite like Toyota's Akio Toyoda (ft.com)

5

Why LLMs Are Not (Yet) the Silver Bullet for Unstructured Data Processing (unstract.com)

5

Show HN: Maybe – The personal finance app for everyone (maybefinance.com)

1

Ts_server: A web server proposing a REST API to large language models (bellard.org)

1

Scaling Document Data Extraction with LLMs and Vector Databases (timescale.com)

2

Next-Gen Virtual Office App.Remote Work Reimagined (teracy.io)

2

Open Source API service for document layout analysis, OCR and chunking (chunkr.ai)

4

WordPress Bans WP Engine Customers (searchenginejournal.com)

2

MidrasAI: Simple API for text and image retrieval (github.com/ajac-zero)

1

Playground – AI Design and Editor (playground.com)

1

Unstract: Open-source ETL pipelines to structure unstructured documents (unstract.com)

1

Localops: Streamline Private SaaS Deployments (localops.co)

61

Torchchat: Chat with LLMs Everywhere (github.com/pytorch)

1

Frank Duckworth Obituary (theguardian.com)

2

Amazon Is Investigating Perplexity over Claims of Scraping Abuse (wired.com)

72

Every Way to Get Structured Output from LLMs (boundaryml.com)

2

Comparing Approaches for Using LLMs for Structured Data Extraction from PDFs (unstract.com)

1

Vector DB Retrieval: To chunk or not to chunk (unstract.com)

2

PDF Hell and Practical RAG Applications (unstract.com)

6

How Canva Activates Users (useronboard.com)

1

Exploring the Extractive Capabilities of Large Language Models (unstract.com)

0

Speedometer 3.0: The Best Way yet to Measure Browser Performance (webkit.org)

4

Show HN: LLMWhisperer – Prep complex documents ready for use in LLMs (unstract.com)

1

Mubi: VHS Go (vhs-go.com)

5

Tuta (formerly Tutanota) Mail turns ten today (tuta.com)

0

FastAPI 0.109.0 Release Notes (tiangolo.com)

3

An Overdue Apology: Kenneth Reitz (kennethreitz.org)

2

The Creator of Rails on Internet's Hidden Price: Our Private Data [video] (youtube.com)

3

The myth of the myth of learning styles (nedbatchelder.com)

2

Tabular Secures $26M for Independent Data Platform Based on Apache Iceberg (tabular.io)

46

FastAPI 0.100.0 release notes (tiangolo.com)

2

Elizabeth Holmes Reports to Prison to Begin More Than 11-Year Sentence (nytimes.com)

3

I was voted “least likely to invent his own programming language.” (twitter.com/gvanrossum)

2

Rob Roy's Glacier, ShotoniPhone (suganth.cc)

3

Threads is a Slack replacement designed for makers (threads.com)