New Datology Research: We expose "The Finetuner's Fallacy"The standard approach to domain adaptation (pretrain on web data, finetune on your data) is leaving performance on the table.Mixing just 1-5% domain data into pretraining, then finetuning, produces a strictly better… pic.twitter.com/dXisEz96Vi— DatologyAI (@datologyai) March 18, 2026
Datacurve | The data engine for frontier AI
Custom data for long-horizon reasoning, software engineering, and data science.
Dictionary Compression is finally here, and it's ridiculously good
Dictionary compression could completely change how applications send data over the web. It's recently gained broad support, and offers absurd real-world...
A Short Guide to Data Strikes and Conscious Data Contribution in the Context of 2026 Frontier AI
Back to the basics of data leverage.
DASL — Data-Addressed Structures & Links
A small set of simple, standard primitives to work on content-addressed data.
DASL — Data-Addressed Structures & Links
A small set of simple, standard primitives to work on content-addressed data.
The Dark Forest and Generative AI
Proving you're a human on a web flooded with generative AI content
DeepL Translator | World's Most Accurate Translator
Translate texts & full document files instantly. Accurate translations for individuals and Teams. Millions translate with DeepL every day.
Related work | Data Counterfactuals
Adjacent research areas and starting papers, curated in Semble and synced into the site.
Nobody knows how the whole system works
One of the surprising (at least to me) consequences of the fall of Twitter is the rise of LinkedIn as a social media site. I saw some interesting posts I wanted to call attention to: First, Simon W…
Physics of LM: Part 4.2, Canon Layers at Scale where Synthetic Pretraining Resonates in Reality
From Feeds to Trails
To design the future of social media, rethink the interface before the algorithm.
Why on earth would you publish data as RDF? – Pieter Colpaert
A rebuttal to a blog post of Anne Schuth, I which I argue that RDF will earn its place in real neurosymbolic systems. Its best argument is eventual interoperability across organizations.
About CRDTs • Conflict-free Replicated Data Types
Resources and community around CRDT technology — papers, blog posts, code and more.
Context Rot: How Increasing Input Tokens Impacts LLM Performance
Claude Sonnet 4, GPT-4.1, Qwen3-32B, and Gemini 2.5 Flash on Repeated Words Task
